notebooks/NERSC-plot_simu_obs_well_data.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T17:27:30.377690Z",
     "start_time": "2018-06-29T17:27:30.371179Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'3.6.1 |Anaconda custom (64-bit)| (default, May 11 2017, 13:09:58) \\n[GCC 4.4.7 20120313 (Red Hat 4.4.7-1)]'"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import sys\n",
    "sys.version"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# import simulation output and observation data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T17:27:33.714188Z",
     "start_time": "2018-06-29T17:27:32.308932Z"
    }
   },
   "outputs": [],
   "source": [
    "# %matplotlib inline\n",
    "# import glob, datetime, ipdb, pickle, pyproj, json, mplleaflet, geopandas, fiona, sys, cartopy, \\\n",
    "#          math, folium, base64\n",
    "import glob\n",
    "import numpy as np\n",
    "import h5py as h5\n",
    "import re\n",
    "import pickle\n",
    "import math\n",
    "\n",
    "# import geopandas as gpd\n",
    "import descartes\n",
    "\n",
    "import matplotlib \n",
    "# matplotlib.use('Agg')\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.gridspec as gridspec\n",
    "import matplotlib.dates as mdates\n",
    "\n",
    "import pandas as pd\n",
    "import ipywidgets as ipyw\n",
    "# import cartopy.crs as ccrs\n",
    "\n",
    "from matplotlib.patches import Patch, Circle\n",
    "from matplotlib.lines import Line2D\n",
    "import datetime\n",
    "from datetime import timedelta\n",
    "from scipy import interpolate\n",
    "# from ipyleaflet import *\n",
    "# from geojson import Feature, FeatureCollection, Point\n",
    "# from geopandas import GeoDataFrame\n",
    "# from shapely.geometry import Point\n",
    "from natsort import natsorted, ns, natsort_keygen\n",
    "from sklearn.metrics import  mean_squared_error\n",
    "# from folium.plugins import FloatImage\n",
    "# from folium import IFrame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T17:27:50.517868Z",
     "start_time": "2018-06-29T17:27:50.503531Z"
    },
    "code_folding": []
   },
   "outputs": [],
   "source": [
    "def batch_delta_to_time(origin, x, time_format, delta_format):\n",
    "    y = []\n",
    "    for ix in x:\n",
    "        if delta_format == \"hours\":\n",
    "            temp_y = origin + timedelta(hours=ix)\n",
    "        elif delta_format == \"days\":\n",
    "            temp_y = origin + timedelta(days=ix)\n",
    "        elif delta_format == \"minutes\":\n",
    "            temp_y = origin + timedelta(minutes=ix)\n",
    "        elif delta_format == \"weeks\":\n",
    "            temp_y = origin + timedelta(weeks=ix)\n",
    "        elif delta_format == \"seconds\":\n",
    "            temp_y = origin + timedelta(seconds=ix)\n",
    "        elif delta_format == \"microseconds\":\n",
    "            temp_y = origin + timedelta(microseconds=ix)\n",
    "        elif delta_format == \"milliseconds\":\n",
    "            temp_y = origin + timedelta(milliseconds=ix)\n",
    "        else:\n",
    "            print(\"Sorry, this naive program only solve single time unit\")\n",
    "        y.append(temp_y.strftime(time_format))\n",
    "    y = np.asarray(y)\n",
    "    return(y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-18T23:11:19.773596Z",
     "start_time": "2018-05-18T23:11:19.771637Z"
    }
   },
   "outputs": [],
   "source": [
    "# print both group and its member names\n",
    "def printname(name):\n",
    "    print(name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def natural_sort(l): \n",
    "    convert = lambda text: int(text) if text.isdigit() else text.lower() \n",
    "    alphanum_key = lambda key: [ convert(c) for c in re.split('([0-9]+)', key) ] \n",
    "    return sorted(l, key = alphanum_key)  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## I/O files and parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-06T19:39:15.838248Z",
     "start_time": "2018-07-06T19:39:15.786645Z"
    }
   },
   "outputs": [],
   "source": [
    "#=======================input==========================\n",
    "case_name = \"HFR_model_100x100x2_cyclic/\"\n",
    "\n",
    "model_dir = \"/global/cscratch1/sd/pshuai/\" + case_name\n",
    "\n",
    "fname_model_origin = model_dir + \"model_origin.txt\"\n",
    "fname_material_h5 = model_dir + \"HFR_material_river.h5\"\n",
    "fname_pflotran_h5 = model_dir + \"pflotran_100x100x2_cyclic_2011_2015.h5\"\n",
    "\n",
    "# fname_300A_h5 = \"/Users/shua784/Paraview/300A/John_case_optim_5/pflotran_bigplume-004.h5\"\n",
    "fname_tec_files = model_dir + \"*.tec\"\n",
    "\n",
    "data_dir = \"/global/project/projectdirs/m1800/pin/Reach_scale_model/data/\"\n",
    "\n",
    "fname_mass1_coord = data_dir + \"MASS1/coordinates.csv\"\n",
    "fname_mass1_data = data_dir + 'MASS1/transient_1976_2016/'\n",
    "fname_300A_well_screen = data_dir + \"well_data/300A_Monitoring_Well_Screen.csv\"\n",
    "fname_300A_well_data = data_dir + 'well_data/SFA_all_wells.csv'\n",
    "fname_300A_well_ID = data_dir + 'well_data/300A_well_coord.csv'\n",
    "fname_300A_spc_u = data_dir + 'well_data/Sample_Data_2015_U.csv'\n",
    "fname_HEIS_well_screen = data_dir + 'well_data/wellCasingAndScreen.csv'\n",
    "fname_HEIS_300A_screen = data_dir + 'well_data/HEIS_300A_well_screen.csv'\n",
    "fname_HEIS_auto_well = data_dir + 'well_data/mvAwln.csv'\n",
    "fname_HEIS_auto_wellID = data_dir + 'well_data/mvAwln_wellID_updated.csv'\n",
    "fname_HEIS_manual_well = data_dir + 'well_data/HYDRAULIC_HEAD_MV.csv'\n",
    "fname_HEIS_manual_wellID = data_dir + 'well_data/HYDRAULIC_HEAD_MV_WellID.csv'\n",
    "fname_GW_chem = data_dir + 'well_data/GW_SAMPLE_RESULT_MV.csv'\n",
    "fname_river_geo = data_dir + 'river_geometry_manual_v2.csv'\n",
    "fname_USGS_well = data_dir + 'well_data/Burns_well_data.csv'\n",
    "fname_USGS_attr = data_dir + 'well_data/Burns_well_attributes.csv'\n",
    "\n",
    "fname_GW_U_summary = data_dir + 'well_data/GW_chemistry/Uranium_summary.csv'\n",
    "fname_GW_Cr_summary = data_dir + 'well_data/GW_chemistry/Chromium_summary.csv'\n",
    "fname_GW_As_summary = data_dir + 'well_data/GW_chemistry/Arsenic_summary.csv'\n",
    "fname_GW_spc_summary = data_dir + 'well_data/GW_chemistry/Specific Conductance_summary.csv'\n",
    "fname_GW_NO3_summary = data_dir + 'well_data/GW_chemistry/Nitrate_summary.csv'\n",
    "# GeoJson file\n",
    "fname_GW_NO3_geojson = data_dir + 'well_data/GW_chemistry/Nitrate_geojson.json'\n",
    "\n",
    "fname_hanford_shp = '/global/project/projectdirs/m1800/pin/Reach_scale_model/data/shapefile/hanfordArea_newCRS.shp'\n",
    "\n",
    "fname_ichemical_csv = data_dir + 'well_data/GW_chemistry/'\n",
    "\n",
    "fname_river_bc_6h = \"/global/project/projectdirs/m1800/pin/Reach_scale_model/Inputs/river_bc/bc_6h_smooth_032807/\"\n",
    "\n",
    "fname_selected_wells_csv = \"/global/project/projectdirs/m1800/pin/Reach_scale_model/results/selected_wells_unique_010107_3m.csv\"\n",
    "#============================output==========================\n",
    "out_dir = \"/global/project/projectdirs/m1800/pin/Reach_scale_model/Outputs/\" + case_name\n",
    "\n",
    "fig_wl = out_dir + 'wl/'\n",
    "fig_simu_obs_wl = out_dir + \"simu_obs_wl/\"\n",
    "\n",
    "fig_obs_head = out_dir + \"obs_head/\"\n",
    "fig_tracer_spc = out_dir + \"tracer_spc/\"\n",
    "fig_tracer_nitrate = out_dir + \"tracer_nitrate/\"\n",
    "fig_tracer_cr = out_dir + 'tracer_cr/'\n",
    "fig_tracer_U = out_dir + 'tracer_U/'\n",
    "fig_tracer_NO3 = out_dir + \"tracer_nitrate/\"\n",
    "\n",
    "result_dir = \"/global/project/projectdirs/m1800/pin/Reach_scale_model/results/\" + case_name\n",
    "\n",
    "# fname_well_simu_data_pk = result_dir + \"well_simu_data.p\"\n",
    "\n",
    "fname_simu_data_df_pickle = result_dir + 'simu_data_df.p'\n",
    "fname_selected_wells_pk = result_dir + 'selected_wells_near_river.p'\n",
    "\n",
    "fname_well_data_df_pickle = result_dir + 'well_data_df_flux_ave.p'\n",
    "# fname_selected_wells_pickle = result_dir + 'selected_wells_for_folium.pk'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-02T18:53:14.640490Z",
     "start_time": "2018-07-02T18:53:14.629109Z"
    }
   },
   "outputs": [],
   "source": [
    "P_atm = 101325\n",
    "rho = 997.16\n",
    "g_const = 9.81\n",
    "\n",
    "z_bot = 50\n",
    "z_top = 150\n",
    "\n",
    "xlen = 60*1000\n",
    "ylen = 60*1000\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T17:32:00.986079Z",
     "start_time": "2018-06-29T17:32:00.976979Z"
    }
   },
   "outputs": [],
   "source": [
    "date_origin = datetime.datetime.strptime(\"2007-03-28 12:00:00\", \"%Y-%m-%d %H:%M:%S\")\n",
    "# read model origin\n",
    "# model_origin = np.genfromtxt(fname_model_origin, delimiter=\" \", skip_header=1)\n",
    "model_origin = [551600, 104500]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## import well data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### import well screen/coordinates"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**combine 300A well screen with HEIS well screen data**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T17:53:29.116640Z",
     "start_time": "2018-06-29T17:53:29.098721Z"
    }
   },
   "outputs": [],
   "source": [
    "screen_300A = pd.read_csv(fname_300A_well_screen, index_col=['Well_ID'])\n",
    "screen_HEIS = pd.read_csv(fname_HEIS_well_screen, index_col=['Well_ID'])\n",
    "\n",
    "screen_HEIS_copy = screen_HEIS[['WellName', 'EASTING', 'NORTHING', 'TopCasing', 'TopScreen', 'BottomScreen']]\n",
    "# rename columns\n",
    "screen_HEIS_copy.columns = ['WellName', 'Easting', 'Northing', 'CASING_ELEVATION', 'SCREEN_ELEV_TOP', 'SCREEN_ELEV_BOTTOM']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**filter out wells without screen information**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-22T18:33:07.547921Z",
     "start_time": "2018-06-22T18:33:07.348243Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# remove duplicates and NaN\n",
    "screen_300A_HEIS = screen_300A.append(screen_HEIS_copy).drop_duplicates('WellName', keep='first') \\\n",
    "                                .dropna(subset=['SCREEN_ELEV_TOP', 'SCREEN_ELEV_BOTTOM']) \\\n",
    "                                .sort_values(by = 'index')\n",
    "# save to csv file\n",
    "screen_300A_HEIS.to_csv(fname_HEIS_300A_screen)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T17:32:22.123497Z",
     "start_time": "2018-06-29T17:32:22.112400Z"
    }
   },
   "outputs": [],
   "source": [
    "screen_300A_HEIS = pd.read_csv(fname_HEIS_300A_screen, index_col=['WellName'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-08T23:32:11.805745Z",
     "start_time": "2018-06-08T23:32:11.788578Z"
    }
   },
   "outputs": [],
   "source": [
    "# well_screen = well_screen.copy()\n",
    "well_screen = pd.read_csv(fname_HEIS_300A_screen, index_col= ['WellName'])\n",
    "\n",
    "well_screen['Easting'] = well_screen['Easting'] - model_origin[0]\n",
    "well_screen['Northing'] = well_screen['Northing'] - model_origin[1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**combine 300A/HEIS/USGS coord **"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T17:32:30.509723Z",
     "start_time": "2018-06-29T17:32:30.316836Z"
    }
   },
   "outputs": [],
   "source": [
    "USGS_well_attr = pd.read_csv(fname_USGS_attr, dtype={'CP_ID_NUM':np.int64})\n",
    "USGS_well_attr.rename(columns={'CP_ID':'WellName','CP_ID_NUM':'Well_ID', 'X_SP_83FT':'Easting', \\\n",
    "                               'Y_SP_83FT':'Northing'}, inplace = True)\n",
    "USGS_well_attr['Well_ID'] = USGS_well_attr['Well_ID'].apply(str)\n",
    "USGS_well_attr['Easting'] = USGS_well_attr['Easting']*0.3048\n",
    "USGS_well_attr['Northing'] = USGS_well_attr['Northing']*0.3048\n",
    "USGS_well_attr.set_index('WellName', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T17:44:03.442028Z",
     "start_time": "2018-06-29T17:44:03.430995Z"
    }
   },
   "outputs": [],
   "source": [
    "screen_300A_HEIS_sub = screen_300A_HEIS[['Well_ID', 'Easting', 'Northing']]\n",
    "\n",
    "USGS_well_attr_sub = USGS_well_attr[['Well_ID', 'Easting', 'Northing']]\n",
    "\n",
    "USGS_well_attr_sub = USGS_well_attr_sub[(USGS_well_attr_sub['Easting'] <= model_origin[0] + xlen) & (USGS_well_attr_sub['Easting'] >= model_origin[0])\\\n",
    "                                           & (USGS_well_attr_sub['Northing'] <= model_origin[1] + ylen) & (USGS_well_attr_sub['Northing'] >= model_origin[1])]\n",
    "\n",
    "coord_300A_HEIS_USGS = screen_300A_HEIS_sub.append(USGS_well_attr_sub).dropna().\\\n",
    "                        drop_duplicates(subset = ['Easting', 'Northing'], keep = 'first')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### import well ID"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-05T17:48:08.557074Z",
     "start_time": "2018-07-05T17:48:08.546777Z"
    }
   },
   "outputs": [],
   "source": [
    "wellID_300A = pd.read_csv(fname_300A_well_ID)\n",
    "wellID_300A.rename(columns = {'wells':'WellName'}, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-05T18:20:38.101791Z",
     "start_time": "2018-07-05T18:20:38.092808Z"
    }
   },
   "outputs": [],
   "source": [
    "wellID_300A.set_index('WellName', inplace=True)\n",
    "wellID_300A_sub = wellID_300A[['Easting', 'Northing']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-02T19:12:13.787533Z",
     "start_time": "2018-07-02T19:12:13.776238Z"
    }
   },
   "outputs": [],
   "source": [
    "HEIS_manual_wellID = pd.read_csv(fname_HEIS_manual_wellID, dtype={'EASTING':np.float64, 'NORTHING':np.float64})\n",
    "HEIS_manual_wellID.rename(columns = {'WELL_ID':'Well_ID', 'WELL_NAME':'WellName', 'NORTHING':'Northing', 'EASTING':'Easting'}, \\\n",
    "                         inplace = True)\n",
    "HEIS_manual_wellID.set_index('WellName', inplace=True)\n",
    "HEIS_manual_wellID_sub = HEIS_manual_wellID[['Well_ID', 'Easting', 'Northing']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-02T19:18:08.309602Z",
     "start_time": "2018-07-02T19:18:08.302179Z"
    }
   },
   "outputs": [],
   "source": [
    "HEIS_auto_wellID = pd.read_csv(fname_HEIS_auto_wellID, na_values=['#REF!'], dtype={'Easting':np.float64, 'Northing':np.float64})\n",
    "HEIS_auto_wellID.rename(columns = {'WellNumber':'Well_ID'}, inplace = True)\n",
    "HEIS_auto_wellID.set_index('WellName', inplace=True)\n",
    "HEIS_auto_wellID_sub = HEIS_auto_wellID[['Well_ID', 'Easting', 'Northing']]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "** store all well ids within domian **"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-02T19:29:54.047475Z",
     "start_time": "2018-07-02T19:29:54.040282Z"
    }
   },
   "outputs": [],
   "source": [
    "All_well_coord = HEIS_manual_wellID_sub.append([wellID_300A_sub, HEIS_auto_wellID_sub, USGS_well_attr_sub]).dropna()\n",
    "\n",
    "All_well_coord = All_well_coord[~All_well_coord.index.duplicated(keep='first')]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-02T19:30:01.108812Z",
     "start_time": "2018-07-02T19:30:00.804889Z"
    }
   },
   "outputs": [],
   "source": [
    "All_well_coord['Easting_rint'] = np.rint(All_well_coord['Easting'])\n",
    "All_well_coord['Northing_rint'] = np.rint(All_well_coord['Northing'])\n",
    "\n",
    "All_well_coord = All_well_coord.drop_duplicates(subset = ['Easting_rint', 'Northing_rint'],keep='first')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### import well obs data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**import well water level data**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T17:33:00.628519Z",
     "start_time": "2018-06-29T17:32:51.572478Z"
    }
   },
   "outputs": [],
   "source": [
    "# import obs well data\n",
    "SFA_well = pd.read_csv(fname_300A_well_data, parse_dates=['DateTime'])\n",
    "# SFA_well['DateTime'] = SFA_well['DateTime'].apply(pd.to_datetime)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T17:33:23.693820Z",
     "start_time": "2018-06-29T17:33:04.404850Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/global/homes/p/pshuai/.conda/envs/widget/lib/python3.6/site-packages/IPython/core/interactiveshell.py:3063: DtypeWarning: Columns (21,22) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  interactivity=interactivity, compiler=compiler, result=result)\n"
     ]
    }
   ],
   "source": [
    "# print(well_data.head(5))\n",
    "HEIS_auto = pd.read_csv(fname_HEIS_auto_well, parse_dates=['procDate'])\n",
    "HEIS_manual = pd.read_csv(fname_HEIS_manual_well, parse_dates=['HYD_DATE_TIME_PST'])\n",
    "\n",
    "## rename column names to contain the following label : **Well_ID, WellName, DateTime, WL**\n",
    "HEIS_manual.rename(columns={'WELL_ID':'Well_ID', 'WELL_NAME': 'WellName', 'HYD_DATE_TIME_PST':'DateTime', \\\n",
    "                           'HYD_HEAD_METERS_NAVD88':'WL'}, inplace = True)\n",
    "HEIS_auto.rename(columns={'WellNumber':'Well_ID', 'procWaterElevation':'WL', 'procDate':'DateTime'}, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T17:33:46.192953Z",
     "start_time": "2018-06-29T17:33:45.048243Z"
    }
   },
   "outputs": [],
   "source": [
    "USGS_well = pd.read_csv(fname_USGS_well, parse_dates=['DATE'])\n",
    "USGS_well.rename(columns={'CP_NUM': 'WellID', 'DATE':'DateTime', \\\n",
    "                           'WLELEVft88':'WL'}, inplace = True)\n",
    "# USGS_well['DateTime'] = USGS_well['DateTime'].apply(pd.to_datetime)\n",
    "USGS_well['WellName'] = ['CP' + str(x).zfill(6) for x in USGS_well['WellID']]\n",
    "USGS_well['WL'] = USGS_well['WL']*0.3048  # convert ft to m"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**import groudwater chemistry data**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "spc_u_data = pd.read_csv(fname_300A_spc_u)\n",
    "spc_u_data.rename(columns={'well ID':'WellName', 'date/time':'DateTime', 'Temp(degrees C)':'Temp', \\\n",
    "                           'SpC(uS/cm)':'Spc', 'U (ug/L)':'U'}  , inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-21T18:51:56.218946Z",
     "start_time": "2018-06-21T18:48:59.506289Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/global/homes/p/pshuai/.conda/envs/nersc/lib/python3.6/site-packages/IPython/core/interactiveshell.py:2717: DtypeWarning: Columns (8,9,10,22,23,27,28,31,33,34,39,41,48,51,59,60,64,65,66,67,70,71,83,84,87,88) have mixed types. Specify dtype option on import or set low_memory=False.\n",
      "  interactivity=interactivity, compiler=compiler, result=result)\n"
     ]
    }
   ],
   "source": [
    "# import groundwater chemistry data\n",
    "GW_chem = pd.read_csv(fname_GW_chem)\n",
    "\n",
    "GW_chem.rename(columns={'SAMP_DATE_TIME': 'DateTime', 'STD_CON_LONG_NAME':'Chemical', \\\n",
    "                        'STD_VALUE_RPTD':'Concentration', 'STD_ANAL_UNITS_RPTD':'Unit', \\\n",
    "                       'SAMP_SITE_NAME': 'WellName', 'SAMP_SITE_ID':'Well_ID'}, inplace=True)\n",
    "\n",
    "GW_chem['DateTime'] = GW_chem['DateTime'].apply(pd.to_datetime)\n",
    "\n",
    "# # split one column to two separate columns\n",
    "# GW_chem['WellName'] = GW_chem['WellName_ID'].str.split('[()]').str[0]\n",
    "# GW_chem['Well_ID']= GW_chem['WellName_ID'].str.split('[()]').str[1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### export selected chemicals"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-21T05:07:51.725905Z",
     "start_time": "2018-05-21T05:07:27.681148Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "chemicals = list(GW_chem.Chemical.unique())\n",
    "chemicals.sort()\n",
    "\n",
    "selected_chemicals = ['Specific Conductance', 'pH Measurement', 'Temperature', 'Nitrate', \\\n",
    "                     'Chromium', 'Turbidity', 'Nitrite', 'Arsenic', 'Uranium']\n",
    "\n",
    "# save each chemical as file\n",
    "for ichemical in selected_chemicals[:]:\n",
    "    table = GW_chem[GW_chem.loc[:,'Chemical'] == ichemical].copy()\n",
    "    \n",
    "    if '/' in ichemical:\n",
    "        fname = data_dir + 'well_data/GW_chemistry/' + ichemical.replace('/',' ') + '.csv'\n",
    "    else:\n",
    "        fname = data_dir + 'well_data/GW_chemistry/' + ichemical + '.csv'\n",
    "    \n",
    "    table.to_csv(fname)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### import mass1 coordicates data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-08T23:41:46.358484Z",
     "start_time": "2018-06-08T23:41:46.341450Z"
    }
   },
   "outputs": [],
   "source": [
    "# read mass1 coordinates\n",
    "section_coord = np.genfromtxt(\n",
    "    fname_mass1_coord, delimiter=\",\", skip_header=1)\n",
    "section_coord[:, 1] = section_coord[:, 1] - model_origin[0]\n",
    "section_coord[:, 2] = section_coord[:, 2] - model_origin[1]\n",
    "line1 = section_coord[0, 1:3] / 1000\n",
    "line2 = section_coord[int(len(section_coord[:, 1]) / 2), 1:3] / 1000\n",
    "line3 = section_coord[-1, 1:3] / 1000\n",
    "\n",
    "line1_x = [line1[0]] * 2\n",
    "line1_y = [line1[1] - 5, line1[1] + 5]\n",
    "line2_x = [line2[0] - 5, line2[0] + 5]\n",
    "line2_y = [line2[1]] * 2\n",
    "line3_x = [line3[0] - 5, line3[0] + 5]\n",
    "line3_y = [line3[1]] * 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T18:05:09.106946Z",
     "start_time": "2018-06-29T18:05:09.096829Z"
    }
   },
   "outputs": [],
   "source": [
    "mass1_coord = pd.read_csv(fname_mass1_coord)\n",
    "mass1_coord.rename(columns={'quadrant': 'Mass_ID', 'easting':'Easting', \\\n",
    "                        'northing':'Northing'}, inplace=True)\n",
    "mass1_coord['Easting'] = mass1_coord['Easting'] - model_origin[0]\n",
    "mass1_coord['Northing'] = mass1_coord['Northing'] - model_origin[1]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## read model dim, time index from hdf5 output file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-29T18:05:19.830538Z",
     "start_time": "2018-06-29T18:05:16.176678Z"
    }
   },
   "outputs": [],
   "source": [
    "# all_h5 = glob.glob(fname_pflotran_h5) # find all \"pflotran*.h5\" files\n",
    "# all_h5 = np.sort(all_h5)\n",
    "\n",
    "input_h5 = h5.File(fname_pflotran_h5, \"r\")\n",
    "\n",
    "# input_h5.visit(printname)\n",
    "\n",
    "x_grids = list(input_h5[\"Coordinates\"]['X [m]'])\n",
    "y_grids = list(input_h5[\"Coordinates\"]['Y [m]'])\n",
    "z_grids = list(input_h5[\"Coordinates\"]['Z [m]'])\n",
    "# input_h5.close()\n",
    "\n",
    "dx = np.diff(x_grids)\n",
    "dy = np.diff(y_grids)\n",
    "dz = np.diff(z_grids)\n",
    "\n",
    "nx = len(dx)\n",
    "ny = len(dy)\n",
    "nz = len(dz)\n",
    "\n",
    "# x,y,z coordinates at cell center\n",
    "x = x_grids[0] + np.cumsum(dx) - 0.5 * dx[0]\n",
    "y = y_grids[0] + np.cumsum(dy) - 0.5 * dy[0]\n",
    "z = z_grids[0] + np.cumsum(dz) - 0.5 * dz[0]\n",
    "\n",
    "# generate mesh grid\n",
    "xx, yy = np.meshgrid(x, y)\n",
    "\n",
    "# create grids (nx*ny*nz, 3)--a list of arrays based nx, ny, nz\n",
    "grids = np.asarray([(x, y, z) for z in range(nz)\n",
    "                    for y in range(ny) for x in range(nx)])\n",
    "\n",
    "groups = list(input_h5.keys()) # create a list with group names\n",
    "time_index = [s for s, s in enumerate(groups) if \"Time:\" in s] # enumerate returns its index (index, string)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-22T18:34:17.640261Z",
     "start_time": "2018-06-22T18:34:17.629385Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "real_time_index = []\n",
    "for itime in np.arange(len(time_index)):\n",
    "    itime_str = list(batch_delta_to_time(date_origin, [float(time_index[itime][7:18])], \"%Y-%m-%d %H:%M:%S\", \"hours\"))\n",
    "    itime_str = ''.join(itime_str)\n",
    "    real_time_index.append(datetime.datetime.strptime(itime_str, \"%Y-%m-%d %H:%M:%S\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## find location of cell based on cell_id"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-22T00:17:07.573109Z",
     "start_time": "2018-06-22T00:17:07.569802Z"
    }
   },
   "outputs": [],
   "source": [
    "df = pd.DataFrame(grids, columns = ['x', 'y', 'z'])\n",
    "\n",
    "def cell_id_to_xyz(cell_id):\n",
    "    xyz = df.iloc[cell_id - 1, :] \n",
    "    print('x loc is:{}, y loc is:{}, z loc is:{}'.format(xyz.x+1, xyz.y+1, xyz.z+1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-22T00:17:50.503478Z",
     "start_time": "2018-06-22T00:17:50.501063Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x loc is:314, y loc is:386, z loc is:33\n"
     ]
    }
   ],
   "source": [
    "cell_id_to_xyz(11751314)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## read river cell information from material h5 file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-20T22:13:28.941530Z",
     "start_time": "2018-06-20T22:13:27.379269Z"
    }
   },
   "outputs": [],
   "source": [
    "# open file for reading\n",
    "material_h5 = h5.File(fname_material_h5, \"r\") \n",
    "# read river cell ids\n",
    "river_cells = []\n",
    "for i_region in list(material_h5['Regions'].keys()):\n",
    "    river_cells = np.append(river_cells, np.asarray(\n",
    "        list(material_h5[\"Regions\"][i_region][\"Cell Ids\"])))\n",
    "river_cells = np.unique(river_cells).astype(int)\n",
    "river_cells = river_cells - 1  # need minus 1 as python index started with 0\n",
    "# label river cells in x-y plane with '1'\n",
    "yx_river = np.asarray([np.nan] * (ny * nx)).reshape(ny, nx) # initialize ny*nx array with nan value, because python read column first (y,x)\n",
    "for icell in river_cells:\n",
    "    yx_river[grids[icell, 1], grids[icell, 0]] = 1\n",
    "    \n",
    "material_h5.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# plot well obs head"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-02T20:38:44.318435Z",
     "start_time": "2018-07-02T20:38:44.316339Z"
    }
   },
   "outputs": [],
   "source": [
    "# selected_wells_iniH = pd.read_csv(fname_selected_wells_csv, index_col='WellName')\n",
    "# coord_300A_HEIS_USGS = selected_wells_iniH\n",
    "coord_300A_HEIS_USGS = All_well_coord\n",
    "selected_wells = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-04T12:28:54.518084Z",
     "start_time": "2018-07-04T00:48:52.261639Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# %matplotlib inline \n",
    "\n",
    "for iwell in coord_300A_HEIS_USGS.index[2235:]:\n",
    "#     iwell = '299-W15-9'\n",
    "    print(iwell)\n",
    "    # read obs.head\n",
    "    if iwell in SFA_well['WellName'].unique():\n",
    "        iobs_head_df = SFA_well[ SFA_well.loc[:, 'WellName'] == iwell].copy()\n",
    "\n",
    "    elif iwell in HEIS_auto['WellName'].unique():\n",
    "        iobs_head_df = HEIS_auto[ HEIS_auto.loc[:, 'WellName'] == iwell].copy()\n",
    "\n",
    "    elif iwell in HEIS_manual['WellName'].unique():\n",
    "        iobs_head_df = HEIS_manual[ HEIS_manual.loc[:, 'WellName'] == iwell].copy()\n",
    "        \n",
    "    elif iwell in USGS_well['WellName'].unique():\n",
    "        iobs_head_df = USGS_well[ USGS_well.loc[:, 'WellName'] == iwell].copy()\n",
    "\n",
    "    else:\n",
    "        iobs_head_df = pd.DataFrame()\n",
    "        print('No well obs data available for' + iwell)\n",
    "        continue\n",
    "   \n",
    "    #     iobs_head_df['DateTime'] = iobs_head_df['DateTime'].apply(pd.to_datetime)\n",
    "    iobs_head_df = iobs_head_df.drop_duplicates('DateTime', keep='first') \\\n",
    "                                .dropna(subset = ['DateTime', 'WL']).set_index('DateTime', drop=False).sort_index()\n",
    "    # filter wells outside of dates\n",
    "#     iobs_head_df = iobs_head_df[(iobs_head_df['DateTime'] >= datetime.date(2005, 1, 1)) \\\n",
    "#                                 & (iobs_head_df['DateTime'] <= datetime.date(2018, 1, 1))]\n",
    "    \n",
    "    if iobs_head_df['WL'].count() > 1 :   \n",
    "        selected_wells.append(iwell)\n",
    "        \n",
    "        # get stat for well\n",
    "        istat = iobs_head_df.WL.describe()\n",
    "        coord_300A_HEIS_USGS.loc[iwell, 'min'] = istat['min']\n",
    "        coord_300A_HEIS_USGS.loc[iwell, 'max'] = istat['max']\n",
    "        coord_300A_HEIS_USGS.loc[iwell, 'mean'] = istat['mean']\n",
    "        coord_300A_HEIS_USGS.loc[iwell, 'median'] = istat['50%']\n",
    "        coord_300A_HEIS_USGS.loc[iwell, 'count'] = istat['count']\n",
    "                \n",
    "        # find closest mass1 location\n",
    "        iwell_easting = coord_300A_HEIS_USGS.loc[iwell, 'Easting'] - model_origin[0]\n",
    "        iwell_northing = coord_300A_HEIS_USGS.loc[iwell, 'Northing'] - model_origin[1]\n",
    "        dist = np.sqrt((mass1_coord['Easting']- iwell_easting)**2 + (mass1_coord['Northing']- iwell_northing)**2)\n",
    "        mass_id = mass1_coord.loc[np.argmin(dist), 'Mass_ID']\n",
    "\n",
    "        mass1_datum_files = glob.glob(fname_mass1_data + \"mass1*.csv\")\n",
    "        mass1_datum_files = natsorted(mass1_datum_files) \n",
    "\n",
    "        for ifile in mass1_datum_files:\n",
    "            if '_' + str(mass_id) in ifile:\n",
    "                river_stage = pd.read_csv(ifile, parse_dates=['date']).dropna()\n",
    "            elif mass_id == 333:\n",
    "                river_stage = pd.read_csv(mass1_datum_files[-1],  parse_dates=['date']).dropna()\n",
    "\n",
    "        river_stage['stage'] = river_stage['stage'] + 1.039 \n",
    "        \n",
    "        # find the mode of frequecy (in sec)\n",
    "        res = iobs_head_df['DateTime'].diff().dt.seconds.value_counts().index[0]\n",
    "        val = iobs_head_df['DateTime'].diff().dt.seconds.value_counts().values[0]\n",
    "        if res <= 86400 and val >=100 :    \n",
    "            style = '-k'\n",
    "        else:\n",
    "            style = '-ko'\n",
    "        \n",
    "        ##================= plot simu vs obs head========================##\n",
    "        fig = plt.figure()\n",
    "        ax1 = fig.add_subplot(2,1,1)\n",
    "\n",
    "        river_stage.plot(x= 'date', y= 'stage', legend=False, ax=ax1, color = 'gray' , lw = 1, rot = 0, alpha = 0.5)\n",
    "\n",
    "        ax1.set_ylabel('River stage (m)')\n",
    "        ax2 = fig.add_subplot(2,1,2, sharex = ax1)\n",
    "        iobs_head_df.plot(x= 'DateTime', y= 'WL', legend=False, ax=ax2, style=style ,\\\n",
    "                          markerfacecolor = \"None\", ms = 3, lw = 1, rot = 0)\n",
    "\n",
    "    \n",
    "        ax1.set_title('Well ' + iwell,  fontweight = 'bold', fontsize = 14)\n",
    "        # format the ticks\n",
    "#         ax2.xaxis.set_major_locator(mdates.YearLocator())\n",
    "#         ax2.xaxis.set_major_formatter(mdates.DateFormatter('%Y'))\n",
    "#         ax2.xaxis.set_tick_params('labelbottom')\n",
    "        \n",
    "#         ax.set_xlim([real_time_6h[0], real_time_6h[-1]])\n",
    "#         ax.set_ylim([104.5, 108.5])\n",
    "        ax2.set_ylabel('Water Level (m)')\n",
    "        ax2.set_xlabel('')\n",
    "        \n",
    "        plt.minorticks_off()\n",
    "        fig.tight_layout()\n",
    "        fig.set_size_inches(8, 5)\n",
    "        fname = fig_obs_head + 'Well ' + iwell\n",
    "        fig.savefig(fname , dpi=100)\n",
    "        plt.close(fig)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-05T18:34:27.168151Z",
     "start_time": "2018-07-05T18:34:27.124515Z"
    }
   },
   "outputs": [],
   "source": [
    "coord_300A_HEIS_USGS.to_csv(data_dir + \"well_data/All_well_coord.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-25T18:20:24.909545Z",
     "start_time": "2018-06-25T18:20:24.901251Z"
    }
   },
   "outputs": [],
   "source": [
    "with open(fname_selected_wells_pickle, 'wb') as f:\n",
    "    pickle.dump(selected_wells, f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## interactive map with folium"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "note: some wells ('399-1-32', '399-1-60', '399-2-10', '399-2-25', '399-2-33', '399-3-26',\n",
    "       '399-3-29', '399-3-37', 'NRG', 'SWS-1') in the 300A area has not been included in the map."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-25T22:38:08.299368Z",
     "start_time": "2018-06-25T22:38:08.290365Z"
    }
   },
   "outputs": [],
   "source": [
    "# with open(fname_selected_wells_pickle, \"rb\") as f:\n",
    "#     selected_wells = pickle.load(f)\n",
    "\n",
    "# selected_wells_df = coord_300A_HEIS_USGS.loc[selected_wells, :]\n",
    "\n",
    "# selected_wells_df['Easting_rint'] = np.rint(selected_wells_df['Easting'])\n",
    "# selected_wells_df['Northing_rint'] = np.rint(selected_wells_df['Northing'])\n",
    "\n",
    "# selected_wells_df.drop_duplicates(subset = ['Easting_rint', 'Northing_rint'],keep='first', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-05T18:56:17.456077Z",
     "start_time": "2018-07-05T18:56:17.452208Z"
    }
   },
   "outputs": [],
   "source": [
    "selected_wells_df = coord_300A_HEIS_USGS.copy().dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-05T20:02:40.395918Z",
     "start_time": "2018-07-05T20:02:40.392528Z"
    }
   },
   "outputs": [],
   "source": [
    "selected_wells_df.drop(['299-W15-9'], inplace=True) # not included because of single measurement"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-05T18:56:18.946484Z",
     "start_time": "2018-07-05T18:56:18.622218Z"
    }
   },
   "outputs": [],
   "source": [
    "# convert easting/northing to lat/long\n",
    "NAD83 = pyproj.Proj(init= 'epsg:2856')\n",
    "WGS84 = pyproj.Proj(init= 'epsg:4326')\n",
    "lon, lat= pyproj.transform(NAD83, WGS84, selected_wells_df['Easting'].values, selected_wells_df['Northing'].values)\n",
    "selected_wells_df['lat'] = lat\n",
    "selected_wells_df['lon'] = lon"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "** customize tile using `attribution` and tileLayer address from [here](http://leaflet-extras.github.io/leaflet-providers/preview/) **"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-05T18:48:32.927413Z",
     "start_time": "2018-07-05T18:48:32.925417Z"
    }
   },
   "outputs": [],
   "source": [
    "# Esri.WorldImagery\n",
    "attr = ('Tiles &copy; Esri &mdash; Source: Esri, i-cubed, USDA, USGS, AEX, GeoEye, Getmapping, Aerogrid, IGN, IGP, UPR-EGP, and the GIS User Community')\n",
    "tiles = 'https://server.arcgisonline.com/ArcGIS/rest/services/World_Imagery/MapServer/tile/{z}/{y}/{x}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Stamen.Terrain\n",
    "attr = ('Map tiles by <a href=\"http://stamen.com\">Stamen Design</a>, <a href=\"http://creativecommons.org/licenses/by/3.0\">CC BY 3.0</a> &mdash; Map data &copy; <a href=\"http://www.openstreetmap.org/copyright\">OpenStreetMap</a>')\n",
    "tiles = 'https://stamen-tiles-{s}.a.ssl.fastly.net/terrain/{z}/{x}/{y}{r}.{ext}'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-07-09T19:12:26.670456Z",
     "start_time": "2018-07-09T19:11:12.478862Z"
    }
   },
   "outputs": [],
   "source": [
    "HFR_map = folium.Map(location=[46.6, -119.5], tiles= tiles,  attr=attr,  zoom_start= 10)\n",
    "\n",
    "df = selected_wells_df.copy()\n",
    "# df = df.head(1)\n",
    "fg = folium.FeatureGroup(name= 'Observed well head')\n",
    "\n",
    "def colorcode(ivalue):\n",
    "    if 'CP' in ivalue :\n",
    "        icolor = 'orange'\n",
    "    else:\n",
    "        icolor = 'blue'\n",
    "    return icolor \n",
    "\n",
    "# resolution, width, height = 75, 7, 3\n",
    "\n",
    "for lat, lon, name in zip(df['lat'], df['lon'], df.index):\n",
    "    \n",
    "    iframe = IFrame('<img src=\"data:image/png;base64,{}\">'.format(base64.b64encode(open(\\\n",
    "                'figures/obs_head/Well {}.png'.format(name), 'rb').read()).decode()), \\\n",
    "                width=800, height=500)\n",
    "    fg.add_child(folium.CircleMarker(location=[lat, lon], \n",
    "#                                      popup=folium.Popup(name),\n",
    "                                     popup = folium.Popup(iframe, max_width=2500),\n",
    "#                                     radius = (value - df['median'].min())/(df['median'].max()-df['median'].min())*10,\n",
    "                                     radius = 5,\n",
    "                                     color = colorcode(name),\n",
    "#                                     fill = False,\n",
    "#                                     fill_color = 'crimson',\n",
    "                                    ))  \n",
    "## add legend to map\n",
    "# url = \"https://github.com/pshuai88/notebook/raw/master/figures/RMSE_legend.png\"\n",
    "# FloatImage(url, bottom = 75, left= 75).add_to(HFR_map)\n",
    "\n",
    "# FloatImage(iframe, bottom = 75, left= 75).add_to(fg)\n",
    "    \n",
    "HFR_map.add_child(fg)\n",
    "\n",
    "## add hanford bc\n",
    "hanford_bc = geopandas.read_file(fname_hanford_shp)\n",
    "hanford_bc_wgs84 = hanford_bc.to_crs({'init': 'epsg:4326'})\n",
    "hanford_bc_json = hanford_bc_wgs84.to_json()\n",
    "# you can use either GeoJson or GeoPandas file\n",
    "folium.GeoJson(hanford_bc_json, name = 'hanford_bc',\n",
    "               style_function=lambda feature: {\n",
    "                'fillColor':'blue',\n",
    "               'color':'white',\n",
    "                   'weight':2,\n",
    "                'fillOpacity':0,\n",
    "               }\n",
    "              ).add_to(HFR_map)\n",
    "\n",
    "## add map layer\n",
    "folium.TileLayer('openstreetmap').add_to(HFR_map)\n",
    "\n",
    "HFR_map.add_child(folium.LayerControl())\n",
    "HFR_map.save('figures/all_obs_wells.html')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# plot all chemistry data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## import well data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-08T23:40:28.686786Z",
     "start_time": "2018-06-08T23:40:21.035472Z"
    }
   },
   "outputs": [],
   "source": [
    "# loading objects    \n",
    "with open(fname_well_data_df_pickle, \"rb\") as f:\n",
    "    well_data_df = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(fname_selected_wells_pk, \"rb\") as f:\n",
    "    selected_obs_wells = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-08T23:41:05.618947Z",
     "start_time": "2018-06-08T23:41:05.449833Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "real_time_6h = []\n",
    "for itime in np.arange(len(well_data_df.index)):\n",
    "    itime_str = list(batch_delta_to_time(date_origin, [float(well_data_df.index[itime])], \"%Y-%m-%d %H:%M:%S\", \"hours\"))\n",
    "    itime_str = ''.join(itime_str)\n",
    "    real_time_6h.append(datetime.datetime.strptime(itime_str, \"%Y-%m-%d %H:%M:%S\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# selected_chemical = 'Uranium'\n",
    "iobs_U = GW_chem[GW_chem.loc[:,'Chemical'] == 'Uranium'].copy()\n",
    "iobs_U.rename(columns={'Concentration': 'U'}, inplace = True)\n",
    "\n",
    "iobs_temp = GW_chem[GW_chem.loc[:,'Chemical'] == 'Temperature'].copy()\n",
    "iobs_temp.rename(columns={'Concentration': 'Temp'}, inplace = True)\n",
    "\n",
    "iobs_spc = GW_chem[GW_chem.loc[:,'Chemical'] == 'Specific Conductance'].copy()\n",
    "iobs_spc.rename(columns={'Concentration': 'Spc'}, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "iobs_cr = GW_chem[GW_chem.loc[:,'Chemical'] == 'Chromium'].copy()\n",
    "iobs_cr.rename(columns={'Concentration': 'Cr'}, inplace = True)\n",
    "\n",
    "iobs_sr = GW_chem[GW_chem.loc[:,'Chemical'] == 'Strontium'].copy()\n",
    "iobs_sr.rename(columns={'Concentration': 'Sr'}, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "SFA_WL = SFA_well.loc[:, ['DateTime', 'WellName', 'WL']].copy()\n",
    "SFA_Temp = SFA_well.loc[:, ['DateTime', 'WellName', 'Temp']].copy()\n",
    "SFA_Spc = SFA_well.loc[:, ['DateTime', 'WellName', 'Spc']].copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# with open(fname_selected_well_screen, \"rb\") as f:\n",
    "#     selected_well_screen = pickle.load(f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## plot wl, temp, spc and other chemical data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "ichemical = 'U'\n",
    "\n",
    "if ichemical == 'Cr':\n",
    "    iobs_chem = iobs_cr\n",
    "    icolor = 'c'\n",
    "    icolumn = 'Cr'\n",
    "    ilabel = 'Chromium ($\\mu$g/L)'\n",
    "elif ichemical == 'Sr':\n",
    "    iobs_chem = iobs_sr\n",
    "    icolor = 'm'\n",
    "    icolumn = 'Sr'\n",
    "    ilabel = 'Strontium ($\\mu$g/L)'  \n",
    "elif ichemical == 'Sr':\n",
    "    iobs_chem = iobs_sr\n",
    "    icolor = 'm'\n",
    "    icolumn = 'Sr'\n",
    "    ilabel = 'Strontium ($\\mu$g/L)' \n",
    "elif ichemical == 'U':\n",
    "    iobs_chem = iobs_U\n",
    "    icolor = 'y'\n",
    "    icolumn = 'U'\n",
    "    ilabel = 'Uranium ($\\mu$g/L)' "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**parasite y-axis, follow this [link](https://stackoverflow.com/questions/7733693/matplotlib-overlay-plots-with-different-scales)**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-10T05:12:05.576437Z",
     "start_time": "2018-06-10T05:02:45.124870Z"
    },
    "code_folding": [],
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "399-2-3\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAh0AAAHyCAYAAABPmnwtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n0CFJknph6JAkSb0wdEiSpF78DYxY0sQmXKSHAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2abc406d79e8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline \n",
    "selected_map_well = []\n",
    "for iwell in selected_obs_wells[:1]:\n",
    "    iwell = '399-2-3' # U\n",
    "#     iwell = '199-H4-8' # Cr\n",
    "#     iwell = '199-D4-86' # Cr\n",
    "#     iwell = '199-B2-14' # Sr\n",
    "#     print(iwell)\n",
    "\n",
    "    # import simulated tracer\n",
    "    itracer_df = well_data_df[iwell, 'Total Tracer [M]', 'flux_ave_tracer'].copy()  \n",
    "    \n",
    "    ######### read obs.head (wl) ############\n",
    "    if iwell in SFA_WL['WellName'].unique():\n",
    "        iobs_head_df = SFA_WL[ SFA_WL.loc[:, 'WellName'] == iwell].copy()\n",
    "        style_wl = '-k'\n",
    "    elif iwell in HEIS_auto['WellName'].unique():\n",
    "        iobs_head_df = HEIS_auto[ HEIS_auto.loc[:, 'WellName'] == iwell].copy()\n",
    "        style_wl = '-k'\n",
    "    elif iwell in HEIS_manual['WellName'].unique():\n",
    "        iobs_head_df = HEIS_manual[ HEIS_manual.loc[:, 'WellName'] == iwell].copy()   \n",
    "        style_wl = 'ko'\n",
    "    else:\n",
    "        iobs_head_df = pd.DataFrame()\n",
    "        print('No well obs data available for' + iwell)\n",
    "        continue\n",
    "   \n",
    "    iobs_head_df['DateTime'] = iobs_head_df['DateTime'].apply(pd.to_datetime)\n",
    "    iobs_head_df = iobs_head_df.drop_duplicates('DateTime', keep='first') \\\n",
    "                                .dropna(subset = ['DateTime', 'WL']).set_index('DateTime', drop=False).sort_index()\n",
    "    \n",
    "    ########### read Spc #############\n",
    "    if iwell in SFA_Spc['WellName'].unique():\n",
    "        iobs_Spc_df = SFA_Spc[SFA_Spc.loc[:, 'WellName'] == iwell].copy()\n",
    "        iobs_Spc_df['Spc'] = iobs_Spc_df['Spc']*1000 # convert ms/cm to us/cm\n",
    "        style_spc = '-b'\n",
    "    elif iwell in iobs_spc['WellName'].unique():\n",
    "        iobs_Spc_df = iobs_spc[iobs_spc.loc[:, 'WellName'] == iwell].copy() \n",
    "        style_spc = 'b^'\n",
    "    else:\n",
    "        iobs_Spc_df = pd.DataFrame()\n",
    "        print('No spc data for well ' + iwell)\n",
    "        continue\n",
    "\n",
    "    iobs_Spc_df['DateTime'] = iobs_Spc_df['DateTime'].apply(pd.to_datetime)    \n",
    "    iobs_Spc_df = iobs_Spc_df.dropna(subset = ['DateTime']).drop_duplicates('DateTime', keep='first')\\\n",
    "        .set_index('DateTime', drop=False).sort_index()\n",
    "    \n",
    "    ########### read Temp #############\n",
    "    if iwell in SFA_Temp['WellName'].unique():\n",
    "        iobs_Temp_df = SFA_Temp[SFA_Temp.loc[:, 'WellName'] == iwell].copy()\n",
    "        style_temp = '-r'\n",
    "    elif iwell in iobs_temp['WellName'].unique():\n",
    "        iobs_Temp_df = iobs_temp[iobs_temp.loc[:, 'WellName'] == iwell].copy() \n",
    "        style_temp = 'ro'\n",
    "    else:\n",
    "        iobs_Temp_df = pd.DataFrame()\n",
    "        print('No temp data for well ' + iwell)\n",
    "        continue\n",
    "\n",
    "    iobs_Temp_df['DateTime'] = iobs_Temp_df['DateTime'].apply(pd.to_datetime)    \n",
    "    iobs_Temp_df = iobs_Temp_df.dropna(subset = ['DateTime']).drop_duplicates('DateTime', keep='first')\\\n",
    "        .set_index('DateTime', drop=False).sort_index()    \n",
    "\n",
    "    if ichemical == 'U':\n",
    "    ############ read Uranimum ############\n",
    "        if iwell in spc_u_data['WellName'].unique():\n",
    "            iobs_chem_df = spc_u_data[spc_u_data.loc[:, 'WellName'] == iwell].copy()\n",
    "#             style_U = 'yo'\n",
    "        elif iwell in iobs_U['WellName'].unique():\n",
    "            iobs_chem_df = iobs_U[iobs_U.loc[:, 'WellName'] == iwell].copy()  \n",
    "#             style_U = 'yo'\n",
    "        else:\n",
    "            iobs_chem_df = pd.DataFrame()\n",
    "            print('No uranium data for well ' + iwell)\n",
    "            continue\n",
    "\n",
    "        iobs_chem_df['DateTime'] = iobs_chem_df['DateTime'].apply(pd.to_datetime)    \n",
    "        iobs_chem_df = iobs_chem_df.dropna(subset = ['DateTime']).drop_duplicates('DateTime', keep='first')\\\n",
    "            .set_index('DateTime', drop=False).sort_index()\n",
    "    else:\n",
    "    # read other chemical    \n",
    "        iobs_chem_df = iobs_chem[iobs_chem.loc[:, 'WellName'] == iwell].copy()\n",
    "        iobs_chem_df['DateTime'] = iobs_chem_df['DateTime'].apply(pd.to_datetime)    \n",
    "        iobs_chem_df = iobs_chem_df.dropna(subset = ['DateTime']).drop_duplicates('DateTime', keep='first')\\\n",
    "            .set_index('DateTime', drop=False).sort_index()\n",
    "\n",
    "    # filter wells outside of dates    \n",
    "    iobs_Spc_df = iobs_Spc_df[(iobs_Spc_df['DateTime'] >= '1/1/2011') \\\n",
    "                                & (iobs_Spc_df['DateTime'] <= '1/1/2016')] \n",
    "    iobs_Temp_df = iobs_Temp_df[(iobs_Temp_df['DateTime'] >= '1/1/2011') \\\n",
    "                                & (iobs_Temp_df['DateTime'] <= '1/1/2016')]   \n",
    "#     iobs_U_df = iobs_U_df[(iobs_U_df['DateTime'] >= '1/1/2011') \\\n",
    "#                                 & (iobs_U_df['DateTime'] <= '1/1/2016')]   \n",
    "    iobs_head_df = iobs_head_df[(iobs_head_df['DateTime'] >= '1/1/2011') \\\n",
    "                                    & (iobs_head_df['DateTime'] <= '1/1/2016')]   \n",
    "    \n",
    "    \n",
    "    if iobs_head_df['WL'].count() >= 100 and iobs_Spc_df['Spc'].isnull().all() == False and iobs_Temp_df['Temp'].isnull().all() == False:   \n",
    "        \n",
    "        print(iwell)\n",
    "        \n",
    "        selected_map_well.append(iwell)\n",
    "#         # find closest mass1 location\n",
    "#         iwell_easting = well_screen.loc[iwell, 'Easting']\n",
    "#         iwell_northing = well_screen.loc[iwell, 'Northing']\n",
    "#         dist = np.sqrt((mass1_coord['Easting']- iwell_easting)**2 + (mass1_coord['Northing']- iwell_northing)**2)\n",
    "#         mass_id = mass1_coord.loc[np.argmin(dist), 'Mass_ID']\n",
    "\n",
    "#         datum_files = glob.glob(fname_river_bc_6h + \"Datum*.txt\")\n",
    "#         datum_files = natsorted(datum_files) \n",
    "\n",
    "#         for ifile in datum_files:\n",
    "#             if '_' + str(mass_id) in ifile:\n",
    "#                 river_stage = pd.read_table(ifile, sep=' ', header=None, names=['time', 'x', 'y', 'wl'])\n",
    "#             elif mass_id == 333:\n",
    "#                 river_stage = pd.read_table(datum_files[-1], sep=' ', header=None, names=['time', 'x', 'y', 'wl'])\n",
    "\n",
    "#         river_stage.time = batch_delta_to_time(date_origin, river_stage.time.astype(np.float64), \"%Y-%m-%d %H:%M:%S\", \"seconds\")\n",
    "#         river_stage['time'] = river_stage['time'].apply(pd.to_datetime)  \n",
    "#         river_stage = river_stage[(river_stage['time'] >= '1/1/2011') \\\n",
    "#                             & (river_stage['time'] <= '1/1/2016')]   \n",
    "        \n",
    "        # find the mode of frequecy (in sec)\n",
    "        res = iobs_head_df['DateTime'].diff().dt.seconds.value_counts().index[0]\n",
    "        val = iobs_head_df['DateTime'].diff().dt.seconds.value_counts().values[0]\n",
    "        if res <= 86400 and val >=100 :    \n",
    "            style = '-k'\n",
    "        else:\n",
    "            style = 'ko'\n",
    "\n",
    " \n",
    "\n",
    "       ## ============= plot simu vs obs temp ===================##\n",
    "        fig, [ax1, ax2]= plt.subplots(2, 1, sharex=True)  # short hand\n",
    "        \n",
    "        ##--------------plot (1,1)---------------##\n",
    "\n",
    "        # plot wl\n",
    "        iobs_head_df.plot(x= 'DateTime', y = 'WL', style=style_wl, lw =1 , ms =3, mfc='none', markeredgewidth= 0.5, ax=ax1)\n",
    "        ax1.set_ylabel('Groundwater level (m)', color='k', fontsize = 14)\n",
    "        ax1.legend(loc=[0.3, 1.01], frameon = False)\n",
    "        ax11 = ax1.twinx()\n",
    "        # plot temp\n",
    "        iobs_Temp_df.plot(x= 'DateTime', y = 'Temp', style=style_temp, lw =2 , ms =5, \n",
    "                          mfc='r', markeredgecolor = 'k', markeredgewidth= 0.5, ax=ax11)        \n",
    "        ax11.set_ylabel('Temperature ($^\\circ$C)', color='r', fontsize = 14)\n",
    "        ax11.tick_params('y', colors='r')\n",
    "        ax11.legend(loc=[0.6, 1.01], frameon = False)\n",
    "        \n",
    "        ##------------ plot (2, 1)---------------##\n",
    "        \n",
    "        axes = [ax2, ax2.twinx(), ax2.twinx()]\n",
    "        fig.subplots_adjust(right=0.75)\n",
    "        axes[2].spines['right'].set_position(('axes', 1.2))\n",
    "        axes[2].set_frame_on(True)\n",
    "        axes[2].patch.set_visible(False)\n",
    "        \n",
    "        # plot spc\n",
    "        iobs_Spc_df.plot(x= 'DateTime', y = 'Spc', style=style_spc, lw =1 , \n",
    "                         ms =5, mfc='b', markeredgecolor = 'k', markeredgewidth= 0.5, ax=axes[0])\n",
    "        axes[0].set_ylabel('Specific conductance ($\\mu$s/cm)', color='b', fontsize = 14)\n",
    "        axes[0].set_xlabel('Year', fontsize = 14)\n",
    "        axes[0].legend(loc=[0, 1.01], frameon = False)\n",
    "\n",
    "#         # plot Uranium\n",
    "#         ax21 = ax2.twinx()\n",
    "#         iobs_U_df.plot(x= 'DateTime', y = 'U', style=style_U, lw =1 , ms =3, mfc='none', markeredgewidth= 0.5, ax=ax21)\n",
    "#         ax21.set_ylabel('Uranium ($\\mu$g/L)', color='y')\n",
    "#         ax21.tick_params('y', colors='y')\n",
    "#         ax21.legend(loc=[0.6, 1.01], frameon = False)\n",
    "        \n",
    "        # plot other chemical     \n",
    "#         ax21 = ax2.twinx()\n",
    "        iobs_chem_df.plot(x= 'DateTime', y = icolumn, style = 'o' + icolor, \n",
    "                          lw =2 , ms =5, mfc=icolor, markeredgecolor = 'k', markeredgewidth= 0.5, ax=axes[1])\n",
    "#         ax21.set_ylim([0, 100])\n",
    "        axes[1].set_ylabel(ilabel, color= icolor, fontsize = 14)\n",
    "        axes[1].tick_params('y', colors= icolor)\n",
    "        axes[1].legend(loc=[0.3, 1.01], frameon = False)  \n",
    "        \n",
    "        # plot groundwater tracer\n",
    "        axes[2].plot(real_time_6h, itracer_df.values, 'g-', label= 'Tracer', lw = 1)\n",
    "        axes[2].set_ylabel('River tracer (-)', color = 'g', fontsize = 14)\n",
    "        axes[2].tick_params('y', colors='g')\n",
    "        axes[2].legend(loc=[0.6, 1.01],  frameon = False)  \n",
    "        \n",
    "        ax1.xaxis.set_major_locator(mdates.YearLocator())\n",
    "        ax1.xaxis.set_major_formatter(mdates.DateFormatter('%Y'))\n",
    "        ax1.xaxis.set_tick_params('labelbottom')\n",
    "        \n",
    "        ax1.set_xlim(['01/01/2011', '01/01/2016'])\n",
    "        plt.setp(axes[0].xaxis.get_majorticklabels(), rotation=0, ha = 'center' )\n",
    "        plt.minorticks_off()\n",
    "        \n",
    "        plt.suptitle('Well ' + iwell, fontweight = 'bold', fontsize = 14)\n",
    "#         fig.tight_layout()\n",
    "\n",
    "        fig.set_size_inches(8, 8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-10T05:12:05.576437Z",
     "start_time": "2018-06-10T05:02:45.124870Z"
    },
    "code_folding": [],
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "        fname = out_dir + 'chemistry/Well_' + iwell + '.png'\n",
    "        fig.savefig(fname, dpi=300)\n",
    "        plt.close(fig)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# export simulated head using .TEC obsvervation files"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will import well observation from .tec outputs."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## assemble all output data file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-24T22:56:15.657058Z",
     "start_time": "2018-05-24T22:56:15.651878Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "tec_files = glob.glob(fname_tec_files)\n",
    "# tec_files = natsorted(tec_files)\n",
    "tec_files = natural_sort(tec_files)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-25T18:04:38.019868Z",
     "start_time": "2018-05-25T18:00:16.977428Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-182.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-213.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-214.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-403.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-405.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-436.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-519.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-551.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-552.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1206.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1238.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1424.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1427.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1460.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1486.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1512.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1515.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1543.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1544.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1545.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1548.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1549.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1550.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1556.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1576.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1579.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1611.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1670.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1671.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1702.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1703.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1705.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1736.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1737.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1738.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1743.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1803.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1836.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1837.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2230.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2261.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2262.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2356.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2387.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2389.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2415.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2416.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2417.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2419.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2420.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2445.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2448.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2449.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2451.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2452.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2478.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2483.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2484.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2507.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2510.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2535.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2536.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2537.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2538.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2539.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2540.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2541.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2567.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2568.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2569.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2570.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2571.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2572.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2573.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2574.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2576.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2578.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2600.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2601.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2603.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2604.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2635.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2694.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2695.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2726.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2727.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2728.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2729.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2760.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2761.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2762.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2763.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2766.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2767.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2793.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2794.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2826.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2827.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2829.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2859.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2860.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2861.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2862.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2892.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-2893.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3441.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3469.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3531.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3559.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3560.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3561.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3562.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3591.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3592.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3595.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3596.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3597.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3623.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3624.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3627.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3628.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3752.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3753.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3784.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3785.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3786.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3817.tec\n",
      "/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-3818.tec\n"
     ]
    }
   ],
   "source": [
    "simu_data_df = pd.DataFrame()\n",
    "\n",
    "for ifile in tec_files[:] :\n",
    "    print(ifile)\n",
    "#     ifile = '/global/cscratch1/sd/pshuai/HFR_model_100x100x2_cyclic/pflotran_100x100x2_cyclic-obs-1238.tec'\n",
    "    \n",
    "    idf_first = pd.read_table(ifile, sep=',',header=None, nrows=1, engine='python')\n",
    "    iheader = list(np.asarray(idf_first.values).flatten())\n",
    "    idf = pd.read_table(ifile, sep='\\s+',header=None, names=iheader, index_col=0, skiprows=1, engine='python')\n",
    "    \n",
    "    simu_data_df = pd.concat([simu_data_df, idf], axis=1, join='outer')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-25T18:06:26.115470Z",
     "start_time": "2018-05-25T18:06:17.326416Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "with open(fname_simu_data_df_pickle, \"wb\") as f:\n",
    "    pickle.dump((simu_data_df), f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## grep well data "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-08T23:39:58.842399Z",
     "start_time": "2018-06-08T23:39:58.819543Z"
    }
   },
   "outputs": [],
   "source": [
    "obs_well_names = [iname for iname in well_screen.index if well_screen.loc[iname, 'SCREEN_ELEV_TOP'] >= z_bot and \\\n",
    "                 well_screen.loc[iname, 'SCREEN_ELEV_BOTTOM'] <= z_top and \\\n",
    "                  (well_screen.loc[iname, 'SCREEN_ELEV_TOP'] - well_screen.loc[iname, 'SCREEN_ELEV_BOTTOM']) >= dz[0]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-25T21:50:32.681717Z",
     "start_time": "2018-05-25T21:50:32.679834Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "obs_var_names = ['Liquid Pressure', 'Total Tracer [M]', 'qlx', 'qly']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**filter out wells far from (>= 2 km) the river**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "selected_obs_wells = []\n",
    "for iwell in obs_well_names[:]:\n",
    "#     print(iwell)\n",
    "    ix = well_screen.loc[iwell, 'Easting']\n",
    "    iy = well_screen.loc[iwell, 'Northing']\n",
    "    min_dist = np.min(np.sqrt((mass1_coord['Easting'] - ix)**2 + (mass1_coord['Northing'] - iy)**2))\n",
    "    if min_dist <= 2000:\n",
    "        selected_obs_wells.append(iwell)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(fname_selected_wells_pk, \"wb\") as f:\n",
    "    pickle.dump((selected_obs_wells), f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**create multiindex for dataframe**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-25T23:02:01.173297Z",
     "start_time": "2018-05-25T23:01:54.703245Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "well_array = []\n",
    "var_array = []\n",
    "well_screen_array = [] # well screen interval sequence\n",
    "\n",
    "for iwell in selected_obs_wells:\n",
    "    for iobs in obs_var_names:\n",
    "        iwell_id =iwell + '_'\n",
    "#         iobs = 'qlx'\n",
    "        # search for columns contain both iwell_id and iobs\n",
    "        iwell_col_ind = [icol for icol in simu_data_df.columns if iwell_id in icol and iobs in icol] \n",
    "        # add screen interval sequence to well name \n",
    "        iwell_screen_array = [iwell_id + str(e) for e in list(np.arange(len(iwell_col_ind)) + 1)]\n",
    "        iwell_array = [iwell]*len(iwell_col_ind)\n",
    "        ivar_array = [iobs]*len(iwell_col_ind)\n",
    "        \n",
    "        var_array += ivar_array\n",
    "        well_array += iwell_array \n",
    "        well_screen_array += iwell_screen_array"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-29T07:19:05.190891Z",
     "start_time": "2018-05-29T07:19:01.171223Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "arrays = [np.array(well_array), np.array(var_array), np.array(well_screen_array)]\n",
    "\n",
    "index = pd.MultiIndex.from_arrays(arrays, names=('WellName', 'VariableName', 'WellInterval'))\n",
    "\n",
    "well_data_df = pd.DataFrame(index=simu_data_df.index, columns=index)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This may take a long time to finish.."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-29T08:44:02.910591Z",
     "start_time": "2018-05-29T07:19:18.224450Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "399-1-1\n",
      "399-1-2\n",
      "399-1-7\n",
      "399-1-10A\n",
      "399-1-11\n",
      "399-1-12\n",
      "399-1-13A\n",
      "399-1-16A\n",
      "399-1-17A\n",
      "399-1-21A\n",
      "399-1-23\n",
      "399-1-32\n",
      "399-1-60\n",
      "399-2-1\n",
      "399-2-2\n",
      "399-2-3\n",
      "399-2-5\n",
      "399-2-7\n",
      "399-2-23\n",
      "399-2-25\n",
      "399-2-32\n",
      "399-2-33\n",
      "399-2-34\n",
      "399-2-37\n",
      "399-3-9\n",
      "399-3-12\n",
      "399-3-18\n",
      "399-3-19\n",
      "399-3-20\n",
      "399-3-21\n",
      "399-3-22\n",
      "399-3-34\n",
      "399-3-37\n",
      "399-4-1\n",
      "399-4-7\n",
      "399-4-9\n",
      "399-4-14\n",
      "399-5-1\n",
      "199-D5-13\n",
      "199-D5-14\n",
      "199-D8-4\n",
      "199-D8-5\n",
      "199-D8-55\n",
      "199-D8-6\n",
      "199-F7-2\n",
      "199-H3-2A\n",
      "199-H3-2C\n",
      "199-H4-10\n",
      "199-H4-11\n",
      "199-H4-12A\n",
      "199-H4-12C\n",
      "199-H4-13\n",
      "199-H4-14\n",
      "199-H4-15A\n",
      "199-H4-16\n",
      "199-H4-17\n",
      "199-H4-18\n",
      "199-H4-6\n",
      "199-H4-7\n",
      "199-H4-8\n",
      "199-H4-9\n",
      "199-K-19\n",
      "199-K-22\n",
      "199-K-30\n",
      "199-K-31\n",
      "199-K-36\n",
      "199-N-14\n",
      "199-N-16\n",
      "199-N-18\n",
      "199-N-25\n",
      "199-N-26\n",
      "199-N-27\n",
      "199-N-28\n",
      "199-N-29\n",
      "199-N-31\n",
      "199-N-32\n",
      "199-N-33\n",
      "199-N-34\n",
      "199-N-36\n",
      "199-N-37\n",
      "199-N-39\n",
      "199-N-40\n",
      "199-N-41\n",
      "199-N-42\n",
      "199-N-49\n",
      "199-N-50\n",
      "199-N-51\n",
      "199-N-52\n",
      "199-N-53\n",
      "199-N-54\n",
      "199-N-55\n",
      "199-N-56\n",
      "199-N-57\n",
      "199-N-58\n",
      "199-N-59\n",
      "199-N-60\n",
      "199-N-61\n",
      "199-N-62\n",
      "199-N-63\n",
      "199-N-64\n",
      "199-N-65\n",
      "199-N-66\n",
      "199-N-67\n",
      "199-N-69\n",
      "199-N-70\n",
      "199-N-71\n",
      "199-N-72\n",
      "199-N-73\n",
      "199-N-74\n",
      "199-N-75\n",
      "199-N-76\n",
      "399-1-15\n",
      "399-1-16B\n",
      "399-1-17B\n",
      "399-1-17C\n",
      "399-1-18A\n",
      "399-1-18B\n",
      "399-1-18C\n",
      "399-1-6\n",
      "399-1-8\n",
      "399-1-9\n",
      "399-3-10\n",
      "399-4-10\n",
      "399-4-11\n",
      "399-1-14A\n",
      "399-1-21B\n",
      "399-8-5A\n",
      "399-8-5B\n",
      "399-8-5C\n",
      "199-N-77\n",
      "199-N-81\n",
      "199-N-22\n",
      "199-N-30\n",
      "199-N-43\n",
      "199-N-47\n",
      "199-N-48\n",
      "399-1-10B\n",
      "399-1-13B\n",
      "399-1-14B\n",
      "399-3-11\n",
      "399-5-4B\n",
      "399-6-2\n",
      "699-37-E4\n",
      "199-K-109A\n",
      "199-K-110A\n",
      "199-K-111A\n",
      "199-K-106A\n",
      "199-K-107A\n",
      "199-K-108A\n",
      "199-N-103A\n",
      "199-N-104A\n",
      "199-N-105A\n",
      "199-N-106A\n",
      "199-D8-68\n",
      "199-D8-69\n",
      "199-D8-70\n",
      "199-D8-71\n",
      "199-H4-63\n",
      "199-H4-64\n",
      "199-H3-3\n",
      "199-H3-4\n",
      "199-H3-5\n",
      "199-K-112A\n",
      "199-K-113A\n",
      "199-K-114A\n",
      "199-K-115A\n",
      "199-K-116A\n",
      "199-K-117A\n",
      "199-K-118A\n",
      "199-K-119A\n",
      "199-K-120A\n",
      "199-K-121A\n",
      "199-K-122A\n",
      "199-K-123A\n",
      "199-K-124A\n",
      "199-D4-1\n",
      "199-D4-4\n",
      "199-D4-5\n",
      "199-D4-6\n",
      "199-D4-7\n",
      "199-D4-8\n",
      "199-D4-9\n",
      "199-D4-10\n",
      "199-D4-11\n",
      "199-D4-12\n",
      "199-D4-13\n",
      "199-D4-14\n",
      "199-D4-15\n",
      "199-D3-2\n",
      "199-D4-17\n",
      "199-D4-18\n",
      "199-K-125A\n",
      "199-D5-36\n",
      "199-D5-37\n",
      "199-D4-19\n",
      "199-D5-38\n",
      "199-D5-39\n",
      "199-D5-40\n",
      "199-D4-20\n",
      "199-D5-41\n",
      "199-D5-42\n",
      "199-D5-43\n",
      "199-D5-44\n",
      "199-D4-21\n",
      "199-H4-65\n",
      "199-K-126\n",
      "199-D4-22\n",
      "199-D4-23\n",
      "199-D4-25\n",
      "199-D4-26\n",
      "199-D4-27\n",
      "199-D4-31\n",
      "199-D4-32\n",
      "199-D4-34\n",
      "199-D4-35\n",
      "199-D4-36\n",
      "199-D4-38\n",
      "199-D4-39\n",
      "199-D2-8\n",
      "199-D4-48\n",
      "199-D4-62\n",
      "199-D4-68\n",
      "199-D4-78\n",
      "199-D4-80\n",
      "199-D4-81\n",
      "199-D3-3\n",
      "199-D4-82\n",
      "199-D3-4\n",
      "199-D4-83\n",
      "199-D4-84\n",
      "199-D4-85\n",
      "199-D4-86\n",
      "199-D8-72\n",
      "199-K-129\n",
      "199-K-130\n",
      "699-96-44\n",
      "699-96-45\n",
      "199-D5-93\n",
      "199-N-136\n",
      "199-N-138\n",
      "199-N-139\n",
      "199-N-146\n",
      "199-K-137\n",
      "199-K-138\n",
      "199-K-139\n",
      "199-K-140\n",
      "199-N-147\n",
      "199-K-141\n",
      "199-K-142\n",
      "199-K-143\n",
      "199-K-144\n",
      "199-K-145\n",
      "199-K-146\n",
      "199-K-147\n",
      "199-K-148\n",
      "199-K-149\n",
      "199-K-150\n",
      "199-K-151\n",
      "199-K-152\n",
      "199-K-153\n",
      "199-K-154\n",
      "199-K-156\n",
      "199-K-157\n",
      "199-D5-97\n",
      "199-D5-98\n",
      "199-D5-99\n",
      "199-D2-11\n",
      "199-D5-103\n",
      "199-D5-104\n",
      "199-K-158\n",
      "199-D5-107\n",
      "199-D5-108\n",
      "199-D5-109\n",
      "199-D5-110\n",
      "199-D5-114\n",
      "199-D5-115\n",
      "699-100-43B\n",
      "699-99-42B\n",
      "699-99-41\n",
      "699-99-44\n",
      "699-98-43\n",
      "699-97-41\n",
      "699-98-46\n",
      "699-97-45\n",
      "699-95-45\n",
      "699-97-48B\n",
      "699-95-51\n",
      "699-97-43B\n",
      "699-94-41\n",
      "699-101-45\n",
      "699-95-48\n",
      "699-96-52B\n",
      "699-98-51\n",
      "199-B8-7\n",
      "199-B8-8\n",
      "199-D5-119\n",
      "199-D5-120\n",
      "199-D5-121\n",
      "199-D5-122\n",
      "199-K-159\n",
      "199-K-160\n",
      "199-K-161\n",
      "199-K-162\n",
      "199-K-163\n",
      "199-N-159\n",
      "199-K-164\n",
      "199-D5-123\n",
      "199-D5-125\n",
      "199-D5-126\n",
      "199-K-165\n",
      "199-K-166\n",
      "199-K-168\n",
      "199-N-165\n",
      "199-K-169\n",
      "199-K-170\n",
      "199-K-171\n",
      "199-K-172\n",
      "199-F8-7\n",
      "199-K-173\n",
      "199-N-167\n",
      "199-N-169\n",
      "199-N-170\n",
      "199-N-171\n",
      "199-N-172\n",
      "199-N-173\n",
      "199-K-174\n",
      "199-K-175\n",
      "199-H3-25\n",
      "199-K-178\n",
      "199-K-179\n",
      "199-K-180\n",
      "199-N-229\n",
      "199-N-211\n",
      "199-N-201\n",
      "199-N-248\n",
      "199-N-268\n",
      "199-N-280\n",
      "199-N-298\n",
      "199-N-316\n",
      "199-N-332\n",
      "199-N-342\n",
      "199-N-349\n",
      "199-N-348\n",
      "199-N-347\n",
      "199-N-346\n",
      "199-N-350\n",
      "199-N-351\n",
      "199-N-352\n",
      "199-N-353\n",
      "199-N-354\n",
      "199-N-355\n",
      "199-N-356\n",
      "199-N-357\n",
      "199-N-358\n",
      "199-N-360\n",
      "199-N-361\n",
      "199-N-362\n",
      "199-N-363\n",
      "199-N-364\n",
      "199-N-365\n",
      "199-N-366\n",
      "199-N-367\n",
      "199-K-181\n",
      "199-K-182\n",
      "199-H1-43\n",
      "199-B5-5\n",
      "199-B3-50\n",
      "199-B5-6\n",
      "199-B8-9\n",
      "199-D3-5\n",
      "199-D5-133\n",
      "199-D5-132\n",
      "199-D6-3\n",
      "199-D5-134\n",
      "199-D5-141\n",
      "199-H3-6\n",
      "199-H3-7\n",
      "199-H6-3\n",
      "199-H6-4\n",
      "199-H1-7\n",
      "199-H2-1\n",
      "199-H3-9\n",
      "199-H3-10\n",
      "399-1-54\n",
      "399-1-55\n",
      "399-1-56\n",
      "399-1-57\n",
      "399-1-58\n",
      "399-6-3\n",
      "399-1-59\n",
      "399-4-15\n",
      "399-3-33\n",
      "199-B2-14\n",
      "199-K-183\n",
      "199-K-184\n",
      "199-K-185\n",
      "199-K-186\n",
      "199-K-187\n",
      "199-K-188\n",
      "199-K-189\n",
      "199-K-190\n",
      "199-K-191\n",
      "199-K-192\n",
      "199-K-193\n",
      "199-K-194\n",
      "199-K-195\n",
      "199-K-196\n",
      "199-K-197\n",
      "199-K-198\n",
      "199-K-199\n",
      "199-B2-15\n",
      "199-B2-16\n",
      "199-B3-51\n",
      "199-B4-14\n",
      "199-F5-52\n",
      "199-F5-53\n",
      "199-F5-54\n",
      "199-K-200\n",
      "199-K-201\n",
      "199-B3-52\n",
      "199-D8-101\n",
      "199-D5-142\n",
      "199-H4-84\n",
      "199-H4-83\n",
      "199-D5-140\n",
      "199-F5-55\n",
      "199-F5-56\n",
      "199-N-182\n",
      "199-N-183\n",
      "199-N-184\n",
      "199-N-186\n",
      "199-N-187\n",
      "199-N-188\n",
      "199-N-189\n",
      "399-6-5\n",
      "199-K-202\n",
      "199-K-203\n",
      "199-K-204\n",
      "199-K-205\n",
      "199-K-206\n",
      "199-K-207\n",
      "199-K-208\n",
      "199-D5-143\n",
      "199-D5-144\n",
      "199-H4-85\n",
      "199-H4-86\n",
      "199-D5-145\n",
      "199-D5-146\n",
      "199-D5-147\n",
      "199-D5-148\n",
      "199-B4-16\n",
      "199-B4-18\n",
      "199-B5-9\n",
      "199-B5-10\n",
      "199-B5-11\n",
      "199-B5-12\n",
      "199-B5-13\n",
      "199-B5-14\n",
      "199-H4-90\n",
      "199-H4-91\n",
      "199-D5-153\n",
      "199-D5-154\n",
      "699-97-60\n",
      "699-97-61\n",
      "199-K-221\n",
      "199-K-222\n",
      "399-1-67\n",
      "199-H6-7\n",
      "199-H6-8\n",
      "199-D5-155\n",
      "199-D5-156\n",
      "199-D5-157\n",
      "199-D5-158\n"
     ]
    }
   ],
   "source": [
    "for iwell in selected_obs_wells[:]:\n",
    "    print(iwell)\n",
    "    for iobs in obs_var_names[:]:\n",
    "        iwell_id =iwell + '_'\n",
    "        # search for columns contain both iwell_id and iobs\n",
    "        iwell_col_ind = [icol for icol in simu_data_df.columns if iwell_id in icol and iobs in icol] \n",
    "        iwell_screen_array = [iwell_id + str(e) for e in list(np.arange(len(iwell_col_ind)) + 1)]\n",
    "\n",
    "        for i in np.arange(len(iwell_col_ind))[:]:\n",
    "            well_data_df[iwell, iobs, iwell_screen_array[i]] = simu_data_df[iwell_col_ind[i]].values\n",
    "\n",
    "well_data_df = well_data_df.sort_index(axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-29T05:11:08.694778Z",
     "start_time": "2018-05-29T05:11:08.669183Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "well_data_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**calculate flux averaged tracer conc.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-29T08:58:47.027002Z",
     "start_time": "2018-05-29T08:44:09.849242Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "399-1-1\n",
      "399-1-2\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/global/common/cori/software/python/3.6-anaconda-5.2/lib/python3.6/site-packages/IPython/core/interactiveshell.py:2817: PerformanceWarning: indexing past lexsort depth may impact performance.\n",
      "  raw_cell, store_history, silent, shell_futures)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "399-1-7\n",
      "399-1-10A\n",
      "399-1-11\n",
      "399-1-12\n",
      "399-1-13A\n",
      "399-1-16A\n",
      "399-1-17A\n",
      "399-1-21A\n",
      "399-1-23\n",
      "399-1-32\n",
      "399-1-60\n",
      "399-2-1\n",
      "399-2-2\n",
      "399-2-3\n",
      "399-2-5\n",
      "399-2-7\n",
      "399-2-23\n",
      "399-2-25\n",
      "399-2-32\n",
      "399-2-33\n",
      "399-2-34\n",
      "399-2-37\n",
      "399-3-9\n",
      "399-3-12\n",
      "399-3-18\n",
      "399-3-19\n",
      "399-3-20\n",
      "399-3-21\n",
      "399-3-22\n",
      "399-3-34\n",
      "399-3-37\n",
      "399-4-1\n",
      "399-4-7\n",
      "399-4-9\n",
      "399-4-14\n",
      "399-5-1\n",
      "199-D5-13\n",
      "199-D5-14\n",
      "199-D8-4\n",
      "199-D8-5\n",
      "199-D8-55\n",
      "199-D8-6\n",
      "199-F7-2\n",
      "199-H3-2A\n",
      "199-H3-2C\n",
      "199-H4-10\n",
      "199-H4-11\n",
      "199-H4-12A\n",
      "199-H4-12C\n",
      "199-H4-13\n",
      "199-H4-14\n",
      "199-H4-15A\n",
      "199-H4-16\n",
      "199-H4-17\n",
      "199-H4-18\n",
      "199-H4-6\n",
      "199-H4-7\n",
      "199-H4-8\n",
      "199-H4-9\n",
      "199-K-19\n",
      "199-K-22\n",
      "199-K-30\n",
      "199-K-31\n",
      "199-K-36\n",
      "199-N-14\n",
      "199-N-16\n",
      "199-N-18\n",
      "199-N-25\n",
      "199-N-26\n",
      "199-N-27\n",
      "199-N-28\n",
      "199-N-29\n",
      "199-N-31\n",
      "199-N-32\n",
      "199-N-33\n",
      "199-N-34\n",
      "199-N-36\n",
      "199-N-37\n",
      "199-N-39\n",
      "199-N-40\n",
      "199-N-41\n",
      "199-N-42\n",
      "199-N-49\n",
      "199-N-50\n",
      "199-N-51\n",
      "199-N-52\n",
      "199-N-53\n",
      "199-N-54\n",
      "199-N-55\n",
      "199-N-56\n",
      "199-N-57\n",
      "199-N-58\n",
      "199-N-59\n",
      "199-N-60\n",
      "199-N-61\n",
      "199-N-62\n",
      "199-N-63\n",
      "199-N-64\n",
      "199-N-65\n",
      "199-N-66\n",
      "199-N-67\n",
      "199-N-69\n",
      "199-N-70\n",
      "199-N-71\n",
      "199-N-72\n",
      "199-N-73\n",
      "199-N-74\n",
      "199-N-75\n",
      "199-N-76\n",
      "399-1-15\n",
      "399-1-16B\n",
      "399-1-17B\n",
      "399-1-17C\n",
      "399-1-18A\n",
      "399-1-18B\n",
      "399-1-18C\n",
      "399-1-6\n",
      "399-1-8\n",
      "399-1-9\n",
      "399-3-10\n",
      "399-4-10\n",
      "399-4-11\n",
      "399-1-14A\n",
      "399-1-21B\n",
      "399-8-5A\n",
      "399-8-5B\n",
      "399-8-5C\n",
      "199-N-77\n",
      "199-N-81\n",
      "199-N-22\n",
      "199-N-30\n",
      "199-N-43\n",
      "199-N-47\n",
      "199-N-48\n",
      "399-1-10B\n",
      "399-1-13B\n",
      "399-1-14B\n",
      "399-3-11\n",
      "399-5-4B\n",
      "399-6-2\n",
      "699-37-E4\n",
      "199-K-109A\n",
      "199-K-110A\n",
      "199-K-111A\n",
      "199-K-106A\n",
      "199-K-107A\n",
      "199-K-108A\n",
      "199-N-103A\n",
      "199-N-104A\n",
      "199-N-105A\n",
      "199-N-106A\n",
      "199-D8-68\n",
      "199-D8-69\n",
      "199-D8-70\n",
      "199-D8-71\n",
      "199-H4-63\n",
      "199-H4-64\n",
      "199-H3-3\n",
      "199-H3-4\n",
      "199-H3-5\n",
      "199-K-112A\n",
      "199-K-113A\n",
      "199-K-114A\n",
      "199-K-115A\n",
      "199-K-116A\n",
      "199-K-117A\n",
      "199-K-118A\n",
      "199-K-119A\n",
      "199-K-120A\n",
      "199-K-121A\n",
      "199-K-122A\n",
      "199-K-123A\n",
      "199-K-124A\n",
      "199-D4-1\n",
      "199-D4-4\n",
      "199-D4-5\n",
      "199-D4-6\n",
      "199-D4-7\n",
      "199-D4-8\n",
      "199-D4-9\n",
      "199-D4-10\n",
      "199-D4-11\n",
      "199-D4-12\n",
      "199-D4-13\n",
      "199-D4-14\n",
      "199-D4-15\n",
      "199-D3-2\n",
      "199-D4-17\n",
      "199-D4-18\n",
      "199-K-125A\n",
      "199-D5-36\n",
      "199-D5-37\n",
      "199-D4-19\n",
      "199-D5-38\n",
      "199-D5-39\n",
      "199-D5-40\n",
      "199-D4-20\n",
      "199-D5-41\n",
      "199-D5-42\n",
      "199-D5-43\n",
      "199-D5-44\n",
      "199-D4-21\n",
      "199-H4-65\n",
      "199-K-126\n",
      "199-D4-22\n",
      "199-D4-23\n",
      "199-D4-25\n",
      "199-D4-26\n",
      "199-D4-27\n",
      "199-D4-31\n",
      "199-D4-32\n",
      "199-D4-34\n",
      "199-D4-35\n",
      "199-D4-36\n",
      "199-D4-38\n",
      "199-D4-39\n",
      "199-D2-8\n",
      "199-D4-48\n",
      "199-D4-62\n",
      "199-D4-68\n",
      "199-D4-78\n",
      "199-D4-80\n",
      "199-D4-81\n",
      "199-D3-3\n",
      "199-D4-82\n",
      "199-D3-4\n",
      "199-D4-83\n",
      "199-D4-84\n",
      "199-D4-85\n",
      "199-D4-86\n",
      "199-D8-72\n",
      "199-K-129\n",
      "199-K-130\n",
      "699-96-44\n",
      "699-96-45\n",
      "199-D5-93\n",
      "199-N-136\n",
      "199-N-138\n",
      "199-N-139\n",
      "199-N-146\n",
      "199-K-137\n",
      "199-K-138\n",
      "199-K-139\n",
      "199-K-140\n",
      "199-N-147\n",
      "199-K-141\n",
      "199-K-142\n",
      "199-K-143\n",
      "199-K-144\n",
      "199-K-145\n",
      "199-K-146\n",
      "199-K-147\n",
      "199-K-148\n",
      "199-K-149\n",
      "199-K-150\n",
      "199-K-151\n",
      "199-K-152\n",
      "199-K-153\n",
      "199-K-154\n",
      "199-K-156\n",
      "199-K-157\n",
      "199-D5-97\n",
      "199-D5-98\n",
      "199-D5-99\n",
      "199-D2-11\n",
      "199-D5-103\n",
      "199-D5-104\n",
      "199-K-158\n",
      "199-D5-107\n",
      "199-D5-108\n",
      "199-D5-109\n",
      "199-D5-110\n",
      "199-D5-114\n",
      "199-D5-115\n",
      "699-100-43B\n",
      "699-99-42B\n",
      "699-99-41\n",
      "699-99-44\n",
      "699-98-43\n",
      "699-97-41\n",
      "699-98-46\n",
      "699-97-45\n",
      "699-95-45\n",
      "699-97-48B\n",
      "699-95-51\n",
      "699-97-43B\n",
      "699-94-41\n",
      "699-101-45\n",
      "699-95-48\n",
      "699-96-52B\n",
      "699-98-51\n",
      "199-B8-7\n",
      "199-B8-8\n",
      "199-D5-119\n",
      "199-D5-120\n",
      "199-D5-121\n",
      "199-D5-122\n",
      "199-K-159\n",
      "199-K-160\n",
      "199-K-161\n",
      "199-K-162\n",
      "199-K-163\n",
      "199-N-159\n",
      "199-K-164\n",
      "199-D5-123\n",
      "199-D5-125\n",
      "199-D5-126\n",
      "199-K-165\n",
      "199-K-166\n",
      "199-K-168\n",
      "199-N-165\n",
      "199-K-169\n",
      "199-K-170\n",
      "199-K-171\n",
      "199-K-172\n",
      "199-F8-7\n",
      "199-K-173\n",
      "199-N-167\n",
      "199-N-169\n",
      "199-N-170\n",
      "199-N-171\n",
      "199-N-172\n",
      "199-N-173\n",
      "199-K-174\n",
      "199-K-175\n",
      "199-H3-25\n",
      "199-K-178\n",
      "199-K-179\n",
      "199-K-180\n",
      "199-N-229\n",
      "199-N-211\n",
      "199-N-201\n",
      "199-N-248\n",
      "199-N-268\n",
      "199-N-280\n",
      "199-N-298\n",
      "199-N-316\n",
      "199-N-332\n",
      "199-N-342\n",
      "199-N-349\n",
      "199-N-348\n",
      "199-N-347\n",
      "199-N-346\n",
      "199-N-350\n",
      "199-N-351\n",
      "199-N-352\n",
      "199-N-353\n",
      "199-N-354\n",
      "199-N-355\n",
      "199-N-356\n",
      "199-N-357\n",
      "199-N-358\n",
      "199-N-360\n",
      "199-N-361\n",
      "199-N-362\n",
      "199-N-363\n",
      "199-N-364\n",
      "199-N-365\n",
      "199-N-366\n",
      "199-N-367\n",
      "199-K-181\n",
      "199-K-182\n",
      "199-H1-43\n",
      "199-B5-5\n",
      "199-B3-50\n",
      "199-B5-6\n",
      "199-B8-9\n",
      "199-D3-5\n",
      "199-D5-133\n",
      "199-D5-132\n",
      "199-D6-3\n",
      "199-D5-134\n",
      "199-D5-141\n",
      "199-H3-6\n",
      "199-H3-7\n",
      "199-H6-3\n",
      "199-H6-4\n",
      "199-H1-7\n",
      "199-H2-1\n",
      "199-H3-9\n",
      "199-H3-10\n",
      "399-1-54\n",
      "399-1-55\n",
      "399-1-56\n",
      "399-1-57\n",
      "399-1-58\n",
      "399-6-3\n",
      "399-1-59\n",
      "399-4-15\n",
      "399-3-33\n",
      "199-B2-14\n",
      "199-K-183\n",
      "199-K-184\n",
      "199-K-185\n",
      "199-K-186\n",
      "199-K-187\n",
      "199-K-188\n",
      "199-K-189\n",
      "199-K-190\n",
      "199-K-191\n",
      "199-K-192\n",
      "199-K-193\n",
      "199-K-194\n",
      "199-K-195\n",
      "199-K-196\n",
      "199-K-197\n",
      "199-K-198\n",
      "199-K-199\n",
      "199-B2-15\n",
      "199-B2-16\n",
      "199-B3-51\n",
      "199-B4-14\n",
      "199-F5-52\n",
      "199-F5-53\n",
      "199-F5-54\n",
      "199-K-200\n",
      "199-K-201\n",
      "199-B3-52\n",
      "199-D8-101\n",
      "199-D5-142\n",
      "199-H4-84\n",
      "199-H4-83\n",
      "199-D5-140\n",
      "199-F5-55\n",
      "199-F5-56\n",
      "199-N-182\n",
      "199-N-183\n",
      "199-N-184\n",
      "199-N-186\n",
      "199-N-187\n",
      "199-N-188\n",
      "199-N-189\n",
      "399-6-5\n",
      "199-K-202\n",
      "199-K-203\n",
      "199-K-204\n",
      "199-K-205\n",
      "199-K-206\n",
      "199-K-207\n",
      "199-K-208\n",
      "199-D5-143\n",
      "199-D5-144\n",
      "199-H4-85\n",
      "199-H4-86\n",
      "199-D5-145\n",
      "199-D5-146\n",
      "199-D5-147\n",
      "199-D5-148\n",
      "199-B4-16\n",
      "199-B4-18\n",
      "199-B5-9\n",
      "199-B5-10\n",
      "199-B5-11\n",
      "199-B5-12\n",
      "199-B5-13\n",
      "199-B5-14\n",
      "199-H4-90\n",
      "199-H4-91\n",
      "199-D5-153\n",
      "199-D5-154\n",
      "699-97-60\n",
      "699-97-61\n",
      "199-K-221\n",
      "199-K-222\n",
      "399-1-67\n",
      "199-H6-7\n",
      "199-H6-8\n",
      "199-D5-155\n",
      "199-D5-156\n",
      "199-D5-157\n",
      "199-D5-158\n"
     ]
    }
   ],
   "source": [
    "for iwell in selected_obs_wells[:]:\n",
    "    print(iwell)\n",
    "    obs_vec_df = np.sqrt(well_data_df[iwell, 'qlx']**2 + well_data_df[iwell, 'qly']**2)\n",
    "    well_data_df[iwell,'Total Tracer [M]', 'flux_ave_tracer'] = (well_data_df[iwell, 'Total Tracer [M]'] * obs_vec_df).sum(axis = 1) / obs_vec_df.sum(axis = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-29T16:18:03.827179Z",
     "start_time": "2018-05-29T16:18:00.023160Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "with open(fname_well_data_df_pickle, \"wb\") as f:\n",
    "    pickle.dump((well_data_df), f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## plot well simu vs obs head"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-08T23:40:28.686786Z",
     "start_time": "2018-06-08T23:40:21.035472Z"
    }
   },
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "Can't get attribute 'FrozenNDArray' on <module 'pandas.core.indexes.frozen' from '/global/homes/p/pshuai/.conda/envs/widget/lib/python3.6/site-packages/pandas/core/indexes/frozen.py'>",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-14-8437a7578830>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;31m# loading objects\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;32mwith\u001b[0m \u001b[0mopen\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfname_well_data_df_pickle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"rb\"\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mf\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m     \u001b[0mwell_data_df\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpickle\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mload\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m: Can't get attribute 'FrozenNDArray' on <module 'pandas.core.indexes.frozen' from '/global/homes/p/pshuai/.conda/envs/widget/lib/python3.6/site-packages/pandas/core/indexes/frozen.py'>"
     ]
    }
   ],
   "source": [
    "# loading objects    \n",
    "with open(fname_well_data_df_pickle, \"rb\") as f:\n",
    "    well_data_df = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(fname_selected_wells_pk, \"rb\") as f:\n",
    "    selected_obs_wells = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-08T23:41:05.618947Z",
     "start_time": "2018-06-08T23:41:05.449833Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "real_time_6h = []\n",
    "for itime in np.arange(len(well_data_df.index)):\n",
    "    itime_str = list(batch_delta_to_time(date_origin, [float(well_data_df.index[itime])], \"%Y-%m-%d %H:%M:%S\", \"hours\"))\n",
    "    itime_str = ''.join(itime_str)\n",
    "    real_time_6h.append(datetime.datetime.strptime(itime_str, \"%Y-%m-%d %H:%M:%S\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-07T00:34:52.345433Z",
     "start_time": "2018-06-07T00:06:57.436784Z"
    },
    "code_folding": [],
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "399-1-1\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/global/common/cori/software/python/3.6-anaconda-5.2/lib/python3.6/site-packages/IPython/core/interactiveshell.py:2817: PerformanceWarning: indexing past lexsort depth may impact performance.\n",
      "  raw_cell, store_history, silent, shell_futures)\n",
      "/global/common/cori/software/python/3.6-anaconda-5.2/lib/python3.6/site-packages/numpy/core/fromnumeric.py:57: FutureWarning: 'argmin' is deprecated, use 'idxmin' instead. The behavior of 'argmin'\n",
      "will be corrected to return the positional minimum in the future.\n",
      "Use 'series.values.argmin' to get the position of the minimum now.\n",
      "  return getattr(obj, method)(*args, **kwds)\n",
      "/global/common/cori/software/python/3.6-anaconda-5.2/lib/python3.6/site-packages/pandas/core/indexing.py:1027: FutureWarning: \n",
      "Passing list-likes to .loc or [] with any missing label will raise\n",
      "KeyError in the future, you can use .reindex() as an alternative.\n",
      "\n",
      "See the documentation here:\n",
      "https://pandas.pydata.org/pandas-docs/stable/indexing.html#deprecate-loc-reindex-listlike\n",
      "  return getattr(section, self.name)[new_key]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "399-1-2\n",
      "399-1-7\n",
      "399-1-10A\n",
      "399-1-11\n",
      "399-1-12\n",
      "399-1-13A\n",
      "399-1-16A\n",
      "399-1-17A\n",
      "399-1-21A\n",
      "399-1-23\n",
      "399-1-32\n",
      "399-1-60\n",
      "399-2-1\n",
      "399-2-2\n",
      "399-2-3\n",
      "399-2-5\n",
      "399-2-7\n",
      "No well obs data available for399-2-7\n",
      "399-2-23\n",
      "No well obs data available for399-2-23\n",
      "399-2-25\n",
      "399-2-32\n",
      "399-2-33\n",
      "399-2-34\n",
      "No well obs data available for399-2-34\n",
      "399-2-37\n",
      "No well obs data available for399-2-37\n",
      "399-3-9\n",
      "399-3-12\n",
      "399-3-18\n",
      "399-3-19\n",
      "399-3-20\n",
      "399-3-21\n",
      "399-3-22\n",
      "399-3-34\n",
      "399-3-37\n",
      "399-4-1\n",
      "399-4-7\n",
      "399-4-9\n",
      "399-4-14\n",
      "399-5-1\n",
      "199-D5-13\n",
      "199-D5-14\n",
      "199-D8-4\n",
      "199-D8-5\n",
      "199-D8-55\n",
      "199-D8-6\n",
      "199-F7-2\n",
      "199-H3-2A\n",
      "199-H3-2C\n",
      "199-H4-10\n",
      "199-H4-11\n",
      "199-H4-12A\n",
      "199-H4-12C\n",
      "199-H4-13\n",
      "199-H4-14\n",
      "199-H4-15A\n",
      "199-H4-16\n",
      "199-H4-17\n",
      "199-H4-18\n",
      "199-H4-6\n",
      "199-H4-7\n",
      "199-H4-8\n",
      "199-H4-9\n",
      "199-K-19\n",
      "199-K-22\n",
      "199-K-30\n",
      "199-K-31\n",
      "199-K-36\n",
      "199-N-14\n",
      "199-N-16\n",
      "199-N-18\n",
      "199-N-25\n",
      "199-N-26\n",
      "199-N-27\n",
      "199-N-28\n",
      "199-N-29\n",
      "199-N-31\n",
      "199-N-32\n",
      "199-N-33\n",
      "199-N-34\n",
      "199-N-36\n",
      "199-N-37\n",
      "199-N-39\n",
      "199-N-40\n",
      "199-N-41\n",
      "199-N-42\n",
      "199-N-49\n",
      "199-N-50\n",
      "199-N-51\n",
      "199-N-52\n",
      "199-N-53\n",
      "199-N-54\n",
      "199-N-55\n",
      "199-N-56\n",
      "199-N-57\n",
      "199-N-58\n",
      "199-N-59\n",
      "199-N-60\n",
      "199-N-61\n",
      "199-N-62\n",
      "199-N-63\n",
      "199-N-64\n",
      "199-N-65\n",
      "199-N-66\n",
      "199-N-67\n",
      "199-N-69\n",
      "199-N-70\n",
      "199-N-71\n",
      "199-N-72\n",
      "199-N-73\n",
      "199-N-74\n",
      "199-N-75\n",
      "199-N-76\n",
      "399-1-15\n",
      "399-1-16B\n",
      "399-1-17B\n",
      "399-1-17C\n",
      "399-1-18A\n",
      "399-1-18B\n",
      "399-1-18C\n",
      "399-1-6\n",
      "399-1-8\n",
      "399-1-9\n",
      "399-3-10\n",
      "399-4-10\n",
      "399-4-11\n",
      "399-1-14A\n",
      "399-1-21B\n",
      "399-8-5A\n",
      "399-8-5B\n",
      "399-8-5C\n",
      "199-N-77\n",
      "199-N-81\n",
      "199-N-22\n",
      "199-N-30\n",
      "199-N-43\n",
      "199-N-47\n",
      "199-N-48\n",
      "399-1-10B\n",
      "399-1-13B\n",
      "399-1-14B\n",
      "399-3-11\n",
      "399-5-4B\n",
      "399-6-2\n",
      "699-37-E4\n",
      "199-K-109A\n",
      "199-K-110A\n",
      "199-K-111A\n",
      "199-K-106A\n",
      "199-K-107A\n",
      "199-K-108A\n",
      "199-N-103A\n",
      "199-N-104A\n",
      "199-N-105A\n",
      "199-N-106A\n",
      "199-D8-68\n",
      "199-D8-69\n",
      "199-D8-70\n",
      "199-D8-71\n",
      "199-H4-63\n",
      "199-H4-64\n",
      "199-H3-3\n",
      "199-H3-4\n",
      "199-H3-5\n",
      "199-K-112A\n",
      "199-K-113A\n",
      "199-K-114A\n",
      "199-K-115A\n",
      "199-K-116A\n",
      "199-K-117A\n",
      "199-K-118A\n",
      "199-K-119A\n",
      "199-K-120A\n",
      "199-K-121A\n",
      "199-K-122A\n",
      "199-K-123A\n",
      "199-K-124A\n",
      "199-D4-1\n",
      "199-D4-4\n",
      "199-D4-5\n",
      "199-D4-6\n",
      "199-D4-7\n",
      "199-D4-8\n",
      "199-D4-9\n",
      "199-D4-10\n",
      "199-D4-11\n",
      "199-D4-12\n",
      "199-D4-13\n",
      "199-D4-14\n",
      "199-D4-15\n",
      "199-D3-2\n",
      "199-D4-17\n",
      "199-D4-18\n",
      "199-K-125A\n",
      "199-D5-36\n",
      "199-D5-37\n",
      "199-D4-19\n",
      "199-D5-38\n",
      "199-D5-39\n",
      "199-D5-40\n",
      "199-D4-20\n",
      "199-D5-41\n",
      "199-D5-42\n",
      "199-D5-43\n",
      "199-D5-44\n",
      "199-D4-21\n",
      "199-H4-65\n",
      "199-K-126\n",
      "199-D4-22\n",
      "199-D4-23\n",
      "199-D4-25\n",
      "199-D4-26\n",
      "199-D4-27\n",
      "199-D4-31\n",
      "199-D4-32\n",
      "199-D4-34\n",
      "199-D4-35\n",
      "199-D4-36\n",
      "199-D4-38\n",
      "199-D4-39\n",
      "199-D2-8\n",
      "199-D4-48\n",
      "199-D4-62\n",
      "199-D4-68\n",
      "199-D4-78\n",
      "199-D4-80\n",
      "199-D4-81\n",
      "199-D3-3\n",
      "199-D4-82\n",
      "199-D3-4\n",
      "199-D4-83\n",
      "199-D4-84\n",
      "199-D4-85\n",
      "199-D4-86\n",
      "199-D8-72\n",
      "199-K-129\n",
      "199-K-130\n",
      "699-96-44\n",
      "699-96-45\n",
      "199-D5-93\n",
      "199-N-136\n",
      "199-N-138\n",
      "199-N-139\n",
      "199-N-146\n",
      "199-K-137\n",
      "199-K-138\n",
      "199-K-139\n",
      "199-K-140\n",
      "199-N-147\n",
      "199-K-141\n",
      "199-K-142\n",
      "199-K-143\n",
      "199-K-144\n",
      "199-K-145\n",
      "199-K-146\n",
      "199-K-147\n",
      "199-K-148\n",
      "199-K-149\n",
      "199-K-150\n",
      "199-K-151\n",
      "199-K-152\n",
      "199-K-153\n",
      "199-K-154\n",
      "199-K-156\n",
      "199-K-157\n",
      "199-D5-97\n",
      "199-D5-98\n",
      "199-D5-99\n",
      "199-D2-11\n",
      "199-D5-103\n",
      "199-D5-104\n",
      "199-K-158\n",
      "199-D5-107\n",
      "199-D5-108\n",
      "199-D5-109\n",
      "199-D5-110\n",
      "199-D5-114\n",
      "199-D5-115\n",
      "699-100-43B\n",
      "699-99-42B\n",
      "699-99-41\n",
      "699-99-44\n",
      "699-98-43\n",
      "699-97-41\n",
      "699-98-46\n",
      "699-97-45\n",
      "699-95-45\n",
      "699-97-48B\n",
      "699-95-51\n",
      "699-97-43B\n",
      "699-94-41\n",
      "699-101-45\n",
      "699-95-48\n",
      "699-96-52B\n",
      "699-98-51\n",
      "199-B8-7\n",
      "199-B8-8\n",
      "199-D5-119\n",
      "199-D5-120\n",
      "199-D5-121\n",
      "199-D5-122\n",
      "199-K-159\n",
      "199-K-160\n",
      "199-K-161\n",
      "199-K-162\n",
      "199-K-163\n",
      "199-N-159\n",
      "199-K-164\n",
      "199-D5-123\n",
      "199-D5-125\n",
      "199-D5-126\n",
      "199-K-165\n",
      "199-K-166\n",
      "199-K-168\n",
      "199-N-165\n",
      "199-K-169\n",
      "199-K-170\n",
      "199-K-171\n",
      "199-K-172\n",
      "199-F8-7\n",
      "199-K-173\n",
      "199-N-167\n",
      "199-N-169\n",
      "199-N-170\n",
      "199-N-171\n",
      "199-N-172\n",
      "199-N-173\n",
      "199-K-174\n",
      "199-K-175\n",
      "199-H3-25\n",
      "199-K-178\n",
      "199-K-179\n",
      "199-K-180\n",
      "199-N-229\n",
      "199-N-211\n",
      "199-N-201\n",
      "199-N-248\n",
      "199-N-268\n",
      "199-N-280\n",
      "199-N-298\n",
      "199-N-316\n",
      "199-N-332\n",
      "199-N-342\n",
      "199-N-349\n",
      "199-N-348\n",
      "199-N-347\n",
      "199-N-346\n",
      "199-N-350\n",
      "199-N-351\n",
      "199-N-352\n",
      "199-N-353\n",
      "199-N-354\n",
      "199-N-355\n",
      "199-N-356\n",
      "199-N-357\n",
      "199-N-358\n",
      "199-N-360\n",
      "199-N-361\n",
      "199-N-362\n",
      "199-N-363\n",
      "199-N-364\n",
      "199-N-365\n",
      "199-N-366\n",
      "199-N-367\n",
      "199-K-181\n",
      "199-K-182\n",
      "199-H1-43\n",
      "199-B5-5\n",
      "199-B3-50\n",
      "199-B5-6\n",
      "199-B8-9\n",
      "199-D3-5\n",
      "199-D5-133\n",
      "199-D5-132\n",
      "199-D6-3\n",
      "199-D5-134\n",
      "199-D5-141\n",
      "199-H3-6\n",
      "199-H3-7\n",
      "199-H6-3\n",
      "199-H6-4\n",
      "199-H1-7\n",
      "199-H2-1\n",
      "199-H3-9\n",
      "199-H3-10\n",
      "399-1-54\n",
      "399-1-55\n",
      "399-1-56\n",
      "399-1-57\n",
      "399-1-58\n",
      "399-6-3\n",
      "399-1-59\n",
      "399-4-15\n",
      "399-3-33\n",
      "199-B2-14\n",
      "199-K-183\n",
      "199-K-184\n",
      "199-K-185\n",
      "199-K-186\n",
      "199-K-187\n",
      "199-K-188\n",
      "199-K-189\n",
      "199-K-190\n",
      "199-K-191\n",
      "199-K-192\n",
      "199-K-193\n",
      "199-K-194\n",
      "199-K-195\n",
      "199-K-196\n",
      "199-K-197\n",
      "199-K-198\n",
      "199-K-199\n",
      "199-B2-15\n",
      "199-B2-16\n",
      "199-B3-51\n",
      "199-B4-14\n",
      "199-F5-52\n",
      "199-F5-53\n",
      "199-F5-54\n",
      "199-K-200\n",
      "199-K-201\n",
      "199-B3-52\n",
      "199-D8-101\n",
      "199-D5-142\n",
      "199-H4-84\n",
      "199-H4-83\n",
      "199-D5-140\n",
      "199-F5-55\n",
      "199-F5-56\n",
      "199-N-182\n",
      "199-N-183\n",
      "199-N-184\n",
      "199-N-186\n",
      "199-N-187\n",
      "199-N-188\n",
      "199-N-189\n",
      "399-6-5\n",
      "199-K-202\n",
      "199-K-203\n",
      "199-K-204\n",
      "199-K-205\n",
      "199-K-206\n",
      "199-K-207\n",
      "199-K-208\n",
      "199-D5-143\n",
      "199-D5-144\n",
      "199-H4-85\n",
      "199-H4-86\n",
      "199-D5-145\n",
      "199-D5-146\n",
      "199-D5-147\n",
      "199-D5-148\n",
      "199-B4-16\n",
      "199-B4-18\n",
      "199-B5-9\n",
      "199-B5-10\n",
      "199-B5-11\n",
      "199-B5-12\n",
      "199-B5-13\n",
      "199-B5-14\n",
      "199-H4-90\n",
      "199-H4-91\n",
      "199-D5-153\n",
      "199-D5-154\n",
      "699-97-60\n",
      "699-97-61\n",
      "199-K-221\n",
      "199-K-222\n",
      "399-1-67\n",
      "199-H6-7\n",
      "199-H6-8\n",
      "199-D5-155\n",
      "199-D5-156\n",
      "199-D5-157\n",
      "199-D5-158\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline \n",
    "\n",
    "RMSE_df = screen_300A_HEIS.copy()\n",
    "RMSE_df['RMSE'] = np.nan\n",
    "\n",
    "for iwell in selected_obs_wells[:]:\n",
    "#     iwell = '399-1-15'\n",
    "    print(iwell)\n",
    "#     iwell_screen_index = np.asarray(np.where(well_screen.index == iwell)).flatten()\n",
    "    iwell_interval_index = np.where((z >= well_screen.loc[iwell, 'SCREEN_ELEV_BOTTOM']) \\\n",
    "                                    & (z <= well_screen.loc[iwell, 'SCREEN_ELEV_TOP']))\n",
    "    iwell_interval_index = np.asarray(iwell_interval_index).flatten()\n",
    "\n",
    "    z_reshape = z[iwell_interval_index].reshape(1, len(iwell_interval_index))\n",
    "\n",
    "    isimu_head = well_data_df[iwell, 'Liquid Pressure'].copy()\n",
    "    isimu_head = (isimu_head - P_atm) / rho / g_const\n",
    "    # apply broadcasting rule\n",
    "    isimu_head_df = pd.DataFrame(isimu_head.values + z_reshape, columns=isimu_head.columns, index=isimu_head.index)\n",
    "    isimu_head_df = isimu_head_df.sort_index()\n",
    "    \n",
    "    selected_well_simu_head = isimu_head_df.iloc[:, 0].to_frame(name='simu_head') # choose the bottom screen interval as simu head\n",
    "    selected_well_simu_head['DateTime'] = real_time_6h\n",
    "    selected_well_simu_head.set_index('DateTime', inplace=True)\n",
    "    # read obs.head\n",
    "    if iwell in SFA_well['WellName'].unique():\n",
    "        iobs_head_df = SFA_well[ SFA_well.loc[:, 'WellName'] == iwell].copy()\n",
    "    elif iwell in HEIS_auto['WellName'].unique():\n",
    "        iobs_head_df = HEIS_auto[ HEIS_auto.loc[:, 'WellName'] == iwell].copy()\n",
    "    elif iwell in HEIS_manual['WellName'].unique():\n",
    "        iobs_head_df = HEIS_manual[ HEIS_manual.loc[:, 'WellName'] == iwell].copy()\n",
    "    else:\n",
    "        iobs_head_df = pd.DataFrame()\n",
    "        print('No well obs data available for' + iwell)\n",
    "        continue\n",
    "   \n",
    "    #     iobs_head_df = iobs_head_df.sort_index(axis=1)\n",
    "    iobs_head_df['DateTime'] = iobs_head_df['DateTime'].apply(pd.to_datetime)\n",
    "    iobs_head_df = iobs_head_df.drop_duplicates('DateTime', keep='first') \\\n",
    "                                .sort_values(by = 'DateTime').set_index('DateTime', drop=False)\n",
    "    # filter wells outside of dates\n",
    "    iobs_head_df = iobs_head_df[(iobs_head_df['DateTime'] >= '1/1/2011') \\\n",
    "                                & (iobs_head_df['DateTime'] <= '1/1/2016')]\n",
    "    \n",
    "        \n",
    "    \n",
    "    if iobs_head_df['WL'].count() > 1000 :           \n",
    "        # find closest mass1 location\n",
    "        iwell_easting = well_screen.loc[iwell, 'Easting']\n",
    "        iwell_northing = well_screen.loc[iwell, 'Northing']\n",
    "        dist = np.sqrt((mass1_coord['Easting']- iwell_easting)**2 + (mass1_coord['Northing']- iwell_northing)**2)\n",
    "        mass_id = mass1_coord.loc[np.argmin(dist), 'Mass_ID']\n",
    "\n",
    "        datum_files = glob.glob(fname_river_bc_6h + \"Datum*.txt\")\n",
    "#         datum_files = natsorted(datum_files) \n",
    "        datum_files = natural_sort(datum_files)\n",
    "    \n",
    "        for ifile in datum_files:\n",
    "            if '_' + str(mass_id) in ifile:\n",
    "                river_stage = pd.read_table(ifile, sep=' ', header=None, names=['time', 'x', 'y', 'wl'])\n",
    "            elif mass_id == 333:\n",
    "                river_stage = pd.read_table(datum_files[-1], sep=' ', header=None, names=['time', 'x', 'y', 'wl'])\n",
    "\n",
    "    #     idatum = [ifile if '_' + str(mass_id) in ifile else datum_files[-1] for ifile in datum_files]\n",
    "    #     river_stage = pd.read_table(idatum, sep=' ', header=None, names=['time', 'x', 'y', 'wl'])\n",
    "\n",
    "        river_stage.time = batch_delta_to_time(date_origin, river_stage.time.astype(np.float64), \"%Y-%m-%d %H:%M:%S\", \"seconds\")\n",
    "        river_stage['time'] = river_stage['time'].apply(pd.to_datetime)    \n",
    "        ##================== apply moving-average ========================##\n",
    "        # find the mode of frequecy (in sec)\n",
    "        res = iobs_head_df['DateTime'].diff().dt.seconds.value_counts().index[0]\n",
    "        # apply moving average to well with sampling freq.< 3600 sec\n",
    "        win_size = int(6*3600/res) # 6-hour moving average       \n",
    "        if res <= 3600 :\n",
    "            datetime_ind = pd.date_range(start='1/1/2011', end='1/1/2016', freq=str(res) + 'S')\n",
    "            df = pd.DataFrame(np.nan, index=datetime_ind, columns=['WL_nan'])\n",
    "            iobs_head_df = pd.concat([df, iobs_head_df], axis=1, join='outer', join_axes=[df.index])\n",
    "\n",
    "            iobs_head_df['mv_ave_6h'] = iobs_head_df['WL'].rolling(win_size).mean()\n",
    "            style = '-k'\n",
    "#         elif res == 3600 :\n",
    "#             datetime_ind = pd.date_range(start='1/1/2011', end='1/1/2016', freq=str(res) + 'S')\n",
    "#             df = pd.DataFrame(np.nan, index=datetime_ind, columns=['WL_nan'])\n",
    "#             iobs_head_df = pd.concat([df, iobs_head_df], axis=1, join='outer', join_axes=[df.index])        \n",
    "#             win_size = int(6*3600/res)\n",
    "#             iobs_head_df['mv_ave_6h'] = iobs_head_df['WL'].rolling(win_size).mean()\n",
    "#             style = '-k'\n",
    "            ##================= get RMSE====================##\n",
    "            iobs_head_df_6h = iobs_head_df.loc[real_time_6h, 'mv_ave_6h']\n",
    "\n",
    "            df = pd.concat([selected_well_simu_head, iobs_head_df_6h], axis=1).\\\n",
    "                    rename(columns = {'mv_ave_6h': 'obs_head'}).dropna(how = 'any')\n",
    "\n",
    "            iRMSE = math.sqrt(mean_squared_error(df.simu_head, df.obs_head))\n",
    "            \n",
    "            RMSE_df.loc[iwell, 'RMSE'] = iRMSE\n",
    "                \n",
    "        else :\n",
    "            iobs_head_df['mv_ave_6h'] = iobs_head_df['WL']\n",
    "            style = '-ko'\n",
    "            iRMSE = np.nan\n",
    "        \n",
    "        \n",
    "        ##================= plot simu vs obs head========================##\n",
    "\n",
    "        fig, ax = plt.subplots(1, 1) # short hand\n",
    "\n",
    "        river_stage.plot(x= 'time', y= 'wl', legend=False, ax=ax, color = 'gray' , lw = 1, rot = 0, alpha = 0.5)\n",
    "#         ax.plot(real_time_6h, selected_well_simu_head, 'r-', lw= 1)\n",
    "        selected_well_simu_head.plot(y='simu_head', ax=ax, style='-r', lw=1)\n",
    "        iobs_head_df.plot(x= 'DateTime', y= 'mv_ave_6h', legend=False, ax=ax, style=style ,\\\n",
    "                          markerfacecolor = \"None\", ms = 3, lw = 1, rot = 0)\n",
    "\n",
    "        ax.set_title('Well ' + iwell,  fontweight = 'bold', fontsize = 14)\n",
    "        # format the ticks\n",
    "        ax.xaxis.set_major_locator(mdates.YearLocator())\n",
    "        ax.xaxis.set_major_formatter(mdates.DateFormatter('%Y'))\n",
    "        ax.xaxis.set_tick_params('labelbottom')\n",
    "        plt.setp(ax.xaxis.get_majorticklabels(), rotation=0, ha = 'center' )\n",
    "        \n",
    "        ax.set_xlim('01/01/2011', '01/01/2016')\n",
    "#         ax.set_ylim([104.5, 108.5])\n",
    "        ax.set_ylabel('Water Level (m)')\n",
    "        ax.set_xlabel('')\n",
    "        ax.legend(['river stage', 'simu.head', 'obs.head'], frameon = False)\n",
    "        \n",
    "        if iRMSE != np.nan:\n",
    "            ax.text(0.5,0.9,'RMSE = {:.2}'.format(iRMSE), transform = ax.transAxes)\n",
    "#         fig.autofmt_xdate()\n",
    "\n",
    "        fig.tight_layout()\n",
    "        fig.set_size_inches(8, 5)\n",
    "        \n",
    "        fname = fig_simu_obs_wl + 'Well_' + iwell + '.jpg' \n",
    "        fig.savefig(fname , dpi=300)\n",
    "        plt.close(fig)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-07T04:35:21.697073Z",
     "start_time": "2018-06-07T04:35:21.682209Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "fname = fig_simu_obs_wl + 'RMSE.csv'\n",
    "RMSE_df.to_csv(fname)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## plot well RMSE (obs.head, simu.head)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-07T04:35:08.295339Z",
     "start_time": "2018-06-07T04:35:08.282427Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# convert easting/northing to lat/long\n",
    "NAD83 = pyproj.Proj(init= 'epsg:2856')\n",
    "WGS84 = pyproj.Proj(init= 'epsg:4326')\n",
    "lon, lat= pyproj.transform(NAD83, WGS84, RMSE_df['Easting'].values, RMSE_df['Northing'].values)\n",
    "RMSE_df['lat'] = lat\n",
    "RMSE_df['lon'] = lon"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### scatter plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "ini_head_wells = pd.read_csv(fname_selected_wells_csv)\n",
    "\n",
    "river_geometry = pd.read_csv(fname_river_geo)\n",
    "\n",
    "RMSE_df = pd.read_csv(fig_simu_obs_wl + 'RMSE.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-08T04:22:47.892008Z",
     "start_time": "2018-06-08T04:22:47.867089Z"
    }
   },
   "outputs": [],
   "source": [
    "RMSE_df_sub = RMSE_df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-07T05:47:33.473224Z",
     "start_time": "2018-06-07T05:47:32.846139Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAhsAAAH8CAYAAAB4jm0aAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAIABJREFUeJzs3XdcV2X7wPHPYaNsFHCgoKSIqLhwrzT3XqA4MDMt1DIt\n9bHnV5oztza03CNxYKamuCq3OckFTkQRFFSQPb/n9wf4fSRAUfmK43q/XryA+z7nPtfBzIt7Kqqq\nIoQQQgihK3pFHYAQQggh3mySbAghhBBCpyTZEEIIIYROSbIhhBBCCJ2SZEMIIYQQOiXJhhBCCCF0\nSpINIYQQQuiUJBtCCCGE0ClJNoQQQgihUwZFHcCrqkSJEqqTk1NRhyGEEEK8sk6dOnVPVdWST7tO\nko18ODk5cfLkyaIOQwghhHhlKYoSVpDrZBhFCCGEEDolyYYQQgghdEqSDSGEEELolMzZeAbp6emE\nh4eTkpJS1KGIF2BiYkLZsmUxNDQs6lCEEOKtIMnGMwgPD8fc3BwnJycURSnqcMRzUFWV+/fvEx4e\njrOzc1GHI4QQbwUZRnkGKSkp2NraSqLxGlMUBVtbW+mdEkKIl0iSjWckicbrT/4MhRDi5ZJkQ0cS\n0hL4Zv83OM1zQn+SPk7znPhm/zckpCUUdWhCCCHESyXJhg4kpCXQclVLzkefZ4v3FlK/TGWL9xbO\nRZ2j5aqWL5RwBAYGUrlyZVxcXJg+fXqB71uxYgUlS5bEw8MDV1dX5s6dq637+uuvURSFq1evasvm\nzZuHoijajc2WLVtGtWrVqF69Ou7u7vz2228A+Pr64uzsjIeHBx4eHjRs2PC53+3fQkNDqVevHi4u\nLnh5eZGWlpbndfr6+trnd+7cudCeL4QQonBIsqEDc4/OxcnKCf8e/ng4eGCgZ4CHgwfre66nvGV5\n5h6d+/RG8pCZmYmfnx87d+7k4sWLrFu3josXLz7xngcPHmi/9vLyIigoiMOHDzNlyhRu3bqlratW\nrRr+/v7a7zdu3EjVqlWBrImxU6ZM4dChQ5w9e5Zjx45RvXp17bUzZ84kKCiIoKAgjhw5UuB4nmbs\n2LGMGjWKq1evYm1tzdKlS/O8ztTUVPv8rVu3Frh9IYQQL4ckGzqw9MxSxjcen2tugKIojG88nmVB\ny56r3ePHj+Pi4kKFChUwMjLC29tb28PwuLi4OBYvXoynpyezZs3KVW9ra4uLiwuRkZHasq5du2rb\nunbtGpaWlpQoUQKAqKgozM3NMTMzA8DMzOyZVnJkZGSwdetWOnfuTLdu3Qp0j6qq/PHHH/Ts2ROA\ngQMHsmXLlgI/UwghxKtDkg0duBV3C3c79zzr3O3cufXwVp51T3P79m0cHR2135ctW5bbt29rvz90\n6BC+vr7Url2b0NBQ1qxZw9SpU3O1c/PmTVJSUnL0TlhYWODo6Mj58+fx9/fHy8tLW1ejRg3s7e1x\ndnZm0KBBbNu2LUd7n3/+uXYYw8fHR1t+9epVxo8fT5UqVQgICGD06NHs378fgPj4eO09//64ePEi\n9+/fx8rKCgMDgzzf9XEpKSnUqlWL+vXrS0IihBCvINlnQwccLRw5H3UeDwePXHXno87jaOmYx10v\nZuTIkaxevZrvv/+epUuXoq+vn+ua9evXc+DAAUJCQvjuu+8wMTHJUe/t7Y2/vz+7du1i3759LF++\nHMiaExEYGMiJEyfYt28fo0aN4tSpU3z99ddA1jDKox6IRwICAvDy8mLChAmcPn0ac3PzHPXm5uYE\nBQXl+z737t0r8LuHhYVRpkwZrl+/zrvvvku1atWoWLFige8XQgihW9KzoQODaw5m6sGpqKqao1xV\nVaYdmsb7Hu8/V7tlypTJMc8iPDycMmXKAPDZZ5/h5+fHxIkTGTRoEH/++Weu53t5eXH27FmOHDnC\nuHHjuHPnTo76jh07snr1asqVK4eFhUWOOkVR8PT0ZPz48fj7+xMQEPDEWN977z3mz5/P77//To8e\nPVi3bl2OvS2e1rNha2tLbGwsGRkZud41r58LQIUKFWjevDlnzpx5YmxCCCFeLkk2dGBUg1GEPQzD\na5MXZyLPkJ6ZzpnIM3ht8iLsYRijGox6rnbr1q3LlStXCA0NJS0tDX9/f+3qCycnJyZPnszFixfx\n9vZm4cKFuLq6snbt2lzt1KlTh/79+zN//vwc5cWKFWPGjBlMmDAhR3lERASnT5/Wfh8UFET58uWf\nGKuFhQV+fn6cPHmSGTNmcOjQIapUqcIXX3wB/K9nI68PNzc3FEWhRYsWbNq0CYCVK1fSpUuXXM+J\niYkhNTUVyOoNOXz4MG5ubk/7UQohhHiJZBhFB8yMzNg3YB9zj86l+4bu3Hp4C0dLR973eJ9lXZZh\nZmT2XO0aGBjw3Xff0aZNGzIzM3n//fe1K0Ye0dfXp3379rRv356oqCguX76cZ1tjx46lVq1a/Oc/\n/8lR7u3tneva9PR0xowZQ0REBCYmJpQsWZJFixZp6z///HMmT56s/f748eMYGRlpv69Zsybff/89\nKSkp/PHHHwV+3xkzZuDt7c2XX35JzZo1GTx4MAAnT55k0aJFLFmyhODgYIYOHYqenh4ajYZx48ZJ\nsiGEEK8Y5d9d7SJLnTp11Ed7TDwSHBxMlSpViigiUZjkz1IIIV6coiinVFWt87TrZBhFCCGEEDol\nyYYQQgghdErmbAghhBDimZw4cYLDhw8X+Hrp2RBCCCHEM9m9ezc9evQo8PWSbAghhBDimZQuXZrd\nu3cX+HoZRtGV9AQImQvXl0LSLSjmCBUGg+soMHy+pa9CCCHEq8DHx4cbN24U+Hrp2dCF9AT4oyU8\nPA9Nt4BXatbnh+eyytN1e8T8X3/9haWlpXZHzkmTJhW4/ebNm1O5cmVq1KhB3bp1c2wp7uTkRJMm\nTXJc7+Hhgbt71jkwSUlJ+Pj4UK1aNdzd3WncuDEJCVnv+vgx8B4eHvnG/jzatm2LlZUVHTt2LLQ2\nhRBC5M/IyIhKlSoV+Hrp2dCFkLlQ3Aka+cOjk1+tPaDRejjslVVf7b/P3OyjI+b37NlD2bJlqVu3\nLp07d85zE6smTZqwffv2ArWblpZGeno6xYsXB2Dt2rXUqVOH5cuX8/nnn7Nnzx7ttfHx8dy6dQtH\nR0eCg4NztDN//nzs7e05d+4cAJcuXcLQ0BD43zHwBRUTE4O1tXWBrv38889JSkpi8eLFBW5fCCHE\n/2g0Gg4cOEBsbCzNmjUr8P9/C0p6NnTh+lKoOv5/icYjigJu4+G6bo+YL6jg4GBGjx5N5cqV89xp\ntEGDBrlOWu3duzfr168HYN26dfTp00dbFxkZmeP8ksqVK2NsbFzgeKKiopg1axbu7u7aZxREy5Yt\ncx30JoQQouAuXLjA5cuXMTExYceOHYXeviQbupB0CyzzPmIeK/es+ufwtCPmH3fkyBGqV69Ou3bt\nuHDhgrY8MTGR5cuX07hxY4YMGYKbmxtnz56lZs2audoIDAyka9euOcp69OjB5s2bAdi2bRudOnXS\n1r3//vvMmDGDBg0a8OWXX3LlyhVtXXJyco5hlEfJhEajITAwkJ49e9K8eXNSUlIIDAxk2LBhQFYv\nS16Htf37lFkhhBDPz9DQkNTUVBISEnIcN1FYZBhFF4o5Zs3XsM59xDyx57PqdahWrVrcvHkTMzMz\nduzYQdeuXbX/8JcqVYrq1auzZMkSXF1d87zfx8eHtLQ0EhIScg192NraYm1tjb+/P1WqVKFYsWLa\nOg8PD65fv87u3bvZu3cvdevW5ejRo1SpUiXfYZSuXbty+vRplixZQps2bVD+1Rvk4+ODj4/Pi/5I\nhBBCPEHlypVJSUkhNjaWevXqFXr70rOhCxUGw4Wp8O9zZ1QVLk6DCoV/xPzjLCwsMDPLWvHSvn17\n0tPTuXfvHgCbNm2iTJkydO/enUmTJhEWFpbr/rVr13L9+nUGDhzIiBEjctV7eXnh5+eXYwjlETMz\nM7p3784PP/xAv379ntodN23aNHr27MmIESPw8/PjxIkTuWKRng0hhNAtRVHw8PCgefPmmJqaFnr7\nkmzogusoSAzLmgz64Axo0rM+H/bKKnct/CPmH3fnzh0eHbB3/PhxNBoNtra2ALRu3Zr169dz8OBB\nLC0t6dKlC61atcq1hElRFL755huOHTtGSEhIjrpu3brxxRdf0KZNmxzlhw8fJiYmBsiadHrx4sWn\nHkVftWpV5s2bx4ULF2jWrBkTJkygevXq2vXbPj4+eR5D/+joeSGEEK8+GUbRBUMzeHdf1qqTg90f\n22fjfai37Ln32XjSEfOPjnwfNmwYmzZt4scff8TAwABTU1P8/f1zDU/Y2tryySef8Mknn3D8+HH0\n9fVzPc/U1JTRo0czc+ZMli5dqi03Nzdn7Nixua6/du0aH330EaqqotFo6NChg3aHuUdzNh5p27Zt\njuWvRkZGeHl54eXlRVhYmLYnpiCaNGlCSEgICQkJlC1blqVLl+ZKhIQQQhQdOWI+H3LE/JtN/iyF\nEOLFyRHzQgghhHglSLIhhBBCCJ2SZEMIIYQQOiXJhhBCCCF0SpINIYQQQuiUJBu6kpAA33wDTk6g\nr5/1+ZtvssqFEEKIt4gkG7qQkAAtW8L587BlC6SmZn0+dy6r/AUSjvfffx87Ozvtse4FtWLFCkqW\nLImHhweurq7MnTtXW/f111+jKApXr17Vls2bNw9FUXi0/HfZsmVUq1aN6tWr4+7urj0AztfXF2dn\nZ+3Ong0bNnzudxNCCPFmkmRDF+bOzerJ8PcHDw8wMMj6vH49lC+fVf+cfH19CQwMLPD1Dx480H7t\n5eVFUFAQhw8fZsqUKTm2Pq9WrRr+/v7a7zdu3KjdMCw8PJwpU6Zw6NAhzp49y7Fjx6hevbr22pkz\nZ2p39jxy5Mhzv5sQQog3kyQburB0KYzP54j58eNh2fMdMQ/QtGlTbGxsnnhNXFwcixcvxtPTk1mz\nZuWqt7W1xcXFhcjISG1Z165dtb0V165dw9LSkhIlSgBZR7+bm5trz1sxMzPD2dn5ud9BCCHE20WS\nDV24dQvyG+Zwd8+q14FDhw7h6+tL7dq1CQ0NZc2aNUydOjXXdTdv3iQlJSVH74SFhQWOjo6cP38e\nf39/vLy8tHU1atTA3t4eZ2dnBg0axLZt23K09/nnn2uHUeSEViGEEP8myYYuODpmzdfIy/nzWfWF\nbOTIkXTq1InWrVsTEhLC9OnTqVSpUo5r1q9fT/Xq1XFxceHjjz/GxMQkR723tzf+/v5s2bKFbt26\nacv19fUJDAxk06ZNVKpUiVGjRvH1119r6x8fRlm7dm2hv5sQQojXmyQbujB4MEzN54j5adPg/ec7\nYv5JPvvsM/z8/Jg4cSKDBg3izz//5N/n3nh5eXH27FmOHDnCuHHjuHPnTo76jh07snr1asqVK4eF\nhUWOOkVR8PT0ZPz48fj7+xMQEFDo7yCEEOLN9NKSDUVRlimKEqUoyvl/lY9QFCVEUZQLiqJ8+1j5\neEVRriqKcklRlDaPlddWFOVcdt0CJfs4U0VRjBVFWZ9d/reiKE6P3TNQUZQr2R8Ddf6yo0ZBWBh4\necGZM5CenvXZyyurfNTzHTH/JE5OTkyePJmLFy/i7e3NwoULcXV1zbOnoU6dOvTv35/58+fnKC9W\nrBgzZsxgwoQJOcojIiI4ffq09vugoKCnHh0vhBBCPPIyezZWAG0fL1AUpQXQBaihqmpVYFZ2uRvg\nDVTNvucHRVEenYH+IzAEeCf741Gbg4EYVVVdgLnAjOy2bICvgHqAJ/CVoijWunnFbGZmsG8fVKsG\n3buDqWnW52rVssrNnu+IeYA+ffrQoEEDLl26pD1O/XH6+vq0b9+ezZs3c/DgwXyTgrFjx7J8+XLi\n4+NzlHt7e1OrVq0cZenp6YwZMwZXV1c8PDxYv359jkTl8TkbHh4epKWlPff7CSGEePO81CPms3sb\ntquq6p79/QbgJ1VV9/7ruvEAqqpOy/5+F/A1cAP4U1VV1+zyPkBzVVWHPrpGVdWjiqIYAHeAkmQl\nLc1VVR2afc9i4C9VVdc9KVY5Yv7NJn+WQgjx4l6XI+YrAU2yhz32K4pSN7u8DPD4ko3w7LIy2V//\nuzzHPaqqZgAPAdsntJWLoigfKopyUlGUk9HR0S/0YkIIIURBqKrK0aNH2b17N4mJiUUdjk4UdbJh\nANgA9YHPgQ2P5mAUBVVVf1JVtY6qqnVKlixZVGEIIYR4i5w7d46zZ8+SmJjIzp07izocnSjqZCMc\n2KxmOQ5ogBLAbeDx9aFls8tuZ3/973Ievyd7GMUSuP+EtoQQQogip6oqiqKgp1fU/yTrTlG/2Rag\nBYCiKJUAI+AesBXwzl5h4kzWRNDjqqpGAnGKotTP7gEZAPyW3dZW4NFKk57AH2rWhJRdQGtFUayz\nJ4a2zi4TQgghily1atWoWrUqxsbGtG3b9uk3vIYMXtaDFEVZBzQHSiiKEk7WCpFlwLLs5bBpwMDs\nBOFC9uTRi0AG4KeqamZ2Ux+TtbLFFNiZ/QGwFFitKMpV4AFZE0NRVfWBoijfACeyr5ukqur/DgwR\nQgghipCenh6NGjUq6jB06qUlG6qq9smnql8+108BpuRRfhLItRe4qqopQK982lpGVmLz0qSlpXH0\n6FHOnDlDXFwcFhYW1KxZkwYNGmBkZPQyQxFCCCGKVFEPo7yR0tLSWLVqFdHR0Xh7e/Pll1/i7e1N\nVFQUq1ateqF9KBRFYfTo0drvZ82alWPr8MjISFq3bq39ft68eZiYmPDw4cPnfqYQQgjxIiTZ0IGj\nR49iZWVFjx49cHBwQE9PDwcHB3r27ImlpSVHjx597raNjY3ZvHkz9+7dy7M+MDCQNm20G66ybt06\n6taty+bNm5/7mUIIIcSLkGRDB86cOUPjxo359ypeRVFo3LgxQUFBz922gYEBH374IXPnzs2zPjAw\nkHbt2gFZR8UnJCQwefJk1q174h5mQgghhM5IsqEDcXFx2NnZ5VlnZ2f3wkMafn5+rF27Nlc7mZmZ\nXLp0CTc3NwD8/f3x9vamSZMmXLp0ibt3777Qc4UQQojnIcmGDlhYWBAVFZVnXVRUFJaWli/c/oAB\nA1iwYEGO8r///pt69eppv1+3bh3e3t7o6enRo0cPNm7c+ELPFUIIIZ6HJBs6ULNmTQ4ePJjriHdV\nVTl06BAeHh4v/IxPP/2UpUuX5tjadufOndo12ufOnePKlSu89957ODk54e/vL0MpQgghioQkGzrQ\noEEDHj58yKZNm4iMjCQzM5PIyEg2bdrEw4cPadCgwQs/w8bGht69e+c49XXfvn20atUKyOrV+Prr\nr7lx4wY3btwgIiKCiIgIwsLCXvjZQgghxLOQZEMHjIyMGDBgAHZ2dmzYsIEpU6awYcMG7OzsGDBg\nQKHtszF69GjtqpTo6GhMTEwwNzcHsuZrdOvWLcf13bp1w9/fv1CeLYQQQhTUS9vU621jZGREs2bN\naNasWaG2m5CQoP3a3t6epKQkANasWZNjf43r16/nunfOnDmFGosQQghREJJsvCH69ctzI1YhhBCi\nyMkwihBCCCF0SpKNZ/TvFSbi9SN/hkII8XJJsvEMTExMuH//vvxj9RpTVZX79+9jYmJS1KEIIcRb\nQ+ZsPIOyZcsSHh5OdHR0UYciXoCJiQlly5Yt6jCEEOKtIcnGMzA0NMTZ2bmowxBCCCFeKzKMIoQQ\nQgidkmRDCCGEEDolyYYQQgghdEqSDSGEEELolCQbQgghhNApSTaEEEIIoVOSbAghhBBCpyTZEEII\nIYROSbIhhBBCCJ2SHUSFEK8sVVUJCQnh6tWrhIaGEhoairGxMc2aNaNx48aYm5sXdYhCiAKQZEMI\n8cqJiopi1apVLFu2jMTERKpWrYqzszPOzs4kJiYyc+ZMevfujZubG82bN6d9+/Y0atQIAwP5X5oQ\nryL5mymEeKUEBwfTokUL2rZty+LFi2ncuDGKouS6LiUlhePHj7Nv3z4+++wzwsLC6NChA927d6dz\n58553lPUNBoN+/bto1ixYjRq1KiowxHipVHkuPS81alTRz158mRRhyHEWyUsLIwmTZowefJkBgwY\n8Ez33rx5k61bt7Jo0SI6dOjAjBkzdBTl81u5ciXh4eFoNBrq1KlDu3btijokIV6IoiinVFWt87Tr\nZIKoEOKVkJGRQZ8+fRg+fPgzJxoA5cqVY/jw4ezfv5/t27czZcoUXrVfpmJiYnBwcMDc3Jy7d+8W\ndThCvDSSbAghXglz587F1NSUMWPGvFA7tra27N69mw0bNtC3b1/i4+MLKcIX16FDB6Kjo8nIyKBD\nhw5FHY4QL43M2RBCvBJiY2MpXbo0enov/jtQmTJlOHbsGCNHjqRSpUo0a9aMunXrUqVKFapUqUL5\n8uUL5TnP6p133mHcuHEv/blCFDWZs5EPmbMhxMuVkJCAu7s769ato0GDBgBkZmaybt06Lly4QK9e\nvahQoQLXrl3D2dkZGxubArV75coVjhw5wpkzZwgODiY4OJj79+9TqVIlmjRpwldffYWtra0uX02I\nN1ZB52xIz4YQ4pVgZmaGk5MTiYmJAKSnp9OxY0fi4uJo1qwZPXv2JDExkVq1aqHRaOjYsSMODg6U\nKFECBwcHXF1d81yB8s477/DOO+8wcOBAbVl8fDyXLl1i9erVuLu7M3v2bPr06fNKrmAR4k0gPRv5\nkJ4NIV6uzMxMHBwcCAoKokyZMvj5+XHjxg22bt2Kvr4+qampjBw5EltbW44dO4azszOxsbHcu3uX\nGxcvYmZhwUdjxtC/f38sLS0L/Nzjx48zZMgQSpcuzY8//oiTk5PuXlKIN4ysRhFCvFa2b99OxYoV\nKVOmDBs3bmTv3r388ssv6OvrA2BsbMzQoUOpUKECU6ZMYenSpQQEBLDfzo4bDRrwfVwcB3/9FRcX\nFxb27EnGwIEQHPzU53p6enLy5EmaNWtGnTp1mD17NhkZGbp+XSHeKtKzkQ/p2RDi5WrVqhWDBg2i\ndevWVKtWjS1btlC/fv2n31ivHvTsCevXw/jxXHj4kJHDhxNtbs53VlY0vXSpwDFcvXqVYcOGERcX\nx99//y3DKkI8hfRsCCFeGxcvXtROAl2/fj2tWrUqWKIB8MMPsG8fNGkCXbpQ1dWVvTY2/Ld6dbqE\nX6OYZzF81vigUTVPbcrFxYWffvqJK1euvHJ7dAjxOpNkQwhR5BYuXMjQoUMxMjIiOTmZkiVLFvzm\n2rUhMBDmzoWHQZA2BeXb2vSs7UHiKD06u3fG/2N/3D52Y2XQyqc25+zszDvvvENAQMALvJEQ4nGS\nbAghilRsbCzr169n6NChABw5coS6des+X2PHh4J9C7C6geLXiH7vVOeG50U0vTTcXH+TQd6D2HVw\n1xObUBSFiRMn8vXXX5OZmfl8cQghcpBkQwhRZDQaDR9//DHdu3enVKlSZGZm8tdff9Ki+CY4PgzS\nE56tQVMHiD4MKXch+FuWWNxlj10sJiWg1pha6Nnr0bZZWz5o2pTECxfybaZt27aYm5uzcePGF3xD\nIQRIsiGEKCKqquLn50dkZCQLFy4Esno1SlmBnb0tsZEXuHfkq2drtOEasGtKssciuHcUpe0pzBUN\nTQ7DgAaebF2ylam1rdl66CB7a7qza+XMPJtRFIVJkyYxadIkNJqnz/UQQjyZJBtCiCIxfvx4Tp06\nxdatWzE1NQVg4sSJjB5Qnwc3jqFJCCMo+BZXr159ckNpabB0KWzeDIZWPLAbyHcBVziV1Bx+dYBi\njvzSHz6MWcuGY1Ppfl/FvZEeH2jAf93/cfr06Tybfe+99zAyMmLv3r2F/OZCvH0k2RBCvHTTp09n\n27Zt7Ny5E3NzcwCSkpI4duwY3qN/ITTDgyjrvoSbtCU0NJTt27cTeuU87GsJG63g2rL/NTZhAixf\nDuPGwbp1REREYG9vj1WLJfwYuwDsm7P6EFBlLGVSo/mmvobvgzWUqQirQlJY9tsykpKScsWoKAp+\nfn58//33L+mnIsSbS5INIcRLtXjxYn766Sf27NmT40ySQ3s24VHdjeIWNji1ns2pmGpYW9ty8uRJ\nLC0tObXjWzTpCdD0Nzg/+X8Nhodn7bXh6grh4bi4uJCWlsb69etp1KgJVPClS20g+FtuR9uxv7IF\nbiMg2NsIjQJL/lqM3yK/PGPt27cvx44dY/LkySQnJ+v4JyPEm0uSDSHES7Np0yYmTZrE7t27KV26\ntLb8YdB3TPvP+3hVOQt392NnY45PC1vaNffQbqwVlVYaJfEG/P0BOLz7v0anTIErV8DaGoYNw8TE\nRHuQ2+7du3mQUZKKo+CfhrvYnPYPrrgCkE4aPxmAzYEMSv6wgt9P/p4r3uLFi3Ps2DGCgoKoUqUK\nGzdulP03hHgOsoNoPmQHUSEK1549e/Dx8WHPnj3UqFFDW37ixAkmj2nGkWvJ/OcLhWolPWllZwbp\nsZB0m5vVfufspQjc3NyoUKo4JIaCTV3Q08/3WbNnz6ZkyZLcu3cPe3t7+vXrR0h0CHV/rEsVTRWu\ncIVKsSbsXBSJa9fiKP6JVBrpwYH5Z/Jt86+//uLTTz/FxsaGFStWUK5cuUL9+QjxOpIdRIUQr4zj\nx4/Tt29fAgICciQai39ZTKtOrfg7NpnT/wUvK5XZNyPh/jGo8x0YFKecTSYdO3akQoUKYGoPJeo/\nMdEAsLa25tatWyQkJGh7UCqXqMyq7qswxph+9KO9+QfcsTXnn+2JfFRSn7+XXeT8+fP5ttm8eXNO\nnTpFmzZtqFOnDv7+/oXzwxHiLSA9G/mQng0hCkdwcDAtWrTg559/plOnTgBERETw6fhP2bhpI+bv\nmpNQOwFjxZAU0viy8pd8U6sUXJgKDq2g3tKnJhf/lrVfx1+Ym5vj6emJoiioqsrZs2fZsmULkLX0\nVi8jgwcxJ7hpY0z6WZWTx0+ya9cuqlWr9sT2T58+TceOHVm/fj1NmjR5vh+MEG+AgvZsSLKRD0k2\nhHhxN2/epHHjxkyePJkBAwYAkJiQQPPmzbkcdgWDngZ0dOjIBjbgjjvebt6M7jW60ON4lGwcOHCA\nmzdvkpLz7UG9AAAgAElEQVSSQlJSEg4ODty+fZvk5GSSk5Nxd3dnxIgRrF27lvfee++JbW7ZsoUR\nI0awZMkS2rRpU+gxC/E6kGEUIUSRio6OpnXr1owaNUqbaIRdPsV7dUtQ0fA0349qTGn7UlTDnwuZ\nxhxt8PETE42UlBRiYmJeKKa6detiaGiIkZERPj4+dOnSBXt7ezIzM0lJScHLy4uAgAD69+/PsmXL\nnthW165d+fnnn/noo4+YPHnyE68V4m0nPRv5kJ4NIZ5ffHw87777Lq1bt2bKlCkA7NqwkP+MG0Ut\nZ4WWrTvStOTf7I1sST+zNWgOg74HpNj9jOkHH+Rq78GDByxduhSAevXq0bRp02eK51HPRn7mzJlD\neHg4c+bMAeDSpUu0b9+evn37MmnSpCceNR8ZGUm1atU4cuQIlSpVeqa4hHjdSc+GEKJIpKam0rVr\nV2rVqsXkyZPRqBpaTmlB509HEmebSZUuGbiX3M61FDdq1aiDog8JdsUhA8Jnz85zx9CwsDAcHR3p\n3r07ISEhzxVXeHh4vh+xsbE5rq1cuTJHjx5lz5499O/fn5SUlHzbLVWqFOPGjePTTz99rriEeBtI\nz0Y+pGdDiGeXmZmJl5cXiqLg7++Pvr4+83+fz2djRmFmpBLXDaoYgt59fQJ+c6TMnSiMa6Vi4JiJ\nugdO9B3Jg7ZtadeuXY52ExISWL58OQ8fPqRDhw7UrFnzmeKqUaMG9+/ff+I148aNY/jw4TnKkpOT\nGTRoEJcuXWLjxo24uLjkeW9aWhpWVlZERUVhZmb2TLEJ8ToraM+GwcsIRgjxdliwYAGRkZH88ccf\n6OvrExkZyeo5M1ETVQy6gHEmuGYorF2QiUnmjaybDkGYkxPGBilcffiQ5tWr52rXzMyM4cOHo9Fo\n0Nd/tpUpAP/8889zvY+pqSnr1q3jhx9+oEGDBgQEBOQ5hGNkZISNjQ0xMTGSbAiRB0k2hBCF4vLl\ny0yZMoVjx45hbGxMfGo8tUdXIfLUQ7r2h6QocEyA//yhYpKZ817Hmze5oadH7DvvYGhomGf7iqLk\nSjTi4+PZu3cvRkZG2oPTCtujM1JcXV3p2bMn+/bty3NprLW1NQ8ePMDR0bHQYxDidSdzNoQQLywz\nM5NBgwbxf//3f9qhhmU7lxG99SGrPtHjlCXsOANj40vjmKxPjsFbKyt+/OILvu3XjxQTEywsLAr8\n3D/++AMDAwMiIyP58ccfCQwM1NmR8C1btmT+/Pm0b9+emzdv5qo3MjIiIyNDJ88W4nUnyYYQ4oUt\nWLAAAwMD7ZyHu3fvMnPkTMw66vO5AXSyUojWL4tBiBE7+g4ixcSEVCMjEkxNWV2iBCv37iVdURgy\nZAgmJiYFfq6JiQkxMTHcuXMHR0dHbty4wbVr13T1mvTp04cBAwbkudRVo9Ggpyf/SxUiLzKMIoR4\nIZcvX2bq1KkcO3YMPT09UlNT6datG4MHDWb4yIEEHZyCq+O7OIzx4cKFC8SePMmlli1x/u47HhYv\njt3QoUwyNcXV1fWZejUgq7fh5MmTJCYmkpSURHx8/DO38ax69OhB//79c5VbWFgQERHxzJNXhXgb\nyGqUfMhqFCGeLjMzk6ZNm+Lt7c2IESNIT09n8ODBJCUlsWHDhpf2m35KSgoXLlzAzs6u0OZM3L17\nl/DwcFxdXSlevLi2PCMjAxsbG27cuIGNjY22fPbs2Vy+fJnFixcXyvOFeB3IPhtCCJ2bP38+BgYG\n+Pn5cenSJbp168aePXsYM2bMSx1SMDExoXbt2oWWaMTHx7Ny5UquXLnC3Llzc9QZGBjg7Oyca7gm\nJiYGBweHQnm+EG8aSTaEEM/l4sWLTJ06lWXLlhEaGsr82VM5fGAfE8d9jN5//8utBg14cPhwUYdJ\nbGwsP/74I3Pnzs1zYmdeEhMTMTQ0pHr16vz3v/9l69at2rr09HSuXr2Kq6urtkyj0fDnn3/KEIoQ\n+XhpyYaiKMsURYlSFOX8Y2VfK4pyW1GUoOyP9o/VjVcU5aqiKJcURWnzWHltRVHOZdctULL3EVYU\nxVhRlPXZ5X8riuL02D0DFUW5kv0x8OW8sRBvrsTERHr16sXMmTOxs7Nj8aZFrNu8mjH9Teh3eRpO\nIcFkuLiQNnhwkcapqipnzpyhbNmyNGjQgL///rtA99nb2+Ph4cHRo0cBmDBhgrZu27ZtVKpUCXNz\nc9LT0wkNDeWjjz7C0NCQtm3b6uQ9hHjdvbQ5G4qiNAUSgFWqqrpnl30NJKiqOutf17oB6wBPoDSw\nF6ikqmqmoijHgZHA38AOYIGqqjsVRfkYqK6q6jBFUbyBbqqqeimKYgOcBOoAKnAKqK2q6hNPdJI5\nG0Lkz9fXF1VVWbFiBSOnj+THTT9S0TCTyHYmhCWpGP5sTLSzC4qi4FREf48SExNZsWIFDx48wMDA\nABMTExo1aoSnp+cztdO8eXP279+vPfvE3d2dTZs24erqSt26dcnIyKB27dqsXLlS55NThXjVvHI7\niKqqeuDx3oan6AL4q6qaCoQqinIV8FQU5QZgoarqMQBFUVYBXYGd2fd8nX3/JuC77F6PNsAeVVUf\nZN+zB2hLVjIjhHhGy5cv5/jx45w4cYLRA6qxJuECmmtQfURZbivhHLBxI61NdWzu3+d8w4Z0Dg3F\n2dmZzMxMAgICuHHjBs2aNaNevXo6jfPq1atYW1vTrl07duzYQY8ePShVqtQzt+Pt7c3+/ftZtGgR\nbm5utG3bloYNG9K1a1e6devG7NmzdRC9EG+WV2HOxghFUc5mD7NYZ5eVAW49dk14dlmZ7K//XZ7j\nHlVVM4CHgO0T2spFUZQPFUU5qSjKyejo6Bd7KyHeQOfPn+eLL75g48aNhIeHcz0+mJQ9hrSpYcDF\nhBh66LUkOLUjl6pW5UiTJqhWVty7dw+A27dvEx0dTb9+/fjrr790Hmu5cuWIiIggICAAd3f350o0\nALp16wZAQEAAixcvZtiwYQQFBXHu3DmmTZtWmCEL8cYq6mTjR6AC4AFEAkX6K4Kqqj+pqlpHVdU6\nJUuWLMpQhHjlJCQk0KtXL2bNmkXVqlU5cOAAEYdN0auSzhWPTCaWdGJhueP4Of1AMeUBGo0GRVGo\nnn3Wia2tLSkpKWzZsoXy5cvrPF5ra2tatGiBsbExSUlJTzxi/kns7e1p1qwZtWrVokuXLnh6enL8\n+HEaNWqkk+3RhXgTFWmyoarqXVVVM1VV1QA/kzVHA+A28PgatrLZZbezv/53eY57FEUxACyB+09o\nSwhRQKqq8tFHH1G/fn0GDhxIemY6a4+uJUTRY0Y1X4aYfkGL0pFsiOjF/TRbyhjfxsDAgJSUFHbs\n2EFmZibFixdn2LBhdOjQgV69er2UuPft20e7du24du0at28//1/7nj17UrZsWebOnYuiKCxfvvyl\nvYMQb4IiTTYURXm8X7Mb8GilylbAO3uFiTPwDnBcVdVIIE5RlPrZ8zEGAL89ds+jlSY9gT/UrF9l\ndgGtFUWxzh6maZ1dJoQooGXLlnHmzBm+//57AMZtHscB/wPYdjfAvtZflLG+ydGY+gxyXIGRXho3\nUiqSkZHBe++9x507dwgLCwOgePHilC9f/rlObn0ednZ2HDlyhJSUFCwtLZ+7HR8fHwIDA4mMjASy\ndgvNzMx8yl1CiEde5mqUdUBzoARwF/gq+3sPslaJ3ACGZicUKIoyAXgfyAA+VVV1Z3Z5HWAFYErW\nxNARqqqqiqKYAKuBmsADwFtV1evZ97wP/Cc7lCmqqi5/WryyGkWILGfPnqVly5bs378fNzc3NBoN\nJV1LklA8gdujoRiZmCqZrLjZlzu4kZaWgaGhIcbGxlhaWhITE8OQIUOwsrJ66bGnpqZy5coVSpUq\nha2t7TPdq6oq4eHhmJqaUqJECfz8/LC2tmby5MmsX7+eCRMmcOTIEezs7HQUvRCvvoKuRpHtyvMh\nyYYQWTtp1qlThwkTJjBgwABUVaWjb0cCfw2k7ey2bCm2g1XHrRhQN4HTe024YeDFQ09PSpQoQWJi\nIp6entjZ2WFtbf30h71iDh48yJkzZ0hNTaV3796kpqbSsGFDfv/9d+rVq8dXX32Fv78/CxYsoE2b\nNk9vUIg3kGxXLoR4IQkJCXTo0IFWrVoxYMAAUlJSqNWpFrt27ML6Q2uauzXnz4jaDKgTBzEq8ftM\n6RYQQPHr10lJScHMzIzKlSu/lokGwM2bN2nYsCEVK1YkPDycSpUqsWLFCjp27MjKlSsZO3YsM2bM\nYPjw4bRv357g4OCiDlmIV5YkG0KIXB4lGpUqVWLhwoUAjBo1iivHr2DZxhKvuzZYrNyE91cX+dN+\nJ8p8J1qNGoehrS0N3dxwcnKiffv2T3nK88nIyODgwYMcOnSIjIwMnTwDoFGjRhw6dIj79+9rV9R0\n7NiRoUOHMmLECD788EPatWvHhQsXeO+992jatCkjR44kLi5OZzEJ8bqSZEMIkUNiYiIdO3akYsWK\n/PTTT6DJ5DO/AWzcuIF2/drRVa84X/52ncrLjrPGxobWrVtjsGYNHDyI8sEHOPfvT8OGDTE2NtZJ\nfAcPHiQ0NJRr165xWIdnrzg5OfHpp58yZMgQzM3Ngax5HEZGRuzcuZPff/+dadOmYWRkxKhRowgO\nDiYpKQl3d3d+//13ncUlxOtIkg0hhFZSUhIdO3bE2dmZJUuWoADTP3Bi84Y1VB2SyF8l/sQmIY6S\ntRrhOfY/tKtSBY4cgUWLwNsbxo2DrOOKdCYjIwNjY2OMjIxe2oqQmJgYOnfuTPv27QkLCyMwMJDO\nnTsze/ZsPD09SUpKokSJEixZsoTly5czcuRI+vTpw5EjR8jMzOT+/fs67YUR4lUnyYYQAshKNDp1\n6kS5cuVYsmQJsbGxvO/bl8XbI5gxbwmHTFKZYO7Lb/XMeehgRbG//0aZNQs6dQJ3d/Dzg9BQncfZ\ntGlTrKyssLGxoXHjxi/UVmZmJleuXOHBgwdPvO7OnTucOXOGkSNH0rNnTxo2bIi3tzc//fQTV69e\nxcHBgePHjwPQsmVLzp07h5ubG8OGDcPOzo7SpUvTtGlTtm/f/kLxCvG6ktUo+ZDVKOJtkpycTKdO\nnShdujTLly9HT0+P3r168MfeXfwyxp73nG8wPLoEix7eo61LW1a2XUlGWgalHBygdGkYNAgWL4ag\nIHB0JDIyku3bt2NmZka3bt0wMTEp6lfM0+bNm7l37x6xsbEMHjw43+WxwcHBdO/ePc9JoJmZmTRt\n2pSjR48yevRoZs6cmaM+PDwcc3Nzfv31V0aNGkVYWJgc2CbeGLIaRQhRIMnJyXTu3BkHBweWL19O\nXPhxPh9YlyN/beP38aY0KX+HNdEjaGo/n4z/y2BRk0UsX7ocf39/jh47Brt2QWoqrF0Ljlmb9f75\n55+4ubmhKApBQUFF/Ib5i4iIoEmTJtjY2PC85yHp6+tz+PBhZsyYwaJFi6hVqxY3b97U1pctWxZL\nS0t8fX3p168fTZs25eDBg4X1CkK8FiTZEOItlpycTJcuXbCzs2PlypVo0hP5rk8Tfl1/hsOdM7ht\n0Ix7aSUorkbhWa8eeooeN2/epFKlSjRt2pTQ0FCoXh1mz4a2bbXtlixZkvPnz3P79m1KlCjx0t9L\nVVXS09Ofel3Lli3Zs2cPFhYWuLi4FPwBmkxQNTmKbG1t8ff3Jy0tDXd3d9asWZMrgVmwYAGfffYZ\nAwYMoG3bthw4cOC5z2wR4nUiwyj5kGEU8aZLSUmha9euWFtbs3r1avT19RkxeAB/rFjDwmaVePfY\nZeJmmxNjVoHSPf/CqFjWDqAPHz7kl19+ITExkR49euDs7JyrbY1Gw+XLlylevDiOjo656nVJo9Gw\nePFioqKiKF++PL6+vi/cZo5hlDt/wMFuoG+K+u5eQiL1iYuL49ChQyQkJKCvr8+ff/5JUFAQvXr1\nYtiwYVSvXh09vf/9bpeWlsaKFSuYNWsWtra2rFu3DicnpxeOU4iXTYZRhBD5SklJoVu3blhaWrJ6\n9WoMDAzwHeXLuq2/8YcKLQ5fITNTjx1GcyjX77Q20QCwtLTko48+YsyYMXkmGgB6enq4urq+9EQD\nso6yj4qKokuXLoSFhZGYmFi4D7j2M9SYCk79uH9qAYGBgdrzV7y9vbVzOJycnPjtt9/o0aMHtra2\ndOvWjcDAQACMjIz48MMPCQkJoXfv3jRq1Aj55Ua8ySTZEOItk5qaSo8ePTA3N2ft2rUYGBgwcvZI\nVi9djaFXKsvaWnDR3YM4K1tKlC9PSEhIUYf8TGxtbdHT02P79u3o6+sX/n4fpTvC2f9DvbaMvRey\nejXi4uLIyMggICCAypUrY2RkRJs2bRg7dizXrl0jJCSEjh078sknn9C7d2/u3bsHZCVlo0aN4ocf\nfqB9+/Zs27atcGMV4hUhyYYQb5H79+/TpUsXihUrpk00Tp06xeKvF9PC15JejgZct4yjytl/ONWy\nJWlpaezcuZOIiIi8G7x1C0aOhO++g1dkSLZYsWJ88MEH1K9fn6FDh2JgYFC4D3D2gXan+av4d9xM\nyerZMTAwQF9fHx8fH7y8vOjXrx8XLlzQLs21t7dn8ODBBAUFUa5cOapXr87Ro0e1TXbp0oXt27cz\ndOhQ7cm6QrxJJNkQ4i2xf/9+PDw8cHNz45dffsHQ0JBTp07RqVMn/D71I8EigcPhqXyUaYlybhGn\nPTw4d+4cmZmZ+W+eNXgwZGTA99/DK7RrZqlSpWjVqhUlS5bUSfv3kovz99mbpKamAlkbjbm7u1Ou\nXDkURcHR0ZEbN25gaGiY4z5TU1MSEhKIjIxk1apVOeo8PT05fPgwEyZM4O7duzqJW4iiIsmGEG+B\n5ORkWrZsyfDhw5kzZw6GhoasXr2aVq1a0bZtW2rYV+DI9Ez+PmhArX3JhASE03TLFir+8w/JycmU\nLVs2q+di69asJa6P74apqlkfOt459FXy8OFDLC0tad68Oebm5vz3v/+la9euKAX4Gbi5uQHw2Wef\n5apzdnamdu3aHDp0qNBjFqIoSbIhxBssLS2NhQsX4uLiQocOHfD29gbg559/ZugHA1j8oQnvuB/l\nyIUxKBoVg/PpKLHpOMz9DvT0aLdtG6UePMj6R3TNGhg9GubNg0mTsh6wbBkYG2cNpTzDwWuqqnLy\n5Em2b9/+1N07X0XOzs44OTlx+fJlunTpkmOlydOMHDmSOXPm0Lt3b5KSknLVDxs2jJkzZ8qSWPFG\nkWRDiDfUP//8g6enJ7///jvbtm3jt99+w9HRka6DuvLZhBEc+saA+BLlGWsfwpyKqWSYQKKRERga\noqcoaIoXR9XTo0HdulkN3rgBHh7QtOn/tiUvWzYr+fj442fq2bhx4wZHjhzByMiIX3/9tfBfXsf0\n9PRo164dgwcPpmLFis98/6effoq7uzsTJkzIVde9e3diYmL466+/CiFSIV4NhTxzSojXk0aj4cKF\nC1y/fp3Q0FBCQ0OJiIggISGBxMREUlJSsLKywt7eHgcHhxyfy5QpQ6VKlQp/IuIzePjwIevXr8fb\n25tixYoxffp05s+fz8yZMxk4cCDno85TY0wNLq2+RJpJGl0GNqWcw36M0sLIVCFtqYJpT5XUX/W5\nOnEiVrVqUXbGDB6MGUO1fv2yHjJ8OAwbBhcvZs3ReAGKoqDRaMjIyCjQ0ENRMzEx0a4oKajr16/n\n2+OhKAoTJ06kfv36zJw5M8d/O/r6+owZM4a5c+fSokWLF45diFeBbOqVD9nU6832KLk4ceIEf/zx\nB7t27cLa2prKlSvj5OSEs7MzZcuWxczMjOLFi2NiYkJsbCx3797lzp072s937tzh1q1bREZGUrt2\nberXr0/jxo1p06ZNrsmBuhAeHs78+fNZtmwZJiYm1KtXj1u3bmFjY8OSJUtwdHTk7NmztOzfEvVu\nMh91Tmdx2QxM08DBRMM3pp7UMzyLuU0K146WxXibwoiqVflt506dxx4UFMSdO3eoX78+VlZWT7+h\niJ06dYrIyMgCX29oaEirVq3Q19fP95p69erxzTff0Lp16xzlSUlJlCtXjhMnTuS7l4kQr4KCbuol\nyUY+JNl4c6WkpNC9e3dCQkJo0KABzZo1o02bNpQvX/6524yJieH48eP8/fff7Nmzh6tXr+Lr68uQ\nIUOoUKFCIUaf5dy5c8yaNYtt27YxcOBAPv30U1RVxd/fn3LlytGnTx8uX76Mj48Pp06dosHgBqxo\nfIYVsSo9zFPZHQeNjKDK2uLseKcHXuvXYZyaTnq5cvx6/z7l9uyhQYMGhR63yMnPzw8XFxdGjRqV\nq27MmDGYmJgwefLkIohMiIIpaLIhwyjirXL48GGGDRtGtWrVuHz5cq6hj7S0NG2yEBUVRVRUFPHx\n8SQmJqLRaLCwsMDc3Bxzc3PKly9Pw4YNqVGjBtbW1rRp04Y2bdrwf//3f4SEhPDzzz9Tr149unbt\nysSJEylduvQLxZ6UlMSWLVtYvnw5Fy5cYMSIEcybNw9ra2vtNePGjdN+3bRpU6KiogDY9+M+Yrd7\n0N+0OBXTzlD5NBxw1cemXBL1xo2DqlXRBASwv3RpnH77jQexsS8Ua1G5f/8+0dHRuLi4FOmwVkGV\nLl2aW7du5Vnn6urKgQMHXnJEQuiG9GzkQ3o23iz3799n3Lhx7Nixg7lz59KrVy/tXAFVVblw4QIr\nVqxg9erVVKpUCQ8PD+zt7bGzs8Pc3JzixYujKArx8fHEx8cTFxfH1atXOXz4MGFhYdSpU4eGDRvS\nuHFjWrRooT1SPSYmhunTp7NkyRKGDRvGF198gaWlZYHjVlWVo0ePsmLFCjZt2kS9evUYOHAg3bp1\ny3dnzJMnT/LVV19x7tw5vvzyS3x9fTEyMoLkuxC6gg0BAXQtewq9ZIVzYX05WKYOBmlpVJowgebp\n6egZG6O3dSu8996L/+BfogcPHrBkyRJsbGwwNzfHy8urqEN6qpMnT9KnTx8uXbqUa37HxYsXad26\nNd9++y19+/YtogiFeLKC9mygqqp85PFRu3ZtVbz+NBqNumrVKtXe3l4dMWKEGhsbq6qqqmZmZqob\nNmxQ+/btq5YqVUp1dHRUx40bp16+fPmZn/HgwQN1x44d6pdffqk2adJEtbKyUvv166du3bpVTU1N\nVVVVVW/evKn6+vqqdnZ26tSpU9WwsLAnthkTE6NOmTJFfeedd1RXV1d1+vTpanh4+BPvOXPmjNql\nSxe1dOnS6nfffaempKTked3NmzdVZ2cndfr0aeqCBQvUL7/8Uu3QoYMa8cEHqjpypKouWqSqPXsW\n+P01Go0aFxenajSaAt+jC5cvX1YXLVqkBgcHq/PmzSvSWApKo9GoNWvWVLdt25Zn/apVq9TevXu/\n5KiEKDjgpFqAf1OlZyMf0rPx+nvw4AFDhw4lJCSE5cuXU6dOHVRVZceOHUyYMAETExM++OADWrRo\nQYUKFQptVcSdO3cICAjA39+f0NBQRo8ezYcffkjx4sU5d+4cCxcuZPPmzTg5OeHq6krFihWpUKEC\nxYoVQ19fn0uXLvHNN9+gp6fHvn378PT0zDe26OhoduzYwaZNmzh58iTjxo3jww8/xNTU9IkxfvLJ\nJ5w7dw4PDw8aNWpEjx494NAh6Nw5awnrnDkwcOBT31VVVdatW0dYWBhlypShX79+z7TnRGHKzMzk\nt99+IyIignfffVe7edarbt26dSxatIj9+/fnqrt06RJt2rThxo0bLz8wIQpAJoi+IEk2Xm/79u3D\n19eXnj17Mm3aNExMTAgNDeXDDz8kIiKCKVOm0KVLF50vuzx9+jTTpk3jwIEDjBgxguHDh2NlZUVa\nWhonTpzgypUrXLt2jdDQUFJSUsjMzKR8+fK0atWK5s2bY2ZmlqM9VVUJCQlh69atbNu2jfPnz9Oq\nVSs6depEr169KFasWIHiiouL4+LFi9SvXz9nxd27kJICBZwsm5yczOzZs/niiy+YM2cOH3/8MRYW\nFgW6V8CWLVuwtLTE19eXLVu2ULNmzRz1MTExODk58fDhwyKKUIgnk2EUGUZ5K926dUsdOHCgWqZM\nGXXXrl2qqqpqRkaGOm/ePNXW1lb99ttv1fT09Bd+TmZmphodHa1mZGQU6PqLFy+qAwYMUG1sbNRx\n48apd+/efeo9scmxanB0sBobG6vu3btXHTVqlFqxYkXV0dFR/fjjj9XAwMB8h0peFo1Go/r7+6uT\nJ09WV69eXeRDKa8bQAXUfv36qSNGjMhVHxISorq4uBRBZEIUDAUcRnn1p2sLUQDx8fF8++23/PDD\nDwwbNoyLFy9iYWFBSEgIgwcPRk9PjyNHjlCpUqUXfpaqqqxatYp79+5hbW3NoEGDnjp0UKVKFVau\nXEloaCgzZ87ExcWF2rVrM3To0FyTPQ9dO8TPW35m887NJF9JRo1UqehUER8fHzZt2kSNGjVemY2w\nFEWhd+/eJCcnY2pq+kJxxcfHY2Bg8NQhoDfJ2rVr8fHxYffu3aSlpTFz5swc/y1ER0djZ2dXhBEK\nUTgk2RCvtbNnz7JkyRJ++eUX2rVrx5kzZyhXrhwZGRlMnz6dWbNmMXHiRD766KNCm0uQnJzM7du3\nGTNmDHPmzCEhIaHAQwfOzs6MHTtWu+Rx0qRJ+Pn50b59ezQaDcfPH+fq5atQEhQnhbbN2xLiHsIH\n6gcMGTIEa3OT/2fvvMOiuNf+fc+ygPTeFBERKSqKChgVO1bsvSWWGGOixl+6OeaU1Dd5NTkmJzHH\nFo0xYuwdRUqiYhdRBBTpKFKkd9jd+f2B7KsC0kXN3NfF5TI7851nVmCeecrnAVEFQu1CUU8bQRDq\nnb6pjYiICI4ePYogCMyZM6dy8NtfgFmzZvHdd99x8eJFoDLt9rC+SUZGhuRsSLwQSM6GxDOFQqHg\nzp07xMfHExcXR1xcHDk5OcjlcjQ1NZHL5RQXF3P37l3i4uLIy8tjwYIFaqVFURQ5cOAAq1atwtbW\nlqq1BFEAACAASURBVMuXL2Nvb9+sNuro6NClSxe++uorunTpgoGBQYOOLykpwdzcnAEDBuDl5cWI\nESNYvXo1mzdvptihGNkSGY7FjsTYxBAgBDC2YiwmFiYYZe2HgKXQxgqGnwHdF+eGHB0dzYgRI0hP\nT+f27dt/GWdDEAT+/e9/M336dC5fvoy1tfUj70vOhsSLguRsSDx1kpKSOHfuHJGRkdy6dYv09HQy\nMzPJyMggNzcXa2trOnXqpP6yt7dHoVCov7S1tfHx8aF9+/b06tVLLQf9559/snLlSoqLi1m9ejWj\nR49ukXSDIAhMmjSJcePGNUo4ysnJidTUVNLT0xk7diyiKLJ161aOHDmCm2c3XH504XbpbborujPf\nfD56OnqVqZqgAeC9BxJ+gbtHofPrTb6WvLw8Ll68iJWVFd27d2/yeo3F3d2dPXv2oKmpybx6dMG8\nSPTt2xdDQ0Pi4+MlZ0PihUVyNiRanMzMTIKDgwkKCiIoKIiCggK8vb1xc3Nj0qRJWFtbY2lpiaWl\nJaampmrn4f79+wQEBBAZGUlmZiaZmZnk5+ejqamJlpYW2traaGlpYW1tTVRUFDExMXz22WfMnDnz\nqbRfNlahUiaTMXToUPX3CQkJmJmZMaS4GGwcuG9pyS+vvEWChgiqylZalUoFdtPg7FyQaUL3L5rl\nGvbv34+5uTnBwcEYGRk1SbK9KTg5OfHBBx8gk8larXW2tRAEgVmzZuHn50e/fv0eeS8rK0uajSLx\nQiA5GxLNTkFBAadOnVI7F4mJiQwcOJBhw4axfPlyunbtWuMNRRQrlTz379/PkSNHuHnzJoMHD6Z3\n79706tULCwsLjIyMUCgUlJeXU1ZWRnl5OcnJyUycOJEFCxZUKmU+ZwiCgCiK8MMPsG4dhITQLjQU\nf1NTXF1d6dq1a+V1ub4DHaaDhj5oN8/gsqpIkYaGBkqlslnWbCxNkRfPysri6tWrdOjQgc6dOzej\nVU+HWbNm4e3tzb///e9HPgdLS8sGDX+TkHhWkZwNiSYjiiLnz5/H39+foKAgrl27hqenJ8OGDeO/\n//0vHh4eNU5AFUWR0tJSwsLCOHDgAPv376eiooJJkybx5Zdf4u3tXask9wvJmDHw3nugUBA7cSL9\ndHXpcOUK7b28/m+fdz6H9ethxgzw86sU4GoCkyZNIjQ0FC8vr+f6CdrPz49OnTpx4MABFi5ciJmZ\nWb2PrZrc6+rqWk3X5Gnh6OhI+/btCQkJYfhDMvGurq5s3bq1VWySkGhOJGdDosGUlJSQnJxMYmIi\n169fZ8uWLahUKiZOnMi//vUv+vfvr+5OyMnJ4dSpU1y5coXLly9z/fp1cnJyKCoqori4GC0tLZyc\nnJg4cSK7d+/G3d29UXUWSqWSsrKyJndFtAbqyMbSpfDTT3D7NtOSktA+d44rWlp0vHAB0tKgoAB+\n/hmyssDREZKT6y2+VRtmZmaMHz++ma6k9aioqMDQ0BCZTIZCoaj3cfn5+Wzbtg17e3vCwsJ4/fWm\n18E0lqpUyuPORnR0dKvZJCHRXEjOhsQTKS8v5+zZs6Snp2Nubo6Pjw9Q+SRmb2+Po6Mj//3vfxkw\nYACCIJCens7+/fsJDAzk9OnTpKen4+7ujoeHBxMmTOCf//wnZmZm6OnpqeW5m0pJSQmbNm2ioKCA\nvn37MmTIkCav+TRROxv790NMDKxdi+ny5RTr6XFryBAGHDkCSiUYGEDfvuDuXulkNHGK7IvEtGnT\nuHjxIkOGDMHKyqrexxUVFaGlpUWXLl04duxYC1pYNzNmzODzzz/np59+Ukf0HB0dSUlJoays7K8V\n5ZN44ZCcDYlasbS0JDMzEy8vL6ysrLhx4wYdOnRg+/bteHt7A6BSqTh16hTvvvsugYGBpKSkMHjw\nYHx8fHjvvfdwcXFpFofiSSQnJ6Onp8e0adPw8/N77pwNc3NzSktLidXXx1EU4e9/R9DVJXvOHAb8\n9hvCxo1Qlcc/eRJu3gQnJ6ghNfUikJWVhY6OToOiVLa2to1ql7W2tsbd3Z1Lly4xYcKEBh/fnLRr\n147u3bvj7+/PxIkTAdDU1MTe3p7bt2/TrVu3VrVPQqIpSM6GRK14eXnh5eXFP/7xj0e2i6JIWFgY\nO3bsYOfOnZibmzNlyhQ2btxI7969m1To1xjat29PYWEhW7ZsoU+fPk/13PVFpVIREBBARkYGQ4YM\noX379ur3dHV1eeedd/j71q34nToF+/bBihUgCAw/epRPlUoWVu2spQWt2KLa0pw6dYqLFy8iiiIL\nFizA3Ny8Rc8nCAKDBw9m8ODBLXqe+lKVSqlyNuD/UimSsyHxPCM5GxK10q5dOywsLNTf3759Gz8/\nP3bs2EFFRQWzZ88mICCg1adr6urqsnTpUkpLS9HT02tVW2rj9u3bJCUl0bNnT44ePcqSJUseeX/Z\nsmU4Ojrykb09bQwNEX/+mbNnz6JQKEhISGglqyuJjIzk3LlzdOrUicGDB7eoVHpMTAzjx4/n6tWr\nJCUltZizIYoihYWF6OvrPzPS7wBTp07lgw8+UNsGlc5GVFRUK1smIdE0/loN7RINQqlUkpmZydq1\na/Hy8mLAgAFkZWXxyy+/EBsby2effdbqjkYVGhoa1RyN5ORkgoKCyMjIaNFzi/WYnGxoaEh+fj6x\nsbEYGRlVe19fX5+dO3eiqamJUqlEFEX69OnD2rVr+X//7/+1hNn15tChQ3h7exMeHt6gzzInJ4fw\n8HAKCwvrfUyfPn3YtWsXmZmZODs7N8bcOlGpVPzyyy/88MMP7Nq1i9DQUL755hsOHz5cr//LlsTM\nzAxvb28OHTqk3ta3b19CQkJa0SoJiaYjRTYkakVTU5M1a9YwefJkvvjiC4YMGfLUUySNpaioCD8/\nP9zd3dm2bRvvvvtuizzBxsTEsHfvXvT19XnllVdqdCQAbGxsmD59OhkZGbi5udW4z7MUzn8YExMT\nwsPDqaioqHfkqLS0lM2bN9OuXTtOnz7NsmXL6vX5u7m50bVrVwRBaLGIQ15eHhkZGbz77rt8/fXX\n3Lx5kyVLlvDbb79x//79R6J5rUFVKmX27NkA+Pj4MHfuXDIzM1vdNgmJxiJFNiRqZc2aNaSnp7N1\n61aGDx/+3DgagFqgytTUlIqKihZ7Yr1w4QK+vr7Y2toSGRn5xH07dOiAp6cnbdq0eWR7eXk54eHh\n3Llzp0VsbCovv/wynTt3Zv78+XXqUJSVlREYGEhwcDAqlYo+ffqQl5dXqYBaT2QyWYumNoyMjLC2\ntmb16tW4uLjQrl07Dh06hCAI9R6o11JERESQmppKcHAwmZmZQOUsnhEjRnD48OFWtU1CoikIrR02\nfFbx8PAQL1++3NpmSDSBiIgIbt26Rc+ePenUqVOLnOPcuXOEhoaiUqmYO3cubRvRjrpr1y5KS0tJ\nT09n7ty52NjYNJt9FRUV3LhxA1NT06ciRR4QEEB2drbawSspKcHT05NevXo1eK2LFy8SGBiIra0t\ns2bNqlEYrrGIokhxcTG6urpUVFSQnJyMjY1Nq9f8fPvtt4wbN46ZM2cyZcoU/va3vwGwY8cO/Pz8\nJIdD4plDEIQroih61LXf8/OoKiHRQNzc3GpNWTQXffv2xdHRkTZt2jR4+msV2dnZeHh4UFZWRm5u\nbrM6G4cOHaKoqIiMjAxmzJjxSBdMU1EoFAiC8Ehrc5VmiEqlwsHBgQEDBjR6/dOnTzNv3jyOHj1K\ncnJyszqMgiCoHQstLS0cHR2bbe2mYGNjw59//knXrl05fvy42tnw9fVlyZIlFBQUNPrnTEKiNZHS\nKBISTcTCwqJJN4DRo0dz7do1rK2tcXJyakbLIDc3FycnJ0xMTMjNzW22dRMSElizZg1r1qzh7t27\n6u2DBg3CysoKe3t7+vbt26RzdOrUib1791JQUNAgoa7nmalTpzJgwAC+/vprIiIiSEpKAipTP/36\n9ePgwYOtbKGEROOQ0ii1IKVRJKCyo0KlUqlnbYiiSH5+PgYGBs/FdNI7d+4QGBiImZkZY8aMaTaB\ntYMHD2JpaUlpaSllZWWMGjWqWdZ9GJVKxb179zAxMXkuZeibymeffUZoaCj+/v4IgsDJkyd58803\niYyMfC4HDkq8mNQ3jSI5G7UgORsSsbGx7Nu3D0EQGDFiBD169OD3338nISEBCwsL5s+fj4aGBiUl\nJfj5+ZGVlcXYsWNxdXVtbdNbnFu3brF//36gUma7piFu58+fJyYmBi8vL1xcXIBKZ+3y5cvcv3+f\nfv361dq901wolcoWU7AtKSnh1q1b1fRomouKigr69u3L4sWLWbx4MVCZThk+fHirt0NLSFRRX2fj\n2X80k5BoJRISEnB3d6d///7Ex8ejUCi4efMmb7/9Njk5Oeq0xM2bN9HW1sbX15czZ860stVPB2dn\nZ5YvX85bb71Vo6ORlZXFmTNn6NWrF3v37mXXrl0EBAQQExPDhQsXUKlULZ4S2LdvH59//jk7d+5s\nkW6k3377jcjISLZs2UJBQUGzr6+pqckvv/zCqlWrSExMBGD16tV8+eWXZGVlNfv5JCRaEsnZkJCo\nBXd3d6Kiojh79iweHh7I5XK6d+/O6tWrsbS0xMTEBKhUWr179y6BgYFPpePjWaFqmF5NaGtro1Kp\nuH37NlDZgpySkqK+abY0qampREREoK+vT1xcXLPWq1Rx//59PDw80NLSahFnA6Br16588MEHLFiw\nAKVSSZcuXViwYAGTJk2iuLi4Rc4pIdESSGmUWpDSKBI1IYoipaWltGnT5hEtiLy8PAoKCmjXrt0z\nJX/dmty7d4/ExESSkpIQRZGMjAxGjRpFfn5+i6dRwsLCCAoKQhRFFAoFs2fPZvfu3WhqajJ37txm\nkUGPjIwkNDQUBwcHhg0b1mL/70qlkjFjxmBmZsYvv/yChoYG8+fPJyMjg4MHD0rTYCVaFalmo4lI\nzobEX5Xy8nICAgIoKSlhxIgRNToEJSUlCIJQTaCsJsrKyrhy5QrGxsZYWVkRFhaGvb09nTt3bgnz\nASguLuaXX34hOzubsWPHEhcXh42NDfn5+WhoaODj49Ni524JSkpKmDJlCjo6Ovj5+SGTyZg+fToy\nmYydO3c+V4J7Ei8WkrPRRCRnQ+KvytmzZ0lISMDU1JSSkhImT578yPuxsbHs2bMHgJkzZ2Jvbw9U\ndo/U1KGTm5uLv78/2tra3LlzBycnJyIiIli4cKG6y6elEEURQRAIDw/n5MmTiKLIlClTWkzkrSUp\nKytjxowZKBQK9uzZgyAITJgwASsrK7Zs2fJcdEdJvHhIBaISEhKNQk9Pj9zcXDIyMmqsyYiKimLA\ngAG89NJLREdHA5UOyueff87mzZspLy9/ZP8//vgDY2NjFAoFxcXF6OvrI5PJ1JLyLUlVasPd3Z0F\nCxawePHi59LRgMo6mN27d6Ovr8/48eNRKpXs27ePhIQE3nrrrVYfIich8SQkZ0NC4hmmvLycY8eO\nceDAgRYrQnyc7t27M3DgQFxdXWtMN7i5uXH69GkuXLhA165dgUrZ9tdffx2VSkVycvIj+xsaGnL3\n7l0yMjLw8PAgOjoaJyenpz5UzNzcHGNj4yavExoaypo1azh06NBTv8Framry22+/YWNjw+jRo1Eq\nlRw+fJjz58+r1UYlJJ5FpDRKLUhpFIlngdDQUBITEzEwMEAURSZMmNCg45OTk9UdIc1JVVSiSsMi\nNjaW7OxsoDKKoKWlRV5eHvfv38fc3JzS0lI0NDTIz8+nqKiIiooK2rVrVy9lUCcnp2aVWW8KKpWK\nzz//nCVLlrB9+3bmzp2LpaVlg9cpLi6mTZs2jU59qFQqlixZQkREBP7+/iiVSry9vXn33XdZtGhR\no9aUkGgM0mwUCYkXgDZt2lBQUIBSqWyUZPd7771HdHR0o26IDUVXV1c99VVDQ4M+ffpw9+5d2rZt\nq9bWcHZ2RqFQ0KZNG27cuFHnpFtRFImKiuL48eO4u7u3+DXUhUwmw9bWlv379yMIQqO6aQICArh0\n6RKmpqYsXLiwUd0kMpmM9evXs2LFCnx8fDhx4gQHDhxgwIABdOvWjZdeeqnBa0pItCSSsyEh8QzT\nq1cvBEGgrKwMD486Hx6qoVKp+Oc//8nUqVNbwLraKS0tZe3atSxatIhTp05x9OhRdbTj5MmTaGtr\nM3LkyHrJbu/du5dRo0Y9Mw7Hyy+/THJyMtbW1o1yFK5evcobb7zBrl27SE1NrVEUrT4IgsB3333H\nypUrGTJkCIGBgWzevJmpU6dy6dKlZh3oJyHRVCRnQ0LiGUYQhEaNZ28tlEolBQUFGBkZMXXqVCIj\nI5k+fbraqaja3hCmTJkC8Mw4HJqamk0qMu3Rowfr16/HyMiItm3bNmoNlUpFaWkpurq6fPXVV+jo\n6DBo0CCCgoJYvHgxU6dOJSQkRJqhIvHMIDkbEhKtTFxcHJcvX6Znz57NPvX1aVJeXs6mTZsoLCyk\nc+fOTJo0qdlGtz9rDkdtiKLIgQMHiIqKwsvLi+HDh1fbZ9SoUXh7e6Ojo9PouS0nTpzA19cXR0dH\nevbsyeeff462tjaDBg3i/PnzhIWFsWLFCn766aemXpKERLMgdaNISLQiaWlpbN++nZs3b7Jz504y\nMzNb26RGk5mZiUKh4I033iAyMrJBx+bm5pKenv7EfaZMmcKPP/7IqFGjCA8Pr3PN8vLyp9Je+zA5\nOTnExsayfPlyzp07R0VFRY376evrN2lA3OjRo9myZQvJyckkJiYycOBAfHx8GDlyJO+99x7btm3j\njz/+YP369Y0+h4REcyI5GxISrUhCQgJQ2XEhiiKFhYWtbFHjsbKyQl9fn7Vr1zaoviQ5OZkNGzbw\n22+/cf78+SfuW1+HIzw8nNWrV/Pdd9+1yFyU2jA0NERHR4dt27bRvn37FlX2nDdvHqGhoaSnp+Pk\n5MSYMWMYPHgwISEhnDt3jkOHDvHJJ5/w+++/t5gNEhL1RUqjSEi0It27dyckJISYmBiMjY3VapzP\nI3K5nAULFlBRUdGgWoHk5GScnZ2xtbXl1q1bdXZS1JVSKSgo4MyZM4wZM4aEhAT1mPungVwu57XX\nXiMzMxNra+sWn5PTu3dvzpw5w+DBgxkyZAhLly5lxIgRvP7660RERHDixAmGDx+OmZnZcyfRLvFi\nIUU2JCSeQE5ODnFxcahUqgYdV1ZWVq8Qvp6eHu+++y5vvPEGb731VoNvTkqlkqSkpHpHRBQKBbt2\n7eLbb7/l2rVrDTrXk7hz5w4BAQGkpKQ0uCjRzc2NO3fuEBwcXM3RSElJ4caNG9U+y9oiHAqFgk2b\nNiGTyTh06BBxcXE4ODg0/sIagba2Nra2tk9tXomtrS3BwcFcvnyZKVOmcOrUKVJSUti6dStubm6s\nW7eOjz/++KnYIiFRG0/N2RAE4WdBEDIEQbhRw3vvCoIgCoJg/tC2jwRBiBUE4ZYgCCMf2t5bEISI\nB+99Lzz46ywIgrYgCL8/2H5BEAT7h46ZJwjC7Qdf81r2SiVeFLKysti4cSMBAQEcPny43sddu3aN\nb775hrVr15KTk1Pn/tra2lhaWjbqKXjv3r0cPXqU9evX18vhSExMJCcnh4kTJxIcHNzg89VERUUF\nv/32GxoaGvj5+VFaWtqg442MjFi6dCnvv//+I45BYmIiv//+OxcuXMDf37/acTU5HOXl5ZSUlODr\n64tcLmfkyJHcv3//hZfytrOzIzg4mGPHjjF37lymTp2qnmkzYcIEMjIyuHDhQitbKfFX5mlGNrYC\nox7fKAhCe2AEkPzQti7ATKDrg2PWCYJQVU31E/Aa0PnBV9WarwI5oig6Av8Gvn6wlinwT6AP4AX8\nUxAEk2a+NokXkIyMDMzNzRkyZEid4lMPc/XqVSZNmoSDgwMxMTEtaGFlCmLcuHHo6urWq7jUwsKC\n/Px8AgICsLW1bRYbRFFEpVKhp6eHKIrVbuypqamsW7eOLVu2NEhyPSsrCysrK7p06UJWVlaN+zzs\ncBw8eJAffvgBHR0djh49iqurK8HBwYSEhHD69Ol6XceZM2c4cuQIeXl56u1FRUXs2bOHPXv2UFRU\nVG/7W4KcnBz27dtHUFBQtWibvb09wcHBbN++nSFDhtCuXTugUuX1b3/7G/PmzSM1NbU1zJaQeHrO\nhiiKp4DsGt76N/AB8PBfqAnATlEUy0RRTABiAS9BEGwAQ1EUz4uVf9G2ARMfOuaXB6/3AMMeRD1G\nAidFUcwWRTEHOEkNTo+ExOM4Ojqio6PD8ePHGTJkSL2P69atGwcPHiQ2NrbFh34NGjQIPz8/TExM\nsLOzq3N/IyMjFi9ezPDhw6tNc20sWlpaTJs2jXv37jFp0iR0dHQeef/MmTP06NEDY2Njrl69Wu91\n3dzc0NHRISoqil69enH27FmysrKIjo7m+vXr6pttlcNx6dIlDA0NKS4uxsXFBVNTU+zs7HBycuL+\n/ft1ni8yMpLIyEgEQeDYsWOP2C+Xy5HL5fVyWlqSY8eOoa+vT0JCAhEREdXed3BwIDg4mK+//pqv\nvvpK7fgtWrSI+fPnM3jwYFJSUp622RISrVsgKgjCBOCuKIrXHgshtwMeLku/82BbxYPXj2+vOiYF\nQBRFhSAIeYDZw9trOOZxexYDi4F6/eGWeLHR1NRk1qxZDT7Ow8MDFxcXtLS0WlxUydPTE09PzwYd\nY2Rk1CiZ7Sfh6OhYq6aGjY0NYWFhlJWV0a1bt3qvmZubi56eHk5OTvj7++Ps7MypU6fQ19dHS0uL\nnJwcBg0aBFQ6HNHR0cTHx6Onp4eJiQlOTk4cOHCAvLw8xo0bV+f5NDQ0UCgUlJeXq2eWxMfHk5SU\nRElJCTo6OurBc61FmzZtyM3NpaSkpFb10k6dOhEaGsrIkSPJyMhgzZo1yGQyVq5ciZaWFoMHDyYo\nKOi5LkaWeP5otQJRQRB0gb8B/2gtGx5HFMUNoih6iKLo8bQnUv5VebzwT6FQcOjQIbZt28a9e/ca\nvW5OTg5JSUmtlquvuiFKgLe3N2PHjmXOnDl07ty53sf5+fmhqanJ8ePHUSqVODo6UlFRgaWlJTY2\nNo+kOgCWLl2KtbU1586dQxAE9PT0mDNnDgsWLMDc3LyWs/wfLi4u9OnTBwMDA8aOHasupnVzc6Ow\nsBB3d3f69u3b4OtvTnx9fbGxsWHQoEE4OzvXul/btm05deoUFy9eZN68eWq9j3feeYe3336bPn36\n8N1331FeXv60TJf4i9Oa3SidgI7ANUEQEgFbIEwQBGvgLvDwmEfbB9vuPnj9+HYePkYQBDlgBGQ9\nYS2JVubQoUN8/vnn7Nq1S+0UREREkJ2djYODAwEBAY1aNz09nY0bN3L48GFOnDjRnCZLNAJBEOjY\nsWODZ3WIoohcLkcmk9GvXz/CwsIYPXo0ycnJhIWFUVhYiCiKZGdnU1hYiImJCe+88w6vvfZavYW/\nHrfTw8MDHx8f9PT0EAQBmUxGSUkJcrkcNze3Rk9pbS7atGnDgAED6N69e50FxSYmJgQEBJCbm8uo\nUaPIyMgAYNmyZQQGBhIQEICLiwtHjhx5GqZL/MVptd8cURQjRFG0FEXRXhRFeyrTG71EUUwDDgEz\nH3SYdKSyEPSiKIr3gHxBEF56UI/xCnDwwZKHgKpOk6lA8IO6jhPACEEQTB4Uho54sE2iBcnKynpi\nMWBZWRnXrl3jww8/JDExUS28ZGxsTHZ2NnFxcRgbGzfq3Pfu3aNt27ZSfvo5oqioiL1797J37151\nEebMmTMpKChg5MiRDBo0iJdffpmOHTuiUqn46KOPiIuL4+zZs2zevJl169apix8bqjRaGxoaGsyZ\nM4eKigpmzJhRrRblaZKWlsaPP/7I+vXr69XhVIWuri4HDhzgpZdeonfv3pw9exaorIc5evQomzZt\nYsmSJXzzzTcvfMeOROvyNFtf/YBzgLMgCHcEQXi1tn1FUYwEdgFRwHFgqSiKVfH2N4FNVBaNxgFV\nPXGbATNBEGKBd4CVD9bKBj4DLj34+vTBNokW4vLly/z888/89NNPtd7stbS06NChAz/++CPGxsYY\nGhoC0LFjRyZNmoSbmxtjxoxp1PldXFxQKpUEBAQwcODARl+HxKOoVCpOnjyJn58faWlpzbr2mTNn\n0NDQQENDgzNnzgBgbW3NuHHjcHNzAyprOH755RdKSkpYvXo1nTt3Jj4+npEjR+Ls7ExiYqJ6vaY6\nHFeuXOHIkSMYGxszcuTIVq9vOH/+PN26daN9+/Zcvny5QcdqaGjwxRdf8NNPPzFp0iTWrl2rdiyG\nDh3K2bNn2bJlC2+99dZTl3eXeDoUFhYSExNDWVlZq9kgSN5szXh4eIgN/aWWqMTPzw8XFxfu3r2L\noaFhrTd8lUpFVlYWJiYmT00A6UVGqVRWm7cxdepUZs6c2Swj5qOjozl16hRdunQhJiaGV1+t9Xmh\nwZw9e5YbNyoleLp160a/fv2q7XPx4kXu3LmDi4sLoaGhjBkzhuPHj3Pv3j0MDAyYN29etWjY3r17\nWbp0aYOGt504cUItm66pqclHH33U4kqgdREUFMTFixfR0NBg3LhxuLq6NmqdhIQEpk6dioODA5s3\nb1Y7+bm5uUyZMgV9fX127NiBnp5ec5ov0YqUlZWxbt06DA0NUalULFq0qFl/ngVBuCKKYp3zCSQF\nUYlmx9PTk+PHjxMbG/vE7gOZTIaFhYXkaDQRURTZuXMnn3/+eYPExxqKnp4eRUVFpKamNvvN6KWX\nXsLLywsvL69a5codHByIi4vj8OHDdO/enYMHD+Li4oKhoSFjx45FqVRWExSbMmUK33//PbNnz653\n221UVBQABgYGVFRU1DpM7WFEUeTs2bMcPXqU/Px89bbw8HDCwsIarED7MEqlkitXruDq6opSqaRD\nhw6NXqtjx46EhoZiamqKp6enun3W2NgYf39/jI2NGTx4cLNHriRaj4KCAlQqFb6+vty7d6/VXtKU\ncwAAIABJREFU0mWSsyHR7Dg6OrJy5UpWrFiBqalpa5vzTFJYWMi1a9eqdVQ0hpycHFJSUvjwww+5\nevUq9+/fb5HpsXZ2dowbNw47OzsmTpxY4z5lZWXs2LGDH374gfj4+Hqtm5GRQWJiIj169MDd3b3W\nIkxzc3OWL19Oz549SUxMRENDg4yMDMrLy4mIiGDLli38+OOP1QS57t69y5w5c/jqq6/qlVLp1asX\nUPlH2sTEpF5dRVFRUVy/fh1RFNUFlxcuXODChQuEhYURGhpan4+iRkRRRKlUYm1tjUqlYseOHU2q\nRWnTpg3r169n1apVDB06lF9//RWoTG1u3boVX19f+vbt2ySbJZ4dzMzMcHV1xc/PDx8fn1YrcpYe\nKSVahNYOOz/LqFQqfv75Z8zNzQkMDGT58uW13tBUKhUVFRUoFIpa/y0rK0NTU5O1a9diYGDA+vXr\nkclktG3bttmfUDt37vzE9tWoqCiUSiX9+/cnJCSkzrkkSUlJ7Nq1C11dXVxcXBg2bNgT909NTSUm\nJoaePXuSmZmJmZkZ3bp1IyAggOnTp/Pnn3+SkpKi1hG5d+8eenp6zJo1i82bN9c6vO1hBg0ahJ2d\nHYWFhfXW1ZDJZKhUKpRKpfqPeWFhIebm5sjl8iZN85XL5cyYMUNdqzFgwAD27NlDly5dmtRe/cor\nr9CzZ0+mTJnCH3/8wXfffYe+vj7/+te/cHd3Z9q0acyePZvPPvusVYtjJZqGIAiMGTOm0TVwzYXk\nbEhIPEZpaSmFhYWYmZm1iNOkUCjIy8tj7969eHt70717d/Ly8qo5EhUVFYiiiKamJnK5/In/amtr\nY2hoSLt27dDX16egoIC0tDSMjIwand9vDFZWVpw8eZKioiK1MN7NmzeJiYlhwIABmJiYqIfHtWvX\njrS0NOzs7HBwcODmzZt1rq+jo0NpaSnp6eno6+ur64GqikfNzMweUW21s7NDLpezYcMGfHx8cHFx\nqZfD0bFjxwZdt4uLC4WFheTk5Kgdpn79+nHixAkUCkWdTlRdODg4YGtry/fff8+NGzfQ0dGpVp/z\nOBUVFURFRWFmZlarNL2bmxtXrlzhrbfeolevXvj5+dG7d28mTpyIt7c3y5Yto2fPnmzdurXOabwS\nEk9CKhCtBalA9K9Jfn4+GzduBCoLFUeOHFnHEY1j06ZN3Lp1C29vb9zd3dHU1KzRiZDJZCgUCkJD\nQxFFEW9vbzQ1NWtdNzc3l+3bt1NUVMTUqVNbXC69JjIzM8nNzaVTp06kpqayefNmdHR0UCqVrFy5\nkm+++YaioiI0NTV58803OXjwIAUFBYwbN67GeoS8vDyys7Pp0KEDMpmM27dvk5aWRs+ePdHX11fv\nVxVVqMlBFEVRvb0xRaOPk5SURHBwMFZWVowaNeqphaazsrKIi4vDycmpztbwvXv3UlBQQGZmJrNn\nz1bPSqmNnTt3smzZMkaOHMlrr73GoEGDEASBPXv2sGzZMubNm8cnn3xCmzZtmvOSJJ5zpAJRCYlG\ncPfuXczMzJgxY0aLDlEzNzfn9u3bTJgwgQ4dOtC2bVssLCwwMTFBX1+fNm3aoKGhgSAIhIaGkpKS\nQmpqKqdOnQIqbzr79++v1gZpbGzMsmXL+PDDD1vF0YDKYW+dO3dGJpORkZGBTCajZ8+elJeXU1pa\nSlFREStWrKCiooLMzEzmzZvHsmXLanQ0cnJy2LBhA8ePH+fAgQNcvHiR4uJivL29H3E0APXnVRMP\nb28OHY5jx47h5uZGSkoKt2/fbtQajcHMzAwvL696adBkZ2fTtWtXTE1N66XNMXPmTFasWMGNGzd4\n/fXX+eOPP4DKjqbr168TFxdHr169CAwMbOplSPwFkZwNCYmH6NixI2VlZWzZsoU+ffq0tjlA5VN5\n1RN7VSRy06ZNREZGcvTo0SYVCzYHCoWCixcvcuPGjWqV7j169MDKyopLly7Rr18/2rRpg5mZGd99\n9x26urp1pivS09MxNTXFx8eH27dvExUVxdmzZxs00K0mmupwmJiYcPPmTfLz85t9zkxzMXLkSK5f\nv465uXm9U2m2trasWrWKzp07M2XKFD7++GOKi4uxtLRk9+7dfPHFF7z++utMmDCB2NjYFr4CiReJ\nBqVRBEH4uZa3RKCUSqGt30VRfO7nGEtplL8uVWPSWzI0fuDAAbZu3cqBAweeuF9JSQmlpaWEhYUh\niiIDBw5EEAS+/PJLjIyMyMvLw8PDA19f3xaztS78/f3JzMykoKBALaVdF0VFRejo6NT5Gefm5rJ+\n/XrKysqwtLTE1taWgoICbG1tGTBgQJNtryulolQquXXrFtbW1o90VpWVlREZGam26UWhsLCQ8PBw\nbGxs0NbW5r333uP8+fN8++23TJo0CUEQKC0t5bvvvmP16tUsWLCAjz/++Jl1uCRanvqmURpaIGoB\nDABUwI0H27oBAnAFmAx8KgjCAFEUW/dxS0KikQiC8Ex00yQnJ7Nt2zZUKhWdO3fG19cXLS0tSkpK\nkMlkalnvPn36UFxcjCiK6qml+vr6T62DoKioCEtLy0dsepiysjJ+/fVX0tLSEAQBXV1dRo8eTUFB\nAUVFRVy7dg1TU1NGjBiBqamp+rMXRZHIyEjs7e3p0KEDUVFRFBcXo6OjQ/fu3R+pw2gsU6ZMAai1\naHTDhg3qNuJFixbRtm1bALS1tdUtsi8S+vr6eHt7q7/fuXMnISEhLFu2jP/+97/85z//wdnZmQ8/\n/JB58+axatUqXFxc+PTTT1mwYIGkmSNRKw39yfgTKAReFUWxGNTTWzcC14ExwDbgG6Bp5dcSEi84\nJSUlnD59Wh1JAR55ffPmTZRKJXK5nJiYGOLi4ujSpYv6pq1UKrG3tycwMJDY2FhUKpX6WA0NDZyc\nnNTFfA+v6+zs3KwdKj4+Ppw8eRJjY2N69+5d7f3du3erZetVKhV3794lJycHURTVHRU5OTmsW7eO\n4uJiDh48qNbosLOzY+rUqYSGhhIVFcXp06cRBIGKigpGjx7NokWLgMpujfpEVGriSQ5HZmYmvr6+\nBAQEcPPmTbWzAZUS4vHx8RgbG5Oamkr37t3x8vJqlA2Pk52dzfnz52nbtm2ji1ibiyFDhhAeHs4P\nP/xA//79WbRoER9//DHW1tZs3ryZK1eu8P777/PVV1/x0Ucf8corr0gTjyWq0dA0yl3ARxTF6Me2\ndwGCRFG0EQShJxAoiqJZ85r6dJHSKBItyYEDB3jjjTcwMjLCwsJC/YReFVURBEGtIVFFVlYW+fn5\n2NnZce3aNVxcXEhISMDc3JyKigqsra0BKC8vR6VSkZ2dTWZm5iORmoyMDExMTDh27NhTu9Yvv/yS\n8vLyR6IQGhoaGBoa1li4KAgCnp6eDB8+HLlcTmZmJkVFRXTo0AGFQsH9+/cJDw/n+++/R0tLi3v3\n7lFSUqJWw2wsNaVUdu3aRXR0NFpaWixdulQt752ens5vv/1G//79OX78OHPmzGH37t2sWLECXV3d\nJtkBsHHjRuzs7IiOjmby5MnqNuKnTXl5OcePH6ekpISRI0dSUlLCBx98wB9//MGaNWuYPn26+v/1\n9OnTfPrpp+Tl5XHx4sVWsVfi6dNSaRRDwAaIfmy7NVBVGp7fiHUlJP5yaGpq8sEHH7Bw4cJa9wkN\nDeXy5csUFhbi6urK3LlzuXLlCnK5HLlczooVK+jQoQM7duxQOxmampoYGBjw97//vVou3d/fn+++\n+66lL+0RFAqF+rVcLsfCwoJ79+7V2iFRlT4xMjJSC4jZ29ujVCrZvHkzZWVlGBoacuzYMQRBICIi\ngu7du7No0SKWLVvW6EhATRGO6dOno1AoqnW6aGtro1AoSE1NRSaTERYWhpaWVrM+0T8Lqbyqnz1T\nU1OCgoKYMmUKv/76K6dPn2bZsmWsX7+e77//nm7dujFgwABOnDiBsbExWVlZmJk918+bEs1MQyvg\n9gObBUGYJgiC/YOvaVROXN33YB8voOV6BiUkXhC0tbXrlBXv378/K1asYNWqVbz55psYGhrSu3dv\ntLW1eeWVV/jjjz+wtLRk4cKFaGtrq58033rrrScW7V28eJEvvviCX375pV6zP5oLpVJZLyn1oqIi\nTp48yU8//cTPP/9MZGQk+fn55OXl8dprr3Hnzh3Ky8uByrTQli1biI6O5sSJE02yr6YuFblcXu3G\nb2xszKxZs7CwsGD+/Pl06tSJV199tdlqFiZPnkxFRYVazbS10NfXJzs7m7S0tEdajQcMGMCVK1eY\nPHkyQ4cOZcaMGdy4cQOZTMaYMWNYvXp1q9ks8WzS0N+MJcC3wPaHjlUAPwPvPfg+GnitWayTkHiB\nqY+zURMFBQUABAQEYGRkhKamJvr6+uqnfGNj41rVJavaZ0NDQ5k3bx7+/v4kJiY+UYK8qcjlcoqL\ni9HU1FTXmtQXURQpLS1lz5496m2rV6/Gw8MDbW1toHKmx/z584mOjm6WIVN1FY1W0b59e9q3b69+\n3ZyYmZnV2WF0//59Tp48iYGBAaNGjWqR4kw3Nzc0NDQoKSmp9jnI5XKWLVvG/Pnz+emnnxgwYAAR\nERH85z//wd3dnZEjRzJkyJBmt0ni+aRBP50PikKXCILwLlClGBQnimLRQ/tIXSgSEvVAS0uLjIyM\nBh0THh7O4cOHadOmDenp6SxdupQtW7aQmpqKo6MjHTt2ZPTo0U8UtxJFEQcHB/bu3YtCoVDXetRG\nRUUFfn5+pKSkMGzYsAbLVnt6evLnn38CYGRkRElJCRUVFejo6NR7ZoiGhgaampp06tSJtLQ0DAwM\nGD9+vHrKq0KhIDw8nL/97W8Nsq026utwtCYBAQFYWVlx584dwsPD8fCoM23eYARBqHM+jL6+Pu+/\n/z6JiYls2LCBTz/9lJ9//plXXnmF8PBwKZ0iATSytuKBc3G9mW2RkHjuCA0N5fz581hZWTFz5kz1\n02VVZ8iT5lc0NLIRGxtLYGAgKpUKbW1tiouLWb9+PSUlJQDExMRw7949lEolXbp0IS8vj9LSUhwd\nHR8pQBVFkfHjx3Pv3j2MjY3rLGhMTk6mpKSEhQsXsmPHjgY7Gz4+PixevJhPPvmEpKQkDAwMEEWR\nsrIypk+fTmRkJJGRkbRp0wY3NzciIyNRKBTIZDK1M6FUKlEqlURFReHs7ExISAg2NjaMHz9eLd8u\nCILaNpVKRXJyMiYmJjWmk5RKJf7+/mRkZDB06FDs7e2r7fOsOxwGBgakpqaSm5uLgYFBa5vD0qVL\nGTZsGB9//DEjR45k6tSpLF68mD179jwT9ScSrUuDnA1BENoAK6hsa7XksZoPURQb13smIdFMJCcn\nEx8fj5ubW4s/URUUFBAYGIhMJiMuLo5du3Yxe/Zs0tLS+PXXX1EoFMyYMaPWyadaWloNcjaio6Mp\nLi4GKueF6OjoqB2Nh22Kj4/n9OnTyGQydHR01EWMp06dwsDAQN2d8nAb55OwsrKiqKiIXbt2NUoC\nXRAECgsLOXnyJF999RXBwcGUlpbi4+NDx44dcXV1xcvLi/3795OUlIRcLmfKlCk4ODigVCq5f/8+\nWVlZ7NmzB1EU1QPbrKysKCwsZMaMGahUKo4cOcLWrVvp3bs3WVlZJCcnU1RUxKuvvoqZmRlXr17F\n398fbW1thgwZQmpqKr179+b48eMsWbKkRtufZYejqq7EwMAAZ2fnRq2Rl5dHUVERNjY2TXYIunTp\nQp8+ffjkk0/44osv+Oqrr+jfvz9r167l7bffbtLaEs8/DS0QXQesBBKBA8Dex74kJFqNgoIC/Pz8\nKC4uZtu2bc2Sv38SKpUKQRDU0Yv8/HwAIiIicHd3Z+jQoU+UwtbW1m5QGiUhIYGuXbuira2Ns7Nz\nrTd+CwsLiouLKS4upm3btmRkZHD69Glef/11SkpK0NPTa8BVVobJ33jjDaZNm8b48eMbdGwVVRGL\nvLw8ysrK6NWrFwEBAahUKsrKyvDz86O8vJyMjAwqKirUNRAaGhpYWVmp9UgeJy8vj+vXr7N582au\nXr1KcXExISEhJCUl4ePjg66uLrt27SIsLIyAgABsbW0pLCzk6NGj5OTkEBkZiYmJyRNtb45ZKi2B\npqYmnp6euLi4PHG/5ORkNm7cyL59+x4pBk5LS2P9+vXs2rWLkJCQBp1bFEUKCwtRqVSPbN+wYQNb\nt24lJCQEbW1t9u7dy9dff83p06cbtL7Ei0dD0ygTgWmiKEqTeCSeOSoqKpDJZFhZWXH9estn+YyM\njBg1ahQhISHo6OioC/qcnJzYuXMnAGPHjq31+Ko0Sn2VMB0cHIiPj0culzNixAjWr1+Pqakp2dnZ\n6n1kMhnh4eG4urpy7949YmNjefnll1Eqlfj5+SGKYo0qn3Who6PTJEXSK1eu8Mcff2BiYkJOTg4R\nERFoaGjw5ZdfYmhoiFKppKysDE1NTbWNqampxMfH07t37xojQIIgUFBQwIEDB9TFomVlZchkMvr1\n68fhw4cpLi6mQ4cOHDlyBCMjIxISEoBKR7G8vBxra2sGDx6MQqHg+vXr6Ovr4+TkVO1cz3KEoy5O\nnjyJu7s7kZGRREVF0aNHDwDu3LmDnZ0dbm5uXLhwoUFrHj16lOvXr2NmZsbChQvVqSxLS0u2bt3K\n9OnT+frrr1mwYAFbt25l5syZXL58GRsbm1rXFEWRPXv2EB0dTadOnZgzZ07jL1rimaOhkY1iIKUl\nDJGQaCqmpqYMGzaMuLg4pk2b9lTyxF5eXnz44Ye8/fbb6qfxDh06sGzZMt58880nFtdVaTfU9+bv\n6+vLlClTWLJkCcbGxri4uFRLo6hUKry9vYmPj6esrIxZs2ZhZWXFtGnTmDx5Mj179mxQN0hz4uDg\ngI2NDXPmzKFXr14olUp1+2ivXr2wsrJCLpfTo0cPFAoFR48excDAAD8/vxpTUYIg4OPjA/yflkd5\neTkXLlygT58+rFy5ElEUyczMRCaTYWJionZKoPKzOnfuHGVlZQQEBHD9+nWOHTvGrVu3arT/WY1w\n1IW5uTkRERFkZmY+klp0dXWloKCAo0eP0q9fvwatGR4ezvLlyykvLyc9Pf2R94YPH05ISAhr167l\n73//O6NGjWLx4sXMmzfvidHGgoICoqOj6d69O7GxsSQmJjbIJolnm4ZGNv4XeEcQhCViS8eoJSQa\nQa9evZ6JmRX1TVVYWlqSkZFRbVx6TQiCQLt27SgsLGTLli1kZWVhamrK3bt3H9nP3Nyc999/H0EQ\n1IPONDQ0aNeuHTdv3mzx9FJtVJ23bdu2tG3blpycHPbtq5TnGTlyJD169ODYsWPk5+ernQelUokg\nCMyePRt/f/9qypQHDx58RIpdJpPh5eXFZ599RnBwMOnp6eoalezsbObMmcOvv/5KRUUFnp6eXLp0\niYKCAvLy8rCzs0Mul5OXlwdASkoK5eXlODg4qB3XlopwVKXkWsJB9vX15ebNm8jlco4fP45KpWLy\n5MmYm5vz2muNUyno2bMn//nPfzAwMCApKQlzc3O1ND5At27dCAoKwtPTkx49erBq1Sp69uzJnj17\nmDZtWo1r6unpIZPJiImJQUND44nF1RKV3Lhxg/v37+Pp6dng9OjTpqGRjeHADCBREAR/QRAOPfzV\nAvZJSLzQWFhYNFhrY8OGDerix8cdDZlMhoODAxoaGjVOVM3KyuLSpUtNsvlhysvL6+W8VHXBPMyg\nQYN49dVXWbZsGaampgQHB+Ps7IxSqSQpKYmxY8dSWlrKrFmzAGosaK1aU6lUIooi5ubmzJkzR50W\nUCqV6OjoMHjwYPLz8zl16hS9e/dGR0eHsLAwnJ2dadu2LcOHD+fevXsYGBjg7u5OdHQ0u3fv5vjx\n44SGhgKVNTPnzp1j9OjRjY5wlJeXs337dr755huioqLU6/7v//4va9asIS0trUHr1Qe5XE63bt1I\nSkqiXbt2ODg4cO7cuSat6evry9SpUykvLyclJYWDBw9W28fCwoL9+/fz5ptvcvPmTX788Ufefffd\nWtudNTQ0mD9/Pra2tgwfPrzZtUteNOLj4wkMDCQ7O7vGz/9Zo6GRjftUqohKSEg0Aw1xNkpLS/nt\nt9/Uol41MXPmzCc+EVpaWjbbH/HQ0FBCQkIwMTFh4cKFT6zpqO2J/eGx7ebm5kRHR5Ofn4+Xlxd2\ndnakpaURHR2NmZkZnTt3Rk9P75G008MOTJX41L59++jduzdXr15FpVJhY2NDcHAw/fv3Jy8vT93J\nM378eHVxZZWTAlBcXExKSgoODg6YmJiQlpZGWloae/bswd7enpiYGObNmwc0PMJx+/ZtFAoFY8eO\nJSgoiC5dunD16lWGDBlCfn4+ERERdeqeNBRRFElISEBDQ4OoqChkMhn9+/dv8roKhQIDAwMsLCyI\njIykoKCgWgtuz549+eSTT3j77bc5efIkAwcO5IsvvuB//ud/alzT1taW2bNnN9m2vwLl5eVoa2tj\naGiojsY9yzRU1GtBSxkiIfFXQ0NDA1dXVyIiIvDx8XkkDF0TMTExT5y9IZfL66UEamFh0WBba+LK\nlSssXLiQgIAAEhMT65wkW1cEZPjw4URHR2NgYICdnR0hISHqyM3p06cZOnQoK1asICgoiIsXL1Zb\nTxAE2rdvz8iRIzE1NeXq1asAzJ07F5VKhUwmo6ioiF9//ZXy8vIaa1eSkpLYuXMnKpUKY2NjMjIy\nmDBhAkVFRejq6mJnZ0fVgMbGpFSsrKzIyMggMDBQLUPu7OzMkSNHAJg+fXqdazSUsLAwQkND1QJd\nzs7OdOjQocnruri4kJKSop5Ou2nTJpYvX15NyfS1117j22+/JTg4mNWrV+Pm5saCBQtqLMSVqD/O\nzs5kZGRw//59xo0b19rm1Ik0ME1CopWwsrLCyMiIa9euERAQ8MS20qKiIhISEkhOTq51n4cHntVG\nVV3A4x0w9e2IeZiuXbvy66+/oqWlVWe0pKY0yuPIZDK0tLTYs2cPJiYmmJmZkZSUhEqlIi4ujtOn\nT6uVRGUymbrOocr+7OxslEolvr6+1a6lKqWUmpqKIAj4+vpy4sSJagW8t27donfv3hgYGHD37l0m\nT56sXt/V1ZXY2NhH/p8a6nCYm5uzePFicnNz1c5G165dad++PTKZrF61Ow0lIyNDLeymUCiaxdGA\nys904MCBXLlyBScnJ4KCgrh69Sqenp6P7Kepqclnn33GRx99xPnz53nvvfdYtWoVu3fvbhY7/qoI\ngsDAgQNb24x6U6ezIQjCdWCQKIo5giBEALX+xZBEvSQk6k/V07aenl6dUY1jx44himI1XYOHqU9r\n6v379xk4cCD//ve/mT9/PlpaWmzbto3c3FwmTpxIly5d6m3/sGHD8PDwQFdXV936WBuPOxsqlYqA\ngAAyMjIYOHCgWsHz9OnT+Pj4qAevWVhYkJOTU01JVCaTIYqi+l9BEDAyMiItLY1OnTrh6elJYmJi\ntfkiZmZm5OXlcebMmRrTFW5ubvz6668olUpmzJjxiP1Dhw6t8doa6nAYGxtjbGz8yLaq0fUtQd++\nfTlw4ACiKDJx4sRmXVtHR4eBAwcSFBSEi4sLwcHBODk5VVNtnTFjBl9//TXHjh3jrbfewsnJiUuX\nLlVzTCReXOoT2dgLlD14vedJO0pISNSf9PR0XF1d0dXVZdiwYXXuX1hYiIaGBjY2NmpV0Kpujaob\nblFR0ROr0lNTU0lKSsLJyYnIyEh0dHQwNzdnyJAhnDt3rkZn40lRjydNln0Sv//+OzExMRgYGHDw\n4EFWrFgBVA40+/PPPykpKcHHx4eTJ0+qHY3HbSovL8fQ0FAtVOXp6cno0aNZuXIlkZGRQGXNwMOY\nmpry6quvkpmZiaOjY7V1bWxsePfddxFFsUGDzZ5lHQ5jY2Pmz5/fYuv369eP8+fPqx2mmj43mUzG\n8uXL2bRpE76+vvzjH/9g5cqVBAYGtmiLemMidhItQ52/TaIoflLTawkJiaYhiiKdOnXixIkT1Yo6\nIyIiOHz4MGZmZrz88suMHj2aw4cPU1FRwd27dzE3N2fUqFFs375dHTEoLi5m69atLF26lJycHNat\nW4dCocDS0lKdr9fR0aFjx47Ex8djbGyMs7MzwcHB3Llzp9ogryol1qysLMaNG0f37o0PXD4e2YiP\nj0dHR4eysrJHFDx9fHxwcnLCwMAAXV1drly5QmpqarX15HI5MpkMDw8P5HI5ubm5jB49GqjUOXlS\nqsDMzOyJUvaNbblsbYdDqVQSHR2NiYkJ7dq1e2rnlclkzJs3jxs3bjB79uxand3p06fz3nvvkZGR\nwcKFC/nmm2/w9/dnzJgxLWLXrVu32LdvHwYGBrzyyistGj2SqBupZkNCohWxtLTkwoULTJ069RG9\niM6dO3Pv3j3Mzc1ZunQp2dnZGBoa0rZtWxwdHVEqlTXKlefk5ACwZcsWdQ1HRkYGOjo66tRBZmYm\nkydPJisrC0tLS958802Ki4urFY7GxMRgaGiIj48PwcHBTXI2HqdXr15cuXIFuVzO3//+d9544w2O\nHz9Oeno6d+/eZdiwYcTHx5OTk1PNUamq2ygvL+fmzZuoVKpHUh5VlJWVkZubq3a2ngat6XAcPXqU\nzMxMcnJymDFjxlNtHbWwsKhznLyBgQETJkxg+/btvPPOO/z000/MmTOH8+fPN1sdycNcvHiRsWPH\nEhsbS2RkJH379m32c0jUn4YOYjMFvqD2QWyS6ygh0QB69erF5s2bUSgUj0xmTUtLw8jICJVKhZeX\nl7oeIyQkhDt37uDk5MS6devUhZJVuLm5oVKpqrXHVimN2tjYIJfLuX37NhMmTAAqZ5/o6emRn5+P\ngYGBupiyffv2nDx5kkOHDqklrhuLIAhs3LgRPT09tVN18eJF0tLScHV1RV9fnx9//BFPT080NTXx\n9/d/4npBQUF07dqV5cuX1/h+cXExGzZsQBRFOnbs2Oy1Ck+itRyO+/fv4+bmRnR0NFlqB2owAAAg\nAElEQVRZWc+kTsXChQtZsmQJS5cuZejQobz//vtMmjSJM2fO1Dl9uKF06tSJgIAAlEolffr0ada1\nJRpOQyMbm4GewAYglScUi0pISNSNXC6v8UYoiiLp6eno6+s/0qEQGRlJUVERxcXFDBw4kL1796q7\nJDw8PBg+fHiNLZ1yuRyVSsXNmzcxMjJCLperIxmiKLJ9+3bu3r1L27ZtefnllxEEATMzM958800K\nCwuxtLRs0nV++OGHJCcnP+JQWVlZMXv2bHr06EFqairHjx9HoVCg+P/s3Xdc1ef5+P/Xfc7hsFUE\nFRBUVAQcuHDvuI1GExOj0UaTNJqO2AzTz7dN+8mvadN80jSzTVLTxJFRZ5ZNapxx742oKCIgiiKy\n51nv3x/AqcgI48A56PXs4zwK73G/r6OGc3GP67ZY7JNnb52PUlJSYr+/f//+Na7euX79Oh4eHsyc\nOZPly5c3KPb6cEbCMX78eL7//ntat25dY5l8ZxoxYgQ9evRgyJAh/Otf/+LZZ5/l2LFjPPTQQ3z8\n8ccOrTEydOhQwsPD8fDwqFT/QzS9uiYbY4HxmqbVbdceIe4Q5V2yPXv2rNd267WllKryB69SCpvN\nRlZWFgcOHMDb25v4+HhatGhBcnIyf/nLXyosb4XSypvXr19nzJgxbNu2jd27d3Pfffdx8+ZN2rVr\nR2FhIcnJyXTp0oWLFy9SUFBgT3C8vb0dUgZ5yZIlQOnOuMuXLyc/P5/HH3/cPjQTFRXF2LFjKSws\nJCsri8DAQI4dO0ZSUhLu7u7ExMRUWUG0OiEhIXh4ePDBBx9Uu4qksc2cOZP09HR+//vfM3v2bPLz\n85k2bVqd3kddhIaG1rv8eFNRSrF27Vo+/PBDhg8fzquvvspHH33EH/7wB3r16sVLL73Ez372M4eV\nKndUTRnRcHUtV54OVF1rVog7XHFxMevXr6dNmzasW7cOk8nklDg0TWPBggX4+PjYq2lmZ2dz9epV\niouL6dChg301RfkKFaPRyI0bN1BKMWrUKEpKSuxDM0ajEZvNxoULF350eW1DxcfHExwcTHR0NDt3\n7qxUe8PLy4v27duj1+sZMGCAfVv7un5Au7m5sWDBAn7/+98zfPhwR76FOgkPD6d169a4ubnRs2dP\ndu7c6bRYXIVSikWLFrFr1y7ee+895s6dy7PPPsvOnTv54osvGDBgQJ13oRWur67JxovAy0opx1ee\nEcLFlXfh39qd39Rat27Na6+9RmBgIKdOnULTtErDJvHx8Xh4eGCxWNA0jVOnTvHOO+/w1ltvkZyc\nzLVr1wgLC7MvD9Xr9Xh5edG3b1+MRiNubm7k5eVx5MiROu/b8mPCwsI4f/48J06cwGQy2at8NhZn\nL3ssHxa6fPkyp06davBwlCNdv36dEydOVLm0uCl0796dgwcP0rVrV3r27Mn+/fvZtm0bzz33HDNm\nzODxxx8nOTnZKbEJx/vRZEMpFauUOlVW3OtFYAKQrpQ6W378lvNC3LHc3d2ZM2eOfev2mkqHm0wm\nYmNjK22/3VCLFi3CbDZjMpnw9fWt8sN01KhReHl52bdlDw0N5fz58+zatYtHHnmEyMhIEhIS7D0z\n5UsXfX19mTdvHh4eHqxcuZLExESWL19OYWGhw+IPCAggMjKS7t27ExwcXGGfk8ZWXFxsr8nRFHJy\ncjh27BgdO3akU6dOjB071mlDOrfLyspi5cqVnD59mtWrVzstDnd3d1577TU2b97M0qVLGT16NP37\n9+fs2bMEBQXRr18/Fi9e3Cgb1ImmVZuejfWUFvYqf/0VeA1YfdvxLxopRiFcRseOHZk0aZK91HR1\nvvjiC44dO8bKlSvJyMhwaAwGgwGDwcDDDz+Mj49PpaGI3bt3ExYWRnJysn3uRkJCAlCaiBQWFjJg\nwAAOHDhgX7XSpk0bRo8eba/PkJOTQ+/evdHpdA5NNgAmTJgAlC6FHDhwYKXzWVlZnDhxotrdQesj\nLi6Ot956i7feesv+wXXkyBHefvttNmzYUKuda+sqNTUVLy8vzp07R1paGt26dXN6T0u53Nxce29W\nZmams8OhT58+7N+/n4ceeogRI0bw1ltv8fvf/56zZ8+i1+vp0aMH/+///T+XiFXUz48mG5qm/aG2\nr6YIWIjm4MaNG/Tr148WLVqQnZ3dKM/o3LkzAwcO5OLFi/baE2PGjOHBBx8kISGBwMBA+6qO8uRI\np9PZk5Xy72+nlGL69Ons3buXQYMGERAQ4LCYi4uLcXNzY+bMmUydOhV3d/cK50tKSvj44485e/Ys\ny5cvd0gSkJKSwsaNGwkLC6NHjx7Ex8cDsHnzZmbOnMmlS5dIS0tr8HNu16VLFywWC7169WLw4MEO\nb78hOnToQHh4OPv27atU0t1Z9Ho9Tz/9NCdOnODUqVP07t2bM2fO8NZbb3HixAmysrLo1q0bL7/8\nMrm5uc4OV9RRneZsKKUSlVKVSu8ppVoppRIdF5YQzdvkyZPZt28f7du3tw9nNAa9Xo+Pjw83b94k\nIyODV199lenTp3P8+HFiY2Mxm80EBgbaq3TqdDrmzZuH2Wxm1qxZ1a406dmzJ48//jgjRoxwWKxJ\nSUm8/fbbvP322/aeltsVFRVhs9kYNGgQOTk5VS7jrauvvvqKrl27Eh8fz6lTp4iIiABKV+ls2rQJ\ns9lcaa8SR/Dw8KBPnz7s3bu3UnVWZ1NKMXHiRJ588kn7n4erCAkJ4auvvuK1117jJz/5CY8//jhe\nXl4sXbqUgwcPcuHCBcLDw/noo48apUdKNI66Ln3tBFS1JskdCGlwNELcIcLDw2u13XtDlG8o1qtX\nL/Ly8oiIiGDKlCmMGzcOq9XKxYsXad26NTqdjtTUVNq3b49SiuDg4EZbflmTs2fPMnjwYIxGI3Fx\ncVXuTdKqVSsGDx7Mli1bmDx5cp32J6nK1q1byc7OpqioCE9PT+bNm2dfUvzII4+QlJREYGCgwwtK\niYabMWMGY8eO5Xe/+x09evTg9ddfZ968eXz66aecPHmShQsX8tlnn/Hhhx/KdvXNQK3+S1ZKPXDL\nt/cqpXJu+V5Paf2NS44MTAhRvdjYWPbt24fVaqV379788MMPmEwme0lmvV5Pt27dSE1N5V//+hfu\n7u707duXoUOHEhsbS8uWLRvU42K1Wrl69SoBAQG12m0WSrdSX7VqFZqmMWvWrGqvGzlypEO2zs7O\nzubgwYP07NmT5ORkBg0aVCHJMhqNLvUhVVRUxLfffkuLFi2YMGGCy8zvcCZfX1/eeecd5s2bx5w5\nc3B3d2fWrFn07t2bffv28fe//52hQ4fy/PPPs2TJkh/dfVg4T21/bbh1t9ePbztnBpKA5x0RkBCi\nZjabja+//ppWrVphtVpJTExk7ty5VSYPqampdO7cmc6dOxMXF0d2djaZmZlkZmYyY8aMeicc69ev\n58aNG5jNZhYtWlSrnoEOHTrw3HPPoWlajSt5HGXTpk0EBAQQFxeHl5cX/fv3b/RnNsSKFSvIzc3F\nZDJhs9nsG8s5SnPeAXXAgAEsXbqUhQsXMn36dNzd3dHr9fzqV79i+vTpPPXUU+zfv58NGzY4O1RR\njVrN2dA0Tadpmg5IBtqUf1/2ctc0LULTtG8bN1Qh7jzHjh3j3//+d51WrJR/GPn4+GC1Wpk8eXK1\nSUPPnj3Jyspix44dDBs2jOzsbLp06YK/v3+DJq5eunSJhx56CKPRWKfY3dzcmiTRgNKei1atWuHn\n58fEiRPx8PBokufWV3FxMX5+fri7uzt0AqSmaXzzzTe8/PLLfPXVV812nsPYsWOJjIzk/fffr3C8\nU6dObNiwgcOHD9trxwjXU+sJokopN0oriDpuaroQdzFvb2927dqFp6cnX3xR+5Xjbm5u+Pj4YDKZ\n8PT0pGXLltVe6+Pjw5NPPslzzz1H586dmTBhApcuXaJVq1b06tWr3rGPHDmSZcuW4e/v36TbmdfF\n5MmTCQoKYsSIEfTs2bPRnpOfn8/FixcbXMPj/vvvJy8vDy8vL4euECkqKuL06dP8+te/5ty5cw5d\nUtzU/vKXv/Dqq6/adzcuZzQaeeqpp/jb3/7mpMjEj6n17CtN08xKqTBk8zUhHKK8NPitO77Whl6v\n56c//SkJCQl07ty5Tj0FgYGBPProo7W69ocffuD48eP06tWL8ePHVzg3dOhQhg4dWuvnOoOHh4dD\n5n7UpLCwkKVLl+Lr64uHh0et/2yr0qlTJ55/3vGj0Z6engQGBvKPf/yDgIAAh+x14yw9evRgxowZ\n/PnPf+b111+vcG7RokVERUXx5z//mdatWzspQlGdupYrXwm49k4/QjQThYWFjBs3DoAHH3ywTve2\nbNmS/v3725e0OlpxcTH79u1j3rx5HD16tNKW9aJUVlYWbm5u3HvvvaSmpjo7nCoppViwYAGPPPII\njz32WJW1VZqTP/zhDyxbtoxLlyquSQgMDGT06NFs3LjRSZGJmtT1X503sFApdUIp9bFS6t1bX40R\noBB3sujoaCZNmlTr38SsVis//PADGzdudHhlz1uVz3fYsGEDXl5etV5xcrcJCgqiQ4cOrF692p44\nuiK9Xk+7du0avJTYFQQFBbF48WJefPHFSueGDx/O/v37nRCV+DF1/ZcXBRwr+/r2GWkyvCJEIzty\n5AhJSUm0bNmSrVu3ct9999Xpfk3TOHz4MFlZWYwZM6baIRidTscTTzzB5cuXCQkJuSM+pBqDTqdj\nxowZzg6j0SUmJpKcnEzv3r1dYohiyZIldOvWjaNHj1ZYZTR06FD+9a9/OTEyUZ069Wxomjamhpdr\n7DAkxB1Mp9NhtVoxm8316g7/7rvv2LhxIwcPHmTp0qU1Xuvh4UF4eHiT9mqYTCb27dtn39FWOF9W\nVhbr16+nsLDQZT7Ivb29eeaZZypNCO3Xr1+znwR7p6rXrytKKQ+gK6W9GRc1TXPOHsVC3GX69++P\n2WymuLiYYcOG1fn++Ph4DAYD7u7ujbZnS0Ns2rSJ3NxcsrKy0Ol0jbKKJC0tDQ8Pj0ab73KnMZlM\n6PV6AgICOHfunLPDsVuwYAHh4eFkZWXZ/y7d3d2Jjo7m8OHDjBkzxskRilvVdW8UN6XU60AWcBKI\nBbKUUn8pWxorhGhEOp2OoUOHcs8991TaxKw2IiMjsdlsFBQU0Ldv30aIsGGKiorw9/fH19eXoqIi\nh7e/b98+Vq9ezT//+U9SUlIc3v6dqF27dowaNYqUlBQeeughZ4dj16ZNGyZPnsynn35a4fjQoUNl\n3oYLqmvPxmvAHOApYE/ZsRHAq5QmLkscF5oQwtGmTJlCt27d8PHxISgoyNnhVDJhwgS2bdtGu3bt\nGiUZSkpKYsSIEaSmppKSkmLfDVfULCYmxuU2k4PS5a6/+MUvePrpp+3Lx4cMGcKKFSucG5iopK6D\nvo8AT2iatlLTtItlrxXAT4G5Do9OCOFQSinCw8NdMtGA0o3YZs6cyaRJkxplUuqQIUPYuXMnaWlp\ndSpqdqfOH7FaraSmplJc3DxHwkeNGoXVamXv3r32Y+U9G3fq31lzVdf/mlsCF6s4fhFw/B7NQjRD\nSUlJlJSU0K1bt2qLdbVq1Yq4uDhsNluzr3vQnISFhdWpcFZxcTEfffQRN2/eZMSIEfTv37/Giq3O\npGkaycnJtGzZstbzUVavXk1mZiY2m42nnnqqXkNzzqSUYuHChXz44YcMHz4cgODgYHx9fYmPjycy\nMtLJEYpydf0pdxJYXMXxXwEnGh6OEM3bmTNn+Oqrr9i+fTu7du2q9rrBgwcTFBTEnDlzuHLlCu++\n+y5//etf74p5BHl5eVy7dq1Z/Oa5fPlybt68CcDu3bt55513XLZo1GeffcbKlSv529/+RlJS0o9e\nr2kaFy9eZN68eWiaVqkEeHMxf/58vvjiC2w2m/3Y6NGj2bZtmxOjErera7Lxa2C+UipeKbWy7BUP\nzANecHx4QjQvN27coEOHDkRGRnLjxo1qr/Pw8GDz5s0YDAZeeOEFWrVqxbBhwzh8+HCla5OTk9mz\nZ49DNufSNI2vvvqK1157jd27dze4vbpKT0/ngw8+YNWqVezYsaPJn19XN2/erFCLRNM0Dh061KgF\n1eorJSWFkSNH4uXlxalTp370eqUUI0eO5P333yc4OJi2bds2QZSO5+/vj7e3d4X/3iZMmMCWLVuc\nGJW4XV3rbOwCIijdct6n7LUOiNA0bU9N9wpxN4iJiaG4uJjExERGjRpV47UeHh589tlnREdHExcX\nx969eyvt3pqZmcmaNWu4ceMGq1atanB86enpJCUl8cQTT/DDDz80ee/C5cuX6dChA5MmTSIxMbFJ\nn10fXbt2xWQyVTre0E3XGkNUVBS7du2iqKiIgQMH1uqe0aNH8+KLLzJr1qxmPZwXEhLC5cuX7d+P\nGzeOHTt2YLFYnBiVuFWt5mwopW4tGVcEvFHVNZqmZToqMCGaI29vb+bOrf1caaUURUVF5OXlsWTJ\nEtq0aVPhfHFxMUajkZCQkEp7QdRHq1atUEqxbt06OnbsWKcN4BwhMjKSI0eO8M033zB9+vQGt3f1\n6lWsViuhoaEOiK6iHTt2kJycjJeXFxaLxZ50GAwGl5y38cADDzB27Fg8PDycPvdC0zQsFgtubk1T\nESE0NJTU1FT7ipm2bdvSqVMnDh065PIbBt4tajtBNIMfL0eu1aE9IQRQUFDA+++/z969eyslGlC6\nD8SAAQNISEjggQceaPDz3N3dWbRoEenp6U7ZGt7b25tFixY5pK0zZ86wceNG9Ho9w4YNY8CAAQ5p\n99b2Z8+eza5duyr0wnh5eWG1WtHr9Q59niO4QhKUlpbGhx9+CJT2ODzxxBON/szbezYAxo8fz5Yt\nWyTZcBG1TQ5qKsU2idIJotJfJUQdLV++nOHDh9OtW7cqzyulGDZsWL2qhVbH09OTjh07Oqw9Z7ly\n5QqRkZF4eXlx9epVh7cfExPDp59+WqmXwGw2k52djb+/v8OfeSf48ssvAQgICGiynXBDQkIqPWv8\n+PG8/PLLvPTSS00Sg6hZrQbpNE3befsLyAV+DzwHfAR0acQ4hbjjWCwW3nzzTV54QeZW10dMTAxX\nrlzh7NmzDB482OHtDxw4kP/5n/+p1FvQrl07KXVeg7CwMAAyMjKarPenfBjlVkOHDuX48eNVzrkR\nTa/Owx5KqTDgFeAh4Eugu6ZpVdXeEELU4MsvvyQoKEi6eevJz8+PhQsXNuoz0tLSSEtLQymFpmn0\n7duXadOmNflcl+ZkypQpGI1Grly5wsyZM5vkmVUNo/j4+NClSxdOnTrlktVP7za1TjaUUv7A/1Ja\nqnwvMFTTtMrr9Kq/fxkwFUjXNK1n2bE/AtMBG5AOLNA07WrZud8ATwBWYLGmaZvKjvcHVgCewH+A\nX2mapiml3IFPgP7ATeBhTdOSyu6ZD/yuLJQ/aZq2srZxC9FYNm3axMmTJ+nduzehoaH2V0hISIWv\nPTw8HP7soqIi1q5dS3Z2NlOnTqVLF+mYrMq+ffvQ6XT4+vqSl5fHiBEjJNGohXHjxjXp86oaRoHS\n3qmDBw9KsuECajWMopR6kdIqoaOA6Zqm3VOXRKPMCkrnd9zqdU3TojVN6wN8S2kyg1KqOzAb6FF2\nz/tKqfL+uA+AJ4Hwsld5m08AWZqmdQXeonQfl/KVNC8Bg4CBwEtKKekDFU730UcfkZSUxIoVK1i4\ncCE9e/YkNzeXTZs28dJLLzF+/HhatmxJ27Zt6d+/P08//bTDlqrGxcVhNBoZNWpUjcXH7nahoaF4\nenqSl5dHv3797orhk8zMTM6cOdOshh9CQkK4cuVKhcJeUJpsHDp0yElRiVvVtmfjj5QueU0Ffq6U\n+nlVF2madl91DWiatksp1em2Y7dWKfLmvytepgOrNU0rAS4ppRKAgUqpJKCFpmkHAJRSnwAzgI1l\n9/x/ZfevB/6uSn8FmQhsKV+Wq5TaQmmC0vCiBUI0gFKKgIAAAgICqt10zGazkZ6ezuXLl3nggQdY\ntGiRQ7ZdDwoKYvv27WRmZhIeHt7g9u5UQ4cOJTQ0FHd3d9q1a+fscOwKCwv5/vvvsdlsTJo0CR8f\nH4e0m5uby8cff0xAQABHjhzh0UcfdUi7jc3Dw4MWLVpw48aNCn9PgwYN4s0333RiZKJcbZONT/jx\npa/1opR6BXgUyOG/q17aAwduuSy17Ji57Ovbj5ffcxlA0zSLUioH8L/1eBX33B7LQmAhILtBCpeg\n0+kIDAwkMDCQWbNmsXbtWockG+3bt+fxxx8nLy+PTp06NTzQZmjv3r0cPXqUXr16MWZM1QvulFJN\n/rPAZDKxceNGCgsLmTBhQpWrXvbt24emabi5ubFr1y6mTJnikGdnZ2fj6enJ8OHD+frrrx3S5q00\nTSMuLg6lFN27d6/zkNS+ffs4ePAgERERTJ48ucL95UMptyYbPXr04Pr161y9epXg4GCHvQ9Rd7Vd\njbJA07THfuxVnwA0TXtR07RQ4HPgl/Vpw1E0TftQ07QYTdNiqqp5IIQzPfzww6xdu9ZhQykBAQGE\nhYXddXMQCgoKOHnyJNu3b2fmzJkcPHiQvLw8Z4dld+zYMfLz82nTpk21+3v4+PiQkZHBjRs3HNar\nAaUf2B07dmTz5s1MmnT7qHfDbdy4kS+//JIvvviCrVu31uleq9XK1q1bmTNnDmfOnLHvWVMuNDS0\n0iRRg8HAxIkT+e677xocu2gYVyrC9TmlEz5fAq4At5YEDCk7dqXs69uPc8s9qUopA6U71N4sOz76\ntnt2ODx6IW6TmJjIiRMniIyMpHv37g1ub8CAARQXFxMbG0t0dLQDIrz7WK1WPv74Y/ty1i1btuDh\n4dEok3Drq0WLFty8eROz2Vzt0M2gQYPw9PTEZrPRu3dvhz1bp9Mxbdo0h7V3u8TERNq3b4/ZbK5z\nRdzyXr7vvvsOvV6Pr69vhfPVTRKdNm0aa9eu5cknn2xQ7KJhnFoMXyl162DxdOBc2dcbgNlKKfey\npbbhwCFN09KAXKXU4LL5GI8C39xyz/yyrx8EtmulvwJuAiYopfzKJoZOKDsmRKOxWCysWbOG9u3b\ns2HDBvLz8xvcplLKPpTi6vbu3ct7773Hnj2utWVSSUkJ+fn5jBs3Dr1eT+/evfnpT3/aZGW1ayMq\nKooJEybQq1cvxo8fX+U1Sil69+5N3759m9WeJoMHD+bq1avcuHGjzoXqlFLMnz+fESNG8OSTT1Yq\ntlbV8leAyZMn88MPP1BUVNSg2EXDNFnPhlJqFaU9DAFKqVRKezCmKKUiKF36mkzpslo0TYtTSq0F\nzlBamfQXmqZZy5r6Of9d+rqx7AXwMfBp2WTSTEpXs6BpWmbZEtvy1TMvyx4uorEppdDr9eTl5aGU\nctgHwqxZs3jkkUf44x//6LLDH4WFhezcuZN58+bx2Wef0a9fP7y8vJo0BovFwjfffENaWhrjxo0j\nMjISKC01Pnz4cL788kuGDRtW7cRcZ1JK2eNNT08nNTWViIgIvL29nRxZw8XExNCtWzf0en293o+7\nu3u11XZDQ0OJi4urdLx169b06dOHnTt3NsrQkKidJks2NE2bU8Xhj2u4/hVKi4fdfvwIUGmGnKZp\nxZQWGquqrWXAsloHK0QD6fV6fvKTn3D69Glmz57tsA/b/v37Y7FYOHnyJH369HFIm47m7u6Ol5cX\nu3btwsvLq8IW7U3l4sWLZGZmcs8997B161b7hzfAyJEjGTlyZJPHVFd5eXmsWLGC0NBQDh8+7LA9\nZZytRYsWjdJudcMoAKNGjWLPnj2SbDhR8+l/E6KZCQoKYvz48Q7dh6Q5DKXo9Xp++tOf2ocoDAbH\n/05TXFxMSUlJtef9/f3Jyspi3759BAYGOvz5TaGgoAA3Nzf69u1LZqbzOmOvXr3KwYMHHTIU2Jj8\n/PzIysqq8tzw4cNdbkjvbiPJhhDNzNy5c1mxYkW9f/hbrVbS0tIatWiTj48PvXr1cuhKiXIJCQm8\n/fbbvPXWWyQnJ1d5TUBAAI8//jjDhw9nxowZDo+hKbRr144+ffqwZ88epk+f7pQYcnJy+Oyzz0hJ\nSeHzzz93Sgy1VVRUVG0P4pAhQzhy5EizKlR2p3Gl1ShCiFro1asXo0eP5s033+R///d/63z/mjVr\nuHHjBjqdjkWLFtU4zFFQUMCvf/1rEhMT0ev1Dn3pdLp63XflyhUCAgLQNI3vv/+erl271nh9fHx8\nvZ/lzMmXSinGjBlTbQ2QplBYWIibmxsRERF8//33ToujNiwWS7V/Xy1atCA8PJzjx48zaNCgJo5M\ngCQbQjRLr7zyCjExMSxatKhOlS01TSMhIYFnnnmGZcuWkZmZWe0wQ2pqKvfddx89e/bkl7/8JVar\ntcqXzWar9lxVL7PZTHFxcZ3uufVZXl5edO7cGSgtu75mzZp6tVWbZwENSqAKCwubdYXWwMBA+vfv\nz/Hjx12+h6hbt26cO3cOTdOqnDxdPpQiyYZzKEcVCLrTxMTEaEeOHHF2GEJU67nnnqO4uJj333+/\nTvdt27aNvXv30q1bN2bNmlXlb4OHDh3igQceYPHixbzwwgsut/LFZDKhlGr0Jat1TaSqegUHBzfb\neSOuKi8vD51OV2lFS3BwMAcOHKiy6uvq1atZs2YNX331VVOFeVdQSh3VNO1Hd7qTZKMakmwIV3fz\n5k0iIyPZs2cPERERdbq3ut/+oPSH8tNPP81HH33ktLkCQlQnLi6Ob7/9FoDZs2dXmIA9ZcoUFi1a\nVOW/29TUVPr27Ut6errLJc/NWW2TDRlGEaKZ8vf354UXXmDRokW88cYb9OvXr9Y/RKu6Ljs7mz/9\n6U+sW7eOrVu3OrQy5d3MbDZz+PBhvLy86N27d6N80B09etReWfMnP/mJU/d20jSNN998s9qVIbfT\n6/U888wztd5RNz4+ntGjR5Obm8v58+crJBt9+vThxIkTVSYbISEheHt7c/78+SRy9N4AACAASURB\nVDon56LhZDWKEM3Y4sWLGTlyJLNmzSIiIoKXXnqJs2fP1qmNgoICXn31VcLDw8nMzOTQoUOSaDjQ\nli1bSExMZN++fZw8ebJRnrFz5066detGy5Yt2b59e6M8o7Y0TWPJkiW4u7vbS8HX9Prkk0/q9OfS\np08ftm/fzokTJ+jVq1eFc3379uX48ePV3itLYJ1HejaEaMY8PDx4+eWX+cMf/sDhw4dZtWoVY8eO\npW3btsyZM6dSN/OtSkpKWLp0Ka+++qq96FFz+Y0vKysLpRStWrVqcFsWi4W1a9eSnJzMmDFjGDx4\nsAMi/K/i4mL8/PzQNI3i4mKHtl2uffv2xMfHo2ma0xNFnU6HTqfjt7/9LXq9vsprdu3axf79+4mI\niKBLly5YLJZat9+5c2d+/etf2yvzapqGxWLBYrHQtWvXH0029u7dyxNPPFHn9yUaRpINIe4ASikG\nDhzIwIED+etf/8ru3btZtWoV/fv3Jzw8nNDQUAwGA3q93v7/W7ZsITo6mo0bN7psNdKqxMXF2Xfx\nnDFjRrXlq2srNTWV7OxsFixYwKeffurwZGPcuHFs27aNNm3a0L9/f4e2Xe7hhx/m5MmTeHp6NvjP\nozYSEhJISUnh2rVrZGZm4u7uTlFREbm5uVitVgwGA2azucpkw2azsWPHDn7xi1+wcuVKgoKCuPfe\ne+tU/E3TNKxWKxaLBZvNhsFgwGAwoNPpKCwsZPXq1cyePbvSfeHh4axZs6ZB713Uj0wQrYZMEBV3\nApPJxK5du8jMzLT/9lf+Qzo6OrpZLgPcsGEDrVu3xmQyYTKZGlyCuqCggKVLl2I0GmnXrh0PPVTl\nrgfiFiNGjKB9+/ZERERgsVhQSmG1WlFKYTQa8fb2ZsmSJdXOT1mxYgVFRUWYTCaefPJJ6vM5dGuC\ncetzgoKCeOihh3j33Xcr3fPdd9/x3nvv8Z///KfOzxNVkwmiQgiMRiPjxo1rcDubNm3i1KlT9OnT\np9qdSJtKv379WLVqFTqdjrlz5za4PW9vb5566ilu3rxJ+/btHRDhnSc2NparV68yYMAANm7cyD33\n3IPBYMBms9mLwg0ZMoSMjAxycnIYN25cjRNh582bR2pqKu3atcPT09OhsXbp0oVjx45Vea6mKqOi\ncUmyIYSoUX5+PseOHWPhwoX84x//YNiwYU79gR0SEsKSJUuAqlfV1IeXl5d8CFUjNTWVrVu3EhER\nwZo1a8jOzkbTNHvRM6PRiMlkIjo6muvXr+Pt7V1h47uqGAwGOnXq1Cjx9u3bt9q9gwoLCx2e3Ija\nkdUoQogaeXp60qJFC9avX4+fnx8eHh7ODgmllNRKaCI2mw2dTkdeXh4FBQX24ZLyYnAmkwl3d3eC\ngoKIiYkhKirKqX83999/Pzdu3CAnJ6fSuaKiIkk2nER6NoQQNdLr9TzxxBNcvXqV9u3bO3W/kJpk\nZWVx8eJFunbt6pBVKqJUhw4dGDZsGG+88Qbp6ekYDAb69OmDj48PkydPJjAwkMDAwEbZ3bcmt1d3\nLZ+PFBYWBsDhw4crDSFKsuE8kmwIIX6Uh4eHfT8SV2SxWFi2bBkdO3Zk586d/OpXv2ryD787WUxM\nDLt27eLQoUP2eS01VaGti9/+9re89tprNV4THh5Ojx49OH36NAkJCWiahqZp9j1oyldYlX9d3YRT\nmbPhPK75K4oQ4q5w6dIlVq1axcGDBxvUTvnKlOjoaEpKSjCbzQ6KUJTT6XRYrVb7944aKsnKyuLd\nd9+1/x3e/srLy2P+/PksWbKEBQsWkJeXh9lsxmazYbFYKCkpoaCggNzcXLKysrhx4wbjx4+vsnZH\nQUGB9Gw4iSQbQginWb9+PeHh4ezdu5fr16/Xux0vLy/GjRvH7t27mTBhgnygNAK9Xm+fFOpIBoMB\nq9Va7U66t9bP0Ol09l6MmpKdgoICfHx8Kh3ftm0b/fr1c/h7ED9O+hmFEE7j5eXF1atXsVgsuLu7\nN6itAQMGMGDAAAdFJm6n0+kaJdlwc3OrsYKo0Whk7ty5nDlzhrlz59qX2tYkKyuL1q1bA6UVXN95\n5x327NnDlStXGlyXRdSPJBtCCKeZO3cusbGxPPzwwzKp08XdPozyY9LT09mxYwcBAQGMGTOm2p6I\n8mqjNQkJCSEkJKTWz87MzLRv7Pb111+zatUqXnzxRUaPHi1zeZxE/tSFEE7TqlUrRowY4eww6kXT\nNPbv309mZibDhw+vdbJUUlJCYmIi7du3p0WLFnV+bklJCevWrePmzZtMmTKF8PDwOrdRH3UdRvn2\n22/p3Lkz586dIygoiKioqCqv+7GejbrSNK1CsnH+/HnuvfdeqQzrZJJsCCFEPcTHx3PixAnCwsL4\n97//zU9+8pNa3ffJJ59gMBjIzMzk5z//eZ3nl5w9exZN0xgzZgw7duxosmRDp9OxatUqAgICsFgs\nmM3mal8Wi4VWrVpx/vx5jEYjH3zwASaTqcp2Dx8+zLRp0xwWZ2FhIQaDwV4PJj4+ngkTJjisfVE/\nkmwIIUQ9lM9hMJlM1e5uejtN07h27RqPPfYYa9euJS8vr87JRmBgIJs3byYvL6/RqnBWZeHChZw5\nc4b09HTc3NwqvNzd3fHx8bF/Xz5UUVRUhF6vJzg4uNp2HV0CPzMz0z5fA0p7Np5++mmHtS/qR5IN\nIYSoh/DwcPLy8sjMzOSee+6p1T1KKSZNmsSXX35JVFQUbdq0qfNzAwMDeeKJJ8jJyWnSZOPZZ59t\nsmc1xJUrV2jXrh1QmtzFx8c3yU64omaSbAghRD0opeq1ZbwjVs34+/vj7+/foDbuVCdOnKBPnz4A\nXL9+HXd39wo9HcI5pM6GEEKIO8bx48fp27cvUDqEIr0arkGSDSGEEHcEm83Gvn377IW7yieprlu3\nzsmRCUk2hBBC3BH+8Y9/4OPjw6BBgwCIjo5m06ZNvPDCC/z2t7+tU50Q4ViSbAghhGj2rly5wksv\nvcTy5csrrA7q168fhw8fZt26dezcudOJEd7dJNkQQgjR7G3bto2xY8cSGRlZ6VybNm3qVIFUOJ4k\nG0IIIZq9I0eO1LjKp6ioSDbocyJJNoQQzZ7JZGL79u3s2LHDoaWvRfNx5MgRYmJiqj0vyYZzSbIh\nhGj2du7cybVr10hJSWHfvn3ODueOZLPZKCgoQNM0Z4dSpRMnTtiXvFZFkg3nkqJeQog7glIKpVSj\nfxiazWaWL19OXl4eU6dOJSIiolGf5wqsVisrVqzg+vXrdO/enRkzZjg7pErMZrN9P5SqSLLhXNKz\nIYRo9kaNGkXbtm0JCQlh6NChjfqsLVu2kJGRgb+/P1999VWjPqsqiYmJHD169Ee3Za+P/fv38847\n77Bu3boK7d+8eZOcnByee+45Tp48WafdXxtLfHw8r776Ku+99x65ubm4u7tTUlJS7fWSbDiXJBtC\niGbPaDQyduxYxowZg5ubW6M+q2XLllgsFjIyMnB3d2/UZ90uKSmJr7/+mri4OL799luHtp2RkcG2\nbdvw8fEhPj6eI0eO2M/5+/vTsmVL3nzzTaKjo9HpnP/RcfDgQaZOnUpgYCBnzpzBw8NDkg0XJsMo\nQghRB8OGDaO4uJj09HSmTJnSpM/OycnBz8+Pzp07k5CQ4NC2y2tT5OXloWlahSEJvV7PY489RmFh\nId7e3g59bn116dKFzZs3Y7VaGTp0KO7u7hQXF1d5raZpkmw4mSQbQghRR2PHjnXKc3v06EFKSgqX\nLl1i4sSJDm3bz8+PBx98kIMHD9K/f3/7ZmbldDodPj4+Dn1mQwwbNozw8HA8PT3x9fWtcRjFZDJh\nMBgqFPsSTUuSDSGEaCYMBgPTpk1zaJsWi4XTp09jtVrRNI2ZM2e6VFJRk7Zt29q/rinZkF4N55Nk\nQwgh7jLXrl0jMjKS4uJi+wd0v379uHr1KosXL+Y3v/mNkyOsu5qSjfz8/GaTQN2pJNkQQoi7TF5e\nHq1btyYuLg4o/aDW6XT87ne/a7ablen1+mpXyWRlZeHn59fEEYlbSbIhhBB3mfKaJHfS0IJOp6s2\n2cjOzpZkw8mcv35JCCFEk2qK4mdNraZkIysri1atWjVxROJWkmwIIcRd5m5MNqRnw7lkGEUIIZpY\nXFwcW7ZsISgoiJkzZ2IwNO2PYm9vb5KSkujduzeaptlf169fZ8mSJU0aS11cuHCBU6dO0b17d6Ki\noiqc+7FhFOnZcC5JNoQQoolt376dKVOmsGvXLi5evNjk+6u0a9eOpKQksrKy7PM3yl9dunRp0lhq\ny2w2s379esaOHcs333xDx44d8fLysp+Xng3XJsmGEEI0sfbt27N9+3Zyc3Np06aNU2Lo2LEjHTt2\ndMqz60MphV6vJycnB6VUpZLpNSUb165do2fPnk0RpqiGJBtCCNHEpk+fzqVLl/D395ffuGvJYDAw\nb948zpw5wyOPPFJph9eako0TJ04wd+7cpghTVEOSDSGEaGJ6vZ6uXbs6O4xmJzg4mODg4CrP6XS6\nKmuEWCwWYmNjK5VfF01LVqMIIYRodBkZGdy8ebPR2q8u2bh27RotW7bE19e30Z4tfpz0bAghhGg0\nJSUlrF27lkuXLqHX65k6dSq9e/d2+HP8/f25ceNGpeMmkwl3d3eHP0/UjSQbQghRg8LCQpKTkwkN\nDZX9NepI0zTeffddCgsLAbDZbGzfvt1hyYbNZmPjxo1cv36dtm3bcuHChUrXWCyWJl9aLCqTYRQh\nhKiGzWZj2bJlHDp0iH/+85+YzWZnh9SsFBUVUVhYaP+wt9lsGI1Gh7V//vx5rly5QnR0NLm5uVUm\nG2azGTc3N4c9U9SPJBtCCFENi8VCdnY248aNw2QyUVRU5OyQmhVPT08iIiKwWCwopQgKCuL+++93\nWPutWrUiJyeHs2fP0qVLl2p7NiTZcD7pWxJCiGoYjUbGjBnDF198wYABA2jRooWzQ2pWlFLMnj0b\nTdNQSjm8/cDAQObMmcONGzeYOHEib775JteuXSMwMNB+jdlslmEUFyB/A0IIUYNhw4YxbNgwp8Zg\ns9k4ePAgJpOJIUOGOHQooik0RqJRLiQkhJCQEAAeffRRXn/9dd544w37eRlGcQ0yjCKEEC7u8OHD\nnDlzhitXrrB9+3Znh+OyZs2axd69eysckwmirkGSDSGEcHHl8w48PDywWCzODsel6fX6Ct9Lz4Zr\nkHRPCCFc3KBBgygqKsJkMjF69Ghnh+OybDZbpT1TZM6Ga5C/ASGEcHEGg4Fx48Y5OwyXZ7VaKyUb\nGRkZ+Pv7OykiUa7JhlGUUsuUUulKqdO3HHtdKXVOKXVKKfWVUqrVLed+o5RKUErFK6Um3nK8v1Iq\ntuzcu6ps5pFSyl0ptabs+EGlVKdb7pmvlLpQ9prfNO9YCCGEo1VVkryczWarNIxy+fJlQkNDGzss\n8SOacs7GCmDSbce2AD01TYsGzgO/AVBKdQdmAz3K7nlfKVX+L+gD4EkgvOxV3uYTQJamaV2Bt4DX\nytpqDbwEDAIGAi8ppWSbRSGEaEY0TWPNmjX86U9/YsOGDVVeU9UwiiQbrqHJkg1N03YBmbcd26xp\nWvlspwNASNnX04HVmqaVaJp2CUgABiqlgoAWmqYd0DRNAz4BZtxyz8qyr9cDY8t6PSYCWzRNy9Q0\nLYvSBOf2pEcIIYQLy87OJjk5mf/5n//hxIkTmEymStdUNYwiyYZrcKXVKI8DG8u+bg9cvuVcatmx\n9mVf3368wj1lCUwO4F9DW0II4RB5eXns2bOHS5cuOTuUO1aLFi1o1aoV7733HmFhYVWuMJFhFNfl\nEhNElVIvAhbgcyfHsRBYCNChQwdnhiKEaEbWrl2Ln58f+/fvZ/78+bRt29bZITWqkpISrl+/TlBQ\nUJMtK9Xr9Tz++ONkZWXh7+9fZaGwzMzMSlVeJdlwDU7v2VBKLQCmAnPLhkYArgC3/usIKTt2hf8O\ntdx6vMI9SikD0BK4WUNblWia9qGmaTGapsW0adOmAe9KCHE3KSgooH379ri7u1NcXOzscBqVxWLh\nn//8J//+979ZuXIl//2x3fgMBgNt2rSpNFRS7uzZs0RFRdm/LywsJD8/H/l57nxOTTaUUpOAXwP3\naZpWeMupDcDsshUmYZROBD2kaVoakKuUGlw2H+NR4Jtb7ilfafIgsL0sedkETFBK+ZVNDJ1QdkwI\nIRzi/vvvJzExkb59+97xv0Xn5eVRWFjI3LlzSUtLc6mdcG9PNlJTU2nfvn21yYloOk02jKKUWgWM\nBgKUUqmUrhD5DeAObCnrEjugadpTmqbFKaXWAmcoHV75haZp5eudfk7pyhZPSud4lM/z+Bj4VCmV\nQOlE1NkAmqZlKqX+CBwuu+5lTdMqTFQVQoiGCA0NZc6cOc4Oo0m0atWKiIgIPvjgAwYPHuxS+7Tc\nnmzIEIrrUE3ZBdacxMTEaEeOHHF2GEKIu5jZbOb06dP4+/vLPLIfYTab8fX1JTs7Gw8PDwBWrFjB\ntm3b+PTTT50c3Z1LKXVU07SYH7vOJSaICiGEqGzDhg0UFBSQnp7O7Nmz7bubisoSEhIIDQ21Jxog\nPRuuRAayhBDCRWVlZdGtWzf8/PzIzs52djgu7fYhFJBkw5VIsiGEEC5q0qRJnDt3jrZt21b6IBUV\nSbLh2mQYRQghnCQ/P589e/bQokULBg8eXGnVREhICAsWLHBOcM3M2bNnGT9+fIVjkmy4DunZEEII\nJ/n+++8pLi4mNjaW2NhYZ4fTrJ05c6bKng2Z5+IaJNkQQggXUFVFTFE7NpuN+Ph4IiMj7cdyc3Mx\nm820bt3aiZGJcpJsCCGEk0yePBlPT0+io6Pp1auXs8NptlJSUvDz86tQqjw1NZXQ0FBJ4lyEzNkQ\nQggn8fb2ZuLEic4Oo1HYbDaysrLw8/Nr9AqeZ8+epXv37hWO5efnV9onRTiPJBtCCHEXyMnJIT4+\nns6dOxMQENCoz7JarbzxxhsUFRXRokULfvWrXzks4fjuu+/IysrCarXaX5s3byY4OLjCde7u7lVu\nQy+cQ4ZRhBDiDmez2Vi2bBmXL19m+fLllJSUNOrzUlNTKSoq4vnnnyc3N5f8/HyHtFtcXMzUqVPZ\nuHEj27ZtY/fu3Rw8eJD169eTm5tb4Vqj0djo71PUnvRsCCHEHc5qtVJYWEjPnj1JSEigpKQEd3f3\nRnte+Q64b7zxBt7e3vj4+DikXZ1Oh8Fg4PPPP69wfNWqVYwcObLCMXd3d0k2XIgkG0IIcYdzc3Nj\n6tSp7Nu3j9GjRzf6XAaDwcALL7zA9evXCQwMbNAQSlpaGtevXycqKgo3NzcsFkulawoLCxk3blyF\nY5JsuBZJNoQQ4i7Qu3dvevfu3WTP0+v1leZR1FVGRgaffvopQUFBnD59mnnz5uHm5sZf//pXCgoK\n8Pb25vnnnweo1HtiNBplzoYLkTkbQgghXFJOTg4+Pj5ER0eTmZkJlA7RFBQUcO+991JQUEBSUhJ6\nvZ6EhIQK9/r6+pKfn19lT4hoepJsCCGEcElhYWGEhYVx+PBhpkyZApT2dgD85bu/8B/+Q0pRCp6e\nnuzbt6/CvR4eHrRv356LFy82edyiMhlGEUII4VLy8/PJyMigQ4cOTJ48ucI5TdMYcO8A3vr+LWIM\nMTy4/kHatGnDjRs3KrXTvXt3zpw5Q0RERFOFLqohPRtCCCFcxrVr13jzzTdZuXIly5Ytq3AuJyeH\nCRMmsHf/Xmw2G9dKrlGgCkgJTGH5ueVcSLtQ4fryZEM4nyQbQgghXEZsbCxGo5Hp06eTlpYGQEp6\nCr1f6U2ft/ugBWh4ZHvQWevMGc5QbCvG6mulKK+IV7a8UqGtqKgozp4964y3IW4jyYYQQgiXERMT\ng8Vi4ZtvvqFLly4AvPjFi1g1K10NXTkbdBZN07jABbSy/+EPths28lXF4mHSs+E6JNkQQggXVlxc\nzH/+8x82bdp0Vyzl9PPz4ze/+Q1Dhw4lLS2N1atXY8uwkWpN5azlLC2tLWndujUWblll0gFIhUuZ\nlyq0FRUVRXx8PDabrWnfhKhEkg0hhHBhO3bsoKCggMzMzEorLu5UVquVAwcO8Oijj3LhwgV+MfYX\nLGi9gGkB00hflc748eMJMgT99wZPoDWcPH0STdPsh319ffH39yc5Obnp34SoQJINIYRwYW5ubhQV\nFVFUVITBcHcsIHRzc6NNmzZ8/fXXGI1GDh06RHBRML+b8TusZisGg4F7Ot1T8aZOYE20suXClgqH\no6KiZCjFBdwd/3KFEKKZGjVqFAcOHECv1zNw4EBnh9MklFI89thjXL58meDgYLKysvDy8sLPzw8P\nDw8OHjzI0aSjFW/qBByC9/a8x4RuE+yHy+dt3HvvvU36HkRFkmwIIYQLMxgMDB8+3NlhNDl3d3e6\ndu0KgJeXl/34wIEDuXDhAhEqglOc+u8NHYEvYW/a3grtdO/enf379zdFyKIGMowihBDC5WiaVmli\np8ViIaBrAIe8D2HUjPyMn+GNd+nJsnkbeYl5Fe6RYRTXIMmGEEIIl3Lt2jVef/11/u///o9z587Z\nj+v1ej63fE5GUQZf8zUGvYElLGEwg0sv6ACWyxbSC9Lt94SFhZGUlNTE70DcTpINIYQQLsNsNrN0\n6VKKioqwWCwcPnzYfk4pRZ4pj94deuPfwp++A/vi4e7BH4b8ASNGhnYYii5VR2JWov2edu3acfPm\nTdmQzckk2RBCCOEyUlJSgNI5G5qmERgYWOF85+OduVRyiSVDl/DYhMeYOnUqFouFV0a9QnxoPNbL\nVgYED7BfbzAY8Pf3r3LvFNF0JNkQQgjRYJcuXeK7775rcE2LTp06YTAYKCkpQa/XM3bsWABMJhNr\n1qxhcvRkngt9jqcHPQ1Az549mTJlCktGL+HaX65BAZw6eapCm0FBQfbS58I5JNkQQgjRIMXFxaxZ\nswZfX19WrVqF2Wyud1t6vZ7f/va3LF68mGeeeYbLly9jtVqJi4vDZDJx5coVTp8+TW5ubqV7DQYD\nPj4+/P3vf69wXJIN55NkQwghhEPodI75SFFK4e7uzocffsi3337L2rVradOmDWlpaURERGCz2fjo\no48qzcP429/+RlRUFJcvX6awsNB+XJIN55NkQwghRIN4eHjw0EMPkZWVxezZs3Fzc2twmxkZGRiN\nRsLCwjh//jw//PADI0aMQCmF0WikpKSk0l4xmZmZdO/enatXr7J792778eDgYC5fvtzgmET9SVEv\nIYQQDdalSxf7Lq2OoJRCKcXhw4cZOHAgx44dIykpicuXL6PX64mKiqpQ7Mtms2EwGPDy8iItLY1e\nvXrZz8XExPDBBx84LDZRd5JsCCGEqDez2cz58+dp164dAQEBDmmzuLiYTz75BIPBgFKKU6dKJ3wG\nBQVhMpnQ6/UMHlxaWyM/P5/PP/+ca9eu0aZNG6Kjo3n//ffx8/Oztzd8+HAeffRRLBbLXbO/jKuR\nYRQhhBD1tm7dOg4dOsSyZcvIyspySJv5+flYLBaKi4vRNA0/Pz80TePKlSuUlJTQo0cPQkJCANiz\nZw83b94kPDyc4uJiunXrhl6v5+jR/+6d4u/vT8eOHTl27JhD4hN1J8mGEEKIektPT2fw4MH4+vqS\nmZlZ73bi4uLYvHkzJSUl+Pn5oZQCSodT0tLS8PT05P7778fLywsPDw+sVitmsxlfX1/c3Ny4ePEi\nJpOJ4OBgPD097b0h5UaOHMnOnTsb9F5F/UmyIYQQot4mTpzIDz/8QHBwMGFhYfVq4/Tp06xfv56j\nR4/yj3/8o8IOt5qm4eXlhcFgYPfu3RQUFHDo0CFeeeUVXnvtNby9vRk/fjwDBw7k4YcfJjs7G4PB\nwIULFyo8Y+TIkRUmjYqmJYNXQggh6i0qKoqoqKgGtZGamorRaKRHjx7ExsYCMGnSJPz9/Tl+/DjZ\n2dl06NCBfv368cYbb5CWloaXlxc3b97ko48+Yv369Vy/fp3AwEA6duxIhw4duHTpEpqm2XtIhgwZ\nwi9/+csKx0TTkZ4NIYQQdWI2m9mzZw8HDhzAarU2uL1Ro0bZhz5GjhxpP96vXz+uXbvG448/zsWL\nF/Hx8WHixIkEBQXh4+NDSEgIhw4d4v333yclJYW0tDTmzJnDAw88QFpaWoUN2EJCQnBzc+PSpUsN\njlfUnfRsCCGEqJOdO3eSlpaGyWTCZrMxdOjQBrXn6enJM888U+m4TqcjODiY9evX4+bmhq+vL/fe\ney/Dhw9n/PjxvP/++/z5z3/mvvvus68y8fPzIzMzk7S0NFq0aGFvSynF4MGDOXDgAJ07d25QvKLu\npGdDCCFElTRNY//+/Xz99ddkZGTYj5eUlFBYWEhRURHFxcWN9nylFPPnz2fixIksXLgQo9EIQMuW\nLSkpKWH//v3YbDb0er39nlmzZjFq1ChycnJo3bp1hfbKkw3R9CTZEEIIUaXExESOHj2Kr68vX3/9\ntf24j4+PfXnq7VU8a1JcXExmZiaaptXq+oKCAgoLCwkLC8PT07PCuQ4dOrB48WLmz5+PpmkcP36c\nEydOYDQaueeeezAajVy/fr3CPUOGDJFkw0lkGEUIIUSVDAYDZrOZgoKCCiXIlVIEBgbat4Gvjezs\nbD766CM0TSMmJoYxY8bUeH1ycjKrV68GYPLkyURHR1c4/+9//9v+9e7duzl37hw2m42ioiKGDBlC\nZGQk586dq7BFfb9+/YiLi6OoqKhS8iIal/RsCCGEqFLHjh0ZP3483t7e3H///fbjQ4YMoW3btnh7\nezN69OhatXX58mUCAwOZNWsWZ8+e/dHrExIS6NOnD/fccw/nz5+v8dr8/HzatGlDQEAA+fn5APZk\n41ZeXl5ERUVJcS8nkGRDCCFEtXr27MnYsWMrTLZ0c3Nj/PjxTJ48udY9BF26dCEnJ4dPPvnEXkOj\nJr169SI2NpatW7fSr1+/CueysrI4dOiQvYjYyJEjsVqt6HQ6hg0bBpQmTD6jawAAF0lJREFUG/Hx\n8ZXalXkbziHDKEIIIRqdl5cXP//5zytN6KxO27Ztef7559E0rcLW9RkZGfzzn/8kICCAHTt28Oyz\nz+Lt7c3MmTMr3B8ZGcnWrVsrtTt48GC++eabhr8hUSfSsyGEEKJJKKVqlWjcev2tiUZJSQnLly/H\nZDJx/fr1GlfDVDWMAtKz4SySbAghhHB5NpuN7OxsbDYbOp0Oq9VqT1xsNlul6zt16sS1a9coLCys\ncLxLly4UFxeTmpraJHGLUjKMIoQQwqXl5+fz8ccfk5eXh7e3NwB6vZ6uXbvy97//HYPBwKOPPkq7\ndu3s9+j1erp06cL58+fp06eP/Xh5ca+DBw/ad44VjU96NoQQQrgsTdOIi4vD398fvV5Pbm4uJSUl\nzJs3D4vFwrhx42jXrh0ff/wx69atq1A+vU+fPlWuPBk8eDD79+9vyrdx15NkQwghhEvSNI3PP/+c\n7du3k5SUhMlkwsPDA03T2LNnD926dWPLli1cunSJCRMmcOPGjQrDI9XNz5B5G01Pkg0hhBAuqbCw\nkJSUFJ5//nl0Oh1eXl6YzWYAgoKCOHXqFEajEaPRyOHDhyksLMTf399+f/lwye0GDBjA8ePH7W2J\nxifJhhBCCJfk5eVFWFgYr7/+Op06deKpp54iKiqKe+65h0GDBnHx4kU6d+6MyWTCbDYzbdo0fHx8\n7PdHR0eTkJBAXl5ehXZbtGhBUFAQCQkJTf2WKsjJyeHatWu1ujYlJaXBCdLRo0dZtWoVFy5cqHcb\n9SXJhhBCCJeklGL27Nk8++yzzJkzB19fX2bOnMmIESMwGAwEBgYSGxsLlJZD//bbbyvcbzQa6dOn\nD0eOHKnUdvfu3WtVybSxXL16laVLl/Kvf/2LPXv21Hjt5cuXWbt2LbGxsWzYsKFez8vKymLbtm1E\nRkayfv36WpeZdxRJNoQQQrgspRReXl4opSocP3bsmL2CKJTO78jPz68wQRRg0KBBVc7PiIqKcmqy\nkZKSQpcuXRg7diyJiYk1XpuTk4Ofnx/dunUjKyurXs8r39smNTUVDw+PerXREJJsCCGEcHmappGV\nlYXJZOLSpUts3boVg6Fy9YZbi4BB9ZNBnZ1s9OjRg/T0dDZt2sSQIUNqvDYqKoq2bdty7tw5Jk2a\nVK/n+fj48Mgjj9CyZUvmz59fKXlrbFJnQwghhEtLTEzku+++Iy8vD6PRSEFBQZXXGQyGSh+igwcP\nZvHixWiaVuFc9+7deffddxs17pr4+vrys5/9rFbX6vV6pk2b1uBnhoSEOK22iPRsCCGEcFmaprF2\n7VqKi4vR6XQYjUYAJkyYAIC7u7v92ls3iysXGhoKUKliaPlGbVVVHxWOJ8mGEEIIl3HkyBE+++wz\nkpOT7cc8PDxo3bo1JSUl9o3ZNm/ejMFgoGfPnri5ueHm5kbLli0rtaeUonPnziQlJVU43qJFC/z8\n/EhJSWnstySQYRQhhBAuIjU1le+++46WLVvyySef8Lvf/Q6lFI8++iixsbGMGTOGsLAwNE0jIyOD\ngIAAEhISOHfuHG5uboSHh1fZbmhoKJcvX650PDw8nIsXL9KpU6dGfmfi/2/v/oOrqs88jr8fEgKR\nghJD+CFIkV9FlhUXZK3NBihFsQUUF1jrDkVpiy4dWtt1unUc0VE7U1t3au0u7LhblbKj0FJbZKyz\nBreFKKCCQIBSMfwUUgyQQGAhuSZ59o9zEm+SG36Zm3OT+3nNnMm53/M9h+c+5Mdzv+d7zlGxISIi\nKaGyshIzY+DAgRQXFze05+TkMH78eA4dOsSKFSvo27cvBQUFmBnDhg1jzpw5VFdXN5wyaWrAgAEJ\nH7yWl5fH0aNHk/Z+5BNtdhrFzJ4zszIz2xHXNsvMdppZnZmNbdL/QTMrMbP3zeyWuPYxZrY93PaM\nhTN+zKyLma0I2982s8/G7TPXzD4Il7nJf7ciInIxamtrycnJYdCgQezevZuJEyc2m+y5atUqBg4c\nyPbt2xudFunduzdXX311i1dYtDSy0atXLxUbbaQt52y8ADS9ZmcHcAewLr7RzK4F7gRGhvssNrOM\ncPMS4JvA0HCpP+bXgQp3HwL8FHgyPFYO8Ajwt8A44BEz69mab0xERC5d/TNQXnzxRc6ePcsDDzxA\nQUFBs37dunWjtLSUqqoqsrOzL/j4Kjai12bFhruvA8qbtO1y9/cTdL8NWO7u1e6+DygBxplZX6CH\nu2/04PZnvwRuj9tnabi+EpgUjnrcAhS6e7m7VwCFNC96REQkIrFYjAMHDrBgwQJOnDjR7Pbi9WbO\nnEnv3r2ZMWMGffr0ueDj5+bm8tFHHyVsV7HRNlJ1zsZVQPxdWA6FbR+H603b6/f5EMDda8zsJHBl\nfHuCfUREJGJdunRh9OjRPPnkk4wYMSLhVSUQ3JgqPz//oo+/efNmrrvuuoTHa+meHdK6UrXYiISZ\nzQfmA1x99dURRyMikj6mTZvGV77ylWZ3AG0Na9asYd68ec3ay8vLycnJafV/T5pL1ftsHAbipxX3\nD9sOh+tN2xvtY2aZwOXA8XMcqxl3f9bdx7r72F69erXC2xARkQuVjEIjFotRVFTExIkTm22rv3xW\nki9Vi41XgDvDK0wGEUwEfcfd/wJUmtmN4XyMrwGr4vapv9JkJvC/4byO/wFuNrOe4cTQm8M2ERHp\n4DZu3Mjw4cMTjmCo2Gg7bXYaxcxeAiYAuWZ2iOAKkXLg50Av4FUz2+rut7j7TjP7FfAnoAb4lrvX\nP8pvAcGVLdnAa+EC8AtgmZmVhMe9E8Ddy83sceDdsN9j7t5ooqqIiHQcsViMV199lZMnT7Jt2zYm\nT56csN/x48dVbLSRNis23P2rLWz6bQv9fwj8MEH7JuCvErRXAbNaONZzwHMXHKykpNraWlasWMG+\nffsYP378JU0UE5GOr7i4mNOnTzN48GCeeOIJnn322YT9jh07xpVXXtnG0aWnVD2NItJMaWkp5eXl\n3HvvvRQVFUUdjoikqJ49e3L06FGKi4spLS3lpptuStivtLSUvLy8No4uPanYkHYjNzeXmpoaXnzx\nRa655pqowxGRFDV48GBmzJhBLBYjPz+frl27NutTVlZGWVkZw4cPjyDC9KNLX6VF1dXVbNiwgezs\nbG644YakzBS/GNnZ2dx3330cP36cvn37RhqLiKS2QYMGUVJS0vAo+qbWrVtHfn4+mZn6M9gWlGVp\nUWFhIadPn6ayspLMzEzGjBkTdUh07dqVq67SPdlE5Py2bdvG7NmzE25bu3YtEyZMaNuA0phOo0iL\nampqyM7OJisri5qamqjDERG5KEeOHGlxFPQPf/iDio02pGJDWjR58mQyMjIYMGBASoxqiIhcjJaK\njeLiYk6cOMHo0aMjiCo96TSKtKhbt25MnTo16jBERC7aqVOncHc+85nPNNu2ZMkS5s+fr/kabUiZ\nlnavvLyciooKBg0aFPkkVhFJDUVFRYwZM4bgZtOfqKysZPny5ezcuTOiyNKTig1p18rKynjhhRfo\n3r07/fv3Z9q0aVGHJCIpYNWqVdx2223N2pctW8bkyZPp169fBFGlL30MlHatrKyMvLw8CgoKOHw4\n4fP1RCTN1NXVsXr1aqZPn96o3d1ZvHgxCxYsiCiy9KViQ9q1YcOGkZWVxZo1axI+1VFE0s+mTZu4\n4oorGDp0aKP21157jYyMDMaPHx9RZOlLp1GkXcvKyuKuu+6KOgwRSSGvvPJKwlGNRx55hEWLFjWb\nxyHJp5ENERHpUBLN1/j9739PdXU1d9xxR0RRpTcVGyIi0mHs3buXsrIyxo0b19Dm7jz22GMsWrRI\nV6xFRFkXEZEOY/Xq1UybNo2MjIyGtg0bNnD8+HGNakRIxYbIBYrFYlRWVkYdhoicw8aNGykoKGjU\n9vTTT/Ptb39boxoRUuZFLkBFRQXPPPMMS5Ys4a233oo6HBFpwdatWxvdhvzgwYO88cYb3HPPPRFG\nJSo2RC7A/v37GTBgADNnztSdB0VS1JkzZzhw4ACf+9znGtoWL17M3Llz6d69e4SRNRaLxXj33XfZ\nt29f1KG0GRUbEplYLEZhYSGFhYXEYrGowzmnIUOGcOTIEZYvX66H0omkqB07djB8+HCysrIa2lav\nXp1yl8evWrWK3bt38/LLL3Pw4MGow2kTus+GRKaoqIhjx47h7hQVFTFp0qSoQ2pR9+7dWbhwIbW1\ntXTu3DnqcEQkgV27dnHttdc2vK6oqODgwYMp93TXyspKRowYwdmzZ9NmHphGNiQynTp1oqamhtra\n2nYxcatTp04qNERS2GWXXUZ1dXXD6z179jBkyJCUe7rrrbfeyp49e+jXrx8jRoyIOpw2kVr/A5JW\n8vPzWb9+PQA33XRTxNGISHvXq1cvysrKGl6fOnWKHj16RBhRYv369WPOnDlRh9GmVGxIZDp37qxn\nFIjIpzJr1ix2797NvHnzmDx5MkePHm3YdurUqZSaGJrOUn/sWkREJIG9e/dSWFjIhAkT2LJlS7OR\njdOnT6vYSBEqNkSkwzh9+jSbN2/m2LFjUYciSXbw4EEGDx5MXl4eI0eOpLa2lpycHE6ePElNTQ2g\nkY1UomJDRDoEd2fp0qWUlJTw/PPPc+bMmahDkiQqLi4GoKqqiocffpiMjAwyMjLIyclpKDb79OnD\ne++9h7tHGaqgYkNEOpATJ05w/fXXY2YqNjq4KVOmsGPHDl5//XXWr1/P4sWLARg6dCjbt28HYNq0\nadTV1bFy5cooQxU0QVQkErW1tWzatAkzY+zYse3i0t9UZ2ZMnz6ddevWMW7cOHJzc6MOSZIoMzOT\nkSNHNmufOnUqL730Ev369SMWizFr1iweeughbr/9dl26HiEVGyIRePPNN9m7dy91dXV8/PHHfOEL\nX4g6pA5h1KhRjBo1KuowJEIzZ85k+vTpzJ49m6ysLLKysujfvz8nT55UARohFRsiEaiqqqJHjx7U\n1tZSVVUVdTgiHcbQoUPZtWtX1GFIExq7FYlAQUEBnTt35rLLLtOohoh0eBrZEIlAdnY206dPjzoM\nEZE2oZENERERSSoVGyIiIpJUKjZEREQkqVRsiIiISFKp2BAREZGkUrEhIiIiSaViQ0RERJJKxYaI\niIgklYoNERERSSoVGyIiIpJUKjZEREQkqVRsiIiISFKp2BAREZGkUrEhIiIiSaViQ0RERJJKxYaI\niIgklYoNERERSSoVGyIiIpJUKjZEREQkqVRsiIiISFKZu0cdQ0oys6PAgajjSJJc4FjUQaQg5SUx\n5SUx5aU55SSxjpyXge7e63ydVGykITPb5O5jo44j1SgviSkviSkvzSkniSkvOo0iIiIiSaZiQ0RE\nRJJKxUZ6ejbqAFKU8pKY8pKY8tKccpJY2udFczZEREQkqTSyISIiIkmlYkNERESSSsVGO2Jm+81s\nu5ltNbNNYdujZnY4bNtqZl+O6/+gmZWY2ftmdktc+5jwOCVm9oyZWdjexcxWhO1vm9ln4/aZa2Yf\nhMvctnvX55coL2H7QjP7s5ntNLMfx7WnbV7C91H/vbLfzLbG9U/nvIw2s431bWY2Lq5/OuflOjPb\nELavNrMecf3TJS9XmNnK8HfJLjP7vJnlmFlhGG+hmfWM658Weblo7q6lnSzAfiC3SdujwAMJ+l4L\nbAO6AIOAPUBGuO0d4EbAgNeAW8P2BcB/hOt3AivC9Rxgb/i1Z7jeM+p8nCcvE4E1QJfwdZ7y0mz7\nvwKLlBcHeD3ufX0Z+KPy4gDvAuPD9XnA42mYl6XAN8L1LOAK4MfAD8K2HwBPplteLnbRyEbHdRuw\n3N2r3X0fUAKMM7O+QA933+jBd/Qvgdvj9lkarq8EJoXV9y1AobuXu3sFUAhMacs3cwn+CfiRu1cD\nuHtZ2J7ueQEgjH828FLYlO55caD+U/vlQGm4nu55GQasC9cLgb8P19MiL2Z2OVAA/ALA3WPufoLG\n72Upjd9jh8/LpVCx0b44sMbMNpvZ/Lj2hWZWbGbPxQ3nXQV8GNfnUNh2VbjetL3RPu5eA5wErjzH\nsVJForwMA/4uHJZca2Y3hO3pnpd6fwd85O4fhK/TPS/3Az8xsw+Bp4AHw/Z0z8tOgj+GALOAAeF6\nuuRlEHAUeN7MtpjZf5lZN6C3u/8l7HME6B2up0teLlpm1AHIRcl398NmlgcUmtmfgSXA4wS/KB4n\nGBqfF2GMUUiUl0yC4ccbgRuAX5nZNVEGGYFmeXH3+k+pX+WTUY10k+j7ZSbwXXf/jZnNJvgk+6VI\no2x7ifIyD3jGzB4GXgFikUbY9jKBvwEWuvvbZvYzgtMmDdzdzUz3kDgPjWy0I+5+OPxaBvwWGOfu\nH7l7rbvXAf8J1E9sO8wnn0IA+odth8P1pu2N9jGzTILh5OPnOFZKSJQXgk8BL3vgHaCO4GFI6Z6X\n+vdwB7Airnu652Uu8HLY5dfo56j+98uf3f1mdx9DUJzuCbunS14OAYfc/e3w9UqC4uOj8NQI4df6\n07TpkpeLF/WkES0XtgDdgO5x6+sJzt/1jevzXYLzhQAjaTxRaS8tT1T6ctj+LRpPVPpVuJ4D7COY\npNQzXM+JOifnyct9wGNh+zCC4UhL97yEr6cAa5v0T+u8ALuACWH7JGCz8sIUPplY3YlgnsG8dMpL\nGF8RMDxcfxT4SbjETxD9cbrl5aLzGHUAWi7wPwquCb+JtxGcR30obF8GbAeKCYY544uPhwg+ibxP\nOPM5bB8L7Ai3/Ruf3Em2K8GnupLwB+OauH3mhe0lwD1R5+MC8pIF/Hf4Pt8Dvqi8NGx7AbgvwT5p\nmxcgH9gctr8NjFFeHOA7wO5w+VH9e0yXvISxjQY2EfyO/R3BH/4rgTeADwiuesuJ658WebnYRbcr\nFxERkaTSnA0RERFJKhUbIiIiklQqNkRERCSpVGyIiIhIUqnYEBERkaRSsSEi7ZaZ3W1mp6OOQ0TO\nTcWGiLQKM3vBzDzBsrGVju9mNrNJ8wqCe0SISArTs1FEpDWtAeY0aUva8zTc/SxwNlnHF5HWoZEN\nEWlN1e5+pMlSDmBm3wufTvx/ZnY4fILmFfU7mtnlZrbMzMrMrMrM9prZ/eG2/WG3X4cjHPvD9kan\nUczsUTPbYWZ3mtkeMztlZr8zs9y4Pplm9lMzqzCzcjN7yswWm9kfk58ekfSkYkNE2kodwaPcRwJ3\nETzs7Odx258ARgFTgeEEt2quf/DUDeHXbwJ9414n8lngH4AZwM3A9cAP47Y/ANwNfAP4PNAZ+MdL\nekcickF0GkVEWtOUBBM2/93d/8Xdn45r229m3wdWmdlcD55aPBB4z4On9AIcqO/s7kfNDOCEux85\nTwyZwN3ufhLAzJ4F7onb/h3gSXf/Tbj9foKHjolIkqjYEJHWtA6Y36TtBICZfRF4EBhB8BjtDIIH\n5vUBSoElwEozGwMUAqvdfe0lxHCgvtAIlQJ5YQyXh/9efUGDu7uZvUPjx3mLSCvSaRQRaU1n3L2k\nyXLMzAYCrxI8yn0WMIbgNAkEBQfu/hrB6MZTQC7wqpk9fwkxfNzktaPfdSKR0g+giLSFsQRFxXfd\nfYO77wb6Ne3k7sfcfZm73w18HZhrZl3CzR8TjIZcsnDE4whxcz4sOD9zrjkgIvIp6TSKiLSmLmbW\np0lbLfABwYeb+83sZeBGgsmiDczsMeA9YCfB76Y7gL3uXh122Q9MMrO1BFe9VFxijD8Dvm9mu4E/\nAfcSTDr9yyUeT0TOQyMbItKavkTwRzt+2eLuxQQTM79H8Af+GwRXhcSrJrhqZBvwFtAdmBa3/Z+B\nicCHwJZPEeNTwDLgeWAjYMBvgapPcUwROQdz96hjEBGJlJltAd5094VRxyLSEek0ioiklXCy6i3A\nWoJ7bHwT+Ovwq4gkgYoNEUk3dcDXgJ8QnEr+E3Cru2+KNCqRDkynUURERCSpNEFUREREkkrFhoiI\niCSVig0RERFJKhUbIiIiklQqNkRERCSp/h/8Y/NXlW5aTAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2b53725701d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "def colorcode(value):\n",
    "    color = []\n",
    "    for ivalue in value:\n",
    "        if ivalue <= 0.5 :\n",
    "            icolor = 'green'\n",
    "        elif ivalue <= 1 :\n",
    "            icolor = 'orange'\n",
    "        else:\n",
    "            icolor = 'red'\n",
    "        color.append(icolor)\n",
    "    return color   \n",
    "\n",
    "fig, ax1 = plt.subplots(1, 1)\n",
    "\n",
    "# plot river boundary\n",
    "plt.plot(river_geometry['x'], river_geometry['y'], color = 'black', lw=1, alpha = 1)\n",
    "\n",
    "# plot hanford shapefile\n",
    "hf_shape=gpd.read_file(fname_hanford_shp)\n",
    "hf_shape.plot(facecolor = 'none', edgecolor = 'black', ax=ax1, alpha=1)\n",
    "\n",
    "\n",
    "# plot all wells\n",
    "plt.scatter(ini_head_wells['Easting'], ini_head_wells['Northing'], color = 'gray', facecolors='none', s=5)\n",
    "\n",
    "\n",
    "plt.scatter(RMSE_df_sub['Easting'].values, RMSE_df_sub['Northing'].values, \n",
    "            color= colorcode(RMSE_df_sub.RMSE), facecolors='none', s=5)\n",
    "\n",
    "\n",
    "\n",
    "# # add label for points\n",
    "# for i, txt in enumerate(RMSE_df_sub.index):\n",
    "#     ax1.annotate(txt, (RMSE_df_sub['Easting'].values[i], RMSE_df_sub['Northing'].values[i]))\n",
    "\n",
    "ax1.set_xlim([550000, 600000])\n",
    "ax1.set_ylabel('Northing', fontsize = 14)\n",
    "ax1.set_xlabel('Easting', fontsize = 14)\n",
    "# ax1.set_title('RMSE b/w obs.head and simu.head', fontweight = 'bold', fontsize = 14)\n",
    "ax1.set_aspect(\"equal\", \"datalim\")\n",
    "\n",
    "legend_elements = [Line2D([0], [0], marker='o', color='w', label='0<RMSE<=0.5 ',\n",
    "                          markerfacecolor='None', markeredgecolor=\"green\", markersize=8),\n",
    "                   Line2D([0], [0], marker='o', color='w', label='0.5<RMSE<=1 ',\n",
    "                          markerfacecolor='None', markeredgecolor=\"orange\", markersize=8),\n",
    "                    Line2D([0], [0], marker='o', color='w', label='1<RMSE ',\n",
    "                          markerfacecolor='None', markeredgecolor=\"red\", markersize=8),\n",
    "                    Line2D([0], [0], marker='o', color='w', label='N/A',\n",
    "                          markerfacecolor='None', markeredgecolor=\"gray\", markersize=8),\n",
    "                  ]\n",
    "ax1.legend(handles=legend_elements, loc='best')\n",
    "\n",
    "# ax1.legend(handles=legend_elements, loc='best')\n",
    "fig.tight_layout()\n",
    "fig.set_size_inches(8, 8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-07T05:27:52.505525Z",
     "start_time": "2018-06-07T05:27:52.259764Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "fname = fig_simu_obs_wl + 'RMSE_map.jpg'\n",
    "fig.savefig(fname, dpi=300)\n",
    "plt.close(fig)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### interactive map with folium"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-10T05:51:27.556859Z",
     "start_time": "2018-06-10T05:51:27.549759Z"
    }
   },
   "outputs": [],
   "source": [
    "fname = fig_simu_obs_wl + 'RMSE.csv'\n",
    "RMSE_df = pd.read_csv(fname)\n",
    "RMSE_df_sub = RMSE_df[RMSE_df.RMSE < 10].dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-10T05:51:52.398085Z",
     "start_time": "2018-06-10T05:51:40.033043Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "HFR_map = folium.Map(location=[46.6, -119.5], tiles= 'Stamen Terrain', zoom_start= 10)\n",
    "\n",
    "df = RMSE_df_sub.copy()\n",
    "# df = df.head(1)\n",
    "fg = folium.FeatureGroup(name= 'RMSE b/w simu and obs WL')\n",
    "\n",
    "def colorcode(ivalue):\n",
    "    if ivalue <= 0.5 :\n",
    "        icolor = 'green'\n",
    "    elif ivalue <= 1 :\n",
    "        icolor = 'orange'\n",
    "    else:\n",
    "        icolor = 'red'\n",
    "    return icolor \n",
    "\n",
    "resolution, width, height = 75, 7, 3\n",
    "\n",
    "for lat, lon, name, value in zip(df['lat'], df['lon'], df['WellName'], df['RMSE']):\n",
    "    \n",
    "    iframe = IFrame('<img src=\"data:image/png;base64,{}\">'.format(base64.b64encode(open(\\\n",
    "                'figures/wl/Well {}_6h_ma_rs.png'.format(name), 'rb').read()).decode()), \\\n",
    "                width=1200, height=750)\n",
    "    fg.add_child(folium.CircleMarker(location=[lat, lon], \n",
    "#                                      popup=folium.Popup(name),\n",
    "                                     popup = folium.Popup(iframe, max_width=2500),\n",
    "#                                     radius = (value - df['median'].min())/(df['median'].max()-df['median'].min())*10,\n",
    "                                     radius = 5,\n",
    "                                     color = colorcode(value),\n",
    "#                                     fill = False,\n",
    "#                                     fill_color = 'crimson',\n",
    "                                    ))  \n",
    "## add legend to map\n",
    "url = \"https://github.com/pshuai88/notebook/raw/master/figures/RMSE_legend.png\"\n",
    "FloatImage(url, bottom = 75, left= 75).add_to(HFR_map)\n",
    "\n",
    "# FloatImage(iframe, bottom = 75, left= 75).add_to(fg)\n",
    "    \n",
    "HFR_map.add_child(fg)\n",
    "HFR_map.add_child(folium.LayerControl())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-10T05:52:38.222292Z",
     "start_time": "2018-06-10T05:52:32.089077Z"
    }
   },
   "outputs": [],
   "source": [
    "HFR_map.save('figures/RMSE_popup.html')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## plot well tracer vs SpC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-08T23:30:41.107349Z",
     "start_time": "2018-06-08T23:30:40.542646Z"
    },
    "code_folding": []
   },
   "outputs": [],
   "source": [
    "\n",
    "selected_chemical = 'Specific Conductance'\n",
    "iobs_chemical = GW_chem[GW_chem.loc[:,'Chemical'] == selected_chemical].copy()\n",
    "iobs_chemical.rename(columns={'Concentration': 'Spc'}, inplace = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**export tracer to .csv file**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "well_list = ['199-H1-7', '199-H3-25', '199-H3-5', '199-H3-7']\n",
    "for iwell in well_list[0:]:\n",
    "    print(iwell)\n",
    "    itracer_df = well_data_df[iwell, 'Total Tracer [M]', 'flux_ave_tracer'].copy()\n",
    "    idf = pd.DataFrame({'DateTime':real_time_6h, 'tracer_conc':itracer_df.values})\n",
    "    fname = out_dir + iwell + '_tracer.csv'\n",
    "    idf.to_csv(fname)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-10T05:12:05.576437Z",
     "start_time": "2018-06-10T05:02:45.124870Z"
    },
    "code_folding": [],
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline \n",
    "\n",
    "spc_map_well = []\n",
    "for iwell in obs_well_names[:]:\n",
    "#     iwell = '399-3-9'\n",
    "#     iwell = '399-2-33'\n",
    "    print(iwell)\n",
    "    iobs_data_df = pd.DataFrame()\n",
    "    \n",
    "    itracer_df = well_data_df[iwell, 'Total Tracer [M]', 'flux_ave_tracer'].copy()\n",
    "    \n",
    "    # read obs.head\n",
    "    if iwell in SFA_well['WellName'].unique():\n",
    "        iobs_data_df = SFA_well[SFA_well.loc[:, 'WellName'] == iwell].copy()\n",
    "        iobs_data_df['Spc'] = iobs_data_df['Spc']*1000 # convert ms/cm to us/cm\n",
    "\n",
    "    elif iwell in iobs_chemical['WellName'].unique():\n",
    "        iobs_data_df = iobs_chemical[iobs_chemical.loc[:, 'WellName'] == iwell].copy() \n",
    "\n",
    "    else:\n",
    "        print('No spc data for well ' + iwell)\n",
    "        continue\n",
    "        \n",
    "    iobs_data_df = iobs_data_df.drop_duplicates('DateTime', keep='first') \\\n",
    "                                .sort_values(by = 'DateTime').set_index('DateTime', drop=False).sort_index()\n",
    "    # filter wells outside of dates\n",
    "    iobs_data_df = iobs_data_df[(iobs_data_df['DateTime'] >= datetime.date(2011, 1, 1)) \\\n",
    "                                & (iobs_data_df['DateTime'] <= datetime.date(2016, 1, 1))]   \n",
    "    \n",
    "    if iobs_data_df['Spc'].count() > 10 and iobs_data_df['Spc'].isnull().all() == False and itracer_df.max()>= 0.1:   \n",
    "        spc_map_well.append(iwell)\n",
    "        # find closest mass1 location\n",
    "        iwell_easting = well_screen.loc[iwell, 'Easting']\n",
    "        iwell_northing = well_screen.loc[iwell, 'Northing']\n",
    "        dist = np.sqrt((mass1_coord['Easting']- iwell_easting)**2 + (mass1_coord['Northing']- iwell_northing)**2)\n",
    "        mass_id = mass1_coord.loc[np.argmin(dist), 'Mass_ID']\n",
    "\n",
    "        datum_files = glob.glob(fname_river_bc_6h + \"Datum*.txt\")\n",
    "        datum_files = natsorted(datum_files) \n",
    "\n",
    "        for ifile in datum_files:\n",
    "            if '_' + str(mass_id) in ifile:\n",
    "                river_stage = pd.read_table(ifile, sep=' ', header=None, names=['time', 'x', 'y', 'wl'])\n",
    "            elif mass_id == 333:\n",
    "                river_stage = pd.read_table(datum_files[-1], sep=' ', header=None, names=['time', 'x', 'y', 'wl'])\n",
    "\n",
    "        river_stage.time = batch_delta_to_time(date_origin, river_stage.time.astype(np.float64), \"%Y-%m-%d %H:%M:%S\", \"seconds\")\n",
    "        river_stage['time'] = river_stage['time'].apply(pd.to_datetime)  \n",
    "        river_stage = river_stage[(river_stage['time'] >= datetime.date(2011, 1, 1)) \\\n",
    "                            & (river_stage['time'] <= datetime.date(2016, 1, 1))]   \n",
    "        \n",
    "        # find the mode of frequecy (in sec)\n",
    "        res = iobs_data_df['DateTime'].diff().dt.seconds.value_counts().index[0]\n",
    "        val = iobs_data_df['DateTime'].diff().dt.seconds.value_counts().values[0]\n",
    "        if res <= 86400 and val >=100 :    \n",
    "            style = '-k'\n",
    "        else:\n",
    "            style = '-ko'\n",
    "#     iobs_data_df['Spc'] = (iobs_data_df['Spc'] - spc_u_data['Spc'].min()) / (spc_u_data['Spc'].max() - spc_u_data['Spc'].min())\n",
    "#     iobs_data_df['Spc'] = (iobs_data_df['Spc'] - well_data['Spc'].min()) / (well_data['Spc'].max() - well_data['Spc'].min())\n",
    "\n",
    "\n",
    "        ## ============= plot tracer vs SpC===================##\n",
    "#         fig, ax1 = plt.subplots(1, 1)  # short hand\n",
    "        fig = plt.figure()\n",
    "        ax1 = fig.add_subplot(3, 1, 1)\n",
    "        \n",
    "#         river_stage.plot(x= 'time', y= 'wl', legend=False, ax=ax1, color = 'gray' , lw = 1, rot = 0, alpha = 0.5)\n",
    "        ax1.plot(river_stage.time, river_stage.wl, color = 'gray' , lw = 1, alpha = 0.5)        \n",
    "        ax1.set_xlim([real_time_6h[0], real_time_6h[-1]])\n",
    "        ax1.set_ylabel('River Stage (m)', color='k')\n",
    "#         ax1.legend()\n",
    "        \n",
    "        ax2 = fig.add_subplot(3, 1, (2,3), sharex = ax1)\n",
    "#         plt1 = iobs_data_df.plot(x='DateTime', y='Spc', style = style, lw = 1, ms = 3, \\\n",
    "#                                  mfc='None' , markeredgewidth= 0.5, color='k', legend=False, ax=ax2, rot=0)\n",
    "        ax2.plot(iobs_data_df.DateTime, iobs_data_df.Spc, style, lw = 1, ms = 3, \\\n",
    "                                 mfc='None' , markeredgewidth= 0.5, color='k')\n",
    "        ax2.set_xlim([real_time_6h[0], real_time_6h[-1]])\n",
    "\n",
    "        ax2.set_ylabel(r'SpC ($\\mu$s/cm)', color='k')\n",
    "        ax2.set_xlabel('')\n",
    "        ax2.tick_params('y', colors='k')\n",
    "#         ax2.legend(['SpC'], frameon=False,\n",
    "#                    loc='lower left', bbox_to_anchor=(0, 0.95))\n",
    "        \n",
    "\n",
    "        # create another y-scale using the same x axis\n",
    "        ax3 = ax2.twinx()\n",
    "        plt2 = ax3.plot(real_time_6h, itracer_df.values, 'r-', lw = 1)\n",
    "\n",
    "        \n",
    "        ax3.set_ylim([0, 1])\n",
    "#         ax2.set_xlim([real_time_6h[0], real_time_6h[-1]])\n",
    "        ax3.set_ylabel('tracer (-)', color='r')\n",
    "        ax3.tick_params('y', colors='r')\n",
    "#         ax3.legend(['tracer'], frameon=False,\n",
    "#                    loc='lower left', bbox_to_anchor=(0, 0.90))\n",
    "        \n",
    "        plt.setp(ax1.get_xticklabels(), visible=False)\n",
    "        plt.minorticks_off()\n",
    "        plt.suptitle('Well ' + iwell, fontweight = 'bold', fontsize = 14)\n",
    "        fig.tight_layout()\n",
    "        fname = fig_tracer_spc + 'Well ' + iwell + '_spc_10'\n",
    "        fig.set_size_inches(8, 5)\n",
    "        fig.savefig(fname, dpi=300)\n",
    "        plt.close(fig)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### show wells using folium"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-10T05:48:59.682682Z",
     "start_time": "2018-06-10T05:48:59.672523Z"
    }
   },
   "outputs": [],
   "source": [
    "df = pd.read_csv(fname_GW_spc_summary, index_col='WellName')\n",
    "\n",
    "df = df.loc[spc_map_well]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-10T05:49:08.571330Z",
     "start_time": "2018-06-10T05:49:01.265689Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "HFR_map = folium.Map(location=[46.6, -119.5], tiles= 'Stamen Terrain', zoom_start= 10)\n",
    "\n",
    "# df = df.head(1)\n",
    "fg = folium.FeatureGroup(name= 'Tracer vs SpC')\n",
    "\n",
    "def colorcode(ivalue):\n",
    "    if ivalue <= 100 :\n",
    "        icolor = 'green'\n",
    "    elif ivalue <= 300 :\n",
    "        icolor = 'orange'\n",
    "    elif ivalue <= 500 :\n",
    "        icolor = 'purple'\n",
    "    else:\n",
    "        icolor = 'red'\n",
    "    return icolor   \n",
    "\n",
    "resolution, width, height = 75, 7, 3\n",
    "## add popup image \n",
    "# png = 'figures/{}_map_legend.png'.format('nitrate')\n",
    "# encoded = base64.b64encode(open(png, 'rb').read()).decode()\n",
    "# html = '<img src=\"data:image/png;base64,{}\">'.format\n",
    "# iframe = IFrame(html(encoded), width=(width*resolution)+20, height=(height*resolution)+20)\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "for lat, lon, name, value in zip(df['latitude'], df['longitude'], df.index, df['median']):\n",
    "    \n",
    "    iframe = IFrame('<img src=\"data:image/png;base64,{}\">'.format(base64.b64encode(open(\\\n",
    "                'figures/spc/Well {}_spc_10.png'.format(name), 'rb').read()).decode()), \\\n",
    "                width=1200, height=750)\n",
    "    fg.add_child(folium.CircleMarker(location=[lat, lon], \n",
    "#                                      popup=folium.Popup(name),\n",
    "                                     popup = folium.Popup(iframe, max_width=2500),\n",
    "#                                     radius = (value - df['median'].min())/(df['median'].max()-df['median'].min())*10,\n",
    "                                     radius = 5,\n",
    "                                     color = colorcode(value),\n",
    "#                                     fill = False,\n",
    "#                                     fill_color = 'crimson',\n",
    "                                    ))  \n",
    "## add legend to map\n",
    "HFR_map.save('figures/RMSE_popup.html')\n",
    "url = \"https://github.com/pshuai88/notebook/raw/master/figures/spc_legend.png\"\n",
    "FloatImage(url, bottom = 75, left= 75).add_to(HFR_map)\n",
    "\n",
    "# FloatImage(iframe, bottom = 75, left= 75).add_to(fg)\n",
    "    \n",
    "HFR_map.add_child(fg)\n",
    "HFR_map.add_child(folium.LayerControl())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-10T05:51:14.383302Z",
     "start_time": "2018-06-10T05:51:10.759602Z"
    }
   },
   "outputs": [],
   "source": [
    "HFR_map.save('figures/spc_popup.html')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### show wells using mplleaflet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-31T21:19:27.235595Z",
     "start_time": "2018-05-31T21:19:27.082103Z"
    }
   },
   "outputs": [],
   "source": [
    "# %matplotlib inline\n",
    "\n",
    "ichemical = pd.read_csv(fname_GW_NO3_summary)\n",
    "medianConc = ichemical['median'].values /1000  # convert ug/L to mg/L\n",
    "\n",
    "geometry = [Point(xy) for xy in zip(ichemical.longitude, ichemical.latitude)]\n",
    "crs = {'init': 'epsg:4326'}\n",
    "geo_df = GeoDataFrame(ichemical, crs = crs, geometry = geometry)\n",
    "\n",
    "def colorcode(value):\n",
    "    color = []\n",
    "    for ivalue in value:\n",
    "        if ivalue <= 10 :\n",
    "            icolor = 'green'\n",
    "        elif ivalue <= 30 :\n",
    "            icolor = 'orange'\n",
    "        elif ivalue <= 50 :\n",
    "            icolor = 'purple'\n",
    "        else:\n",
    "            icolor = 'red'\n",
    "        color.append(icolor)\n",
    "    return color \n",
    "\n",
    "ax = geo_df.plot(column = 'median', color = colorcode(medianConc))\n",
    "\n",
    "# ax.legend(handles=legend_elements, loc='best')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-31T21:25:32.524616Z",
     "start_time": "2018-05-31T21:25:31.901048Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# tile option: cartodb_positron, esri_worldtopo, \n",
    "tiles = ['cartodb_positron', 'esri_aerial', 'esri_worldtopo', 'mapbox bright']\n",
    "mplleaflet.show(fig=ax.figure, crs=geo_df.crs, tiles = tiles[1], path= 'nitrate_map_aerial.html')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### show well location using Leaflet"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-22T16:34:07.479898Z",
     "start_time": "2018-05-22T16:34:07.451246Z"
    }
   },
   "source": [
    "create GEOjson file to load into leaflet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-22T16:31:31.151594Z",
     "start_time": "2018-05-22T16:31:31.019275Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "nitrate_coord_agg = pd.read_csv(fname_GW_NO3_coord)\n",
    "\n",
    "# create GEOjson file to load into leaflet\n",
    "json_result_string = nitrate_coord_agg.to_json(\n",
    "    orient='records', \n",
    "    double_precision=12,\n",
    "    date_format='iso'\n",
    ")\n",
    "json_result = json.loads(json_result_string)\n",
    "\n",
    "geojson = {\n",
    "    'type': 'FeatureCollection',\n",
    "    'features': []\n",
    "}\n",
    "for record in json_result:\n",
    "    geojson['features'].append({\n",
    "        'type': 'Feature',\n",
    "        'geometry': {\n",
    "            'type': 'Point',\n",
    "            'coordinates': [record['longitude'], record['latitude']],  # GeoJSON need coord in long and lat order\n",
    "        },\n",
    "        'properties': record,\n",
    "    })\n",
    "\n",
    "# save json file\n",
    "with open(fname_GW_NO3_geojson, 'w') as f:\n",
    "    f.write(json.dumps(geojson, indent=2))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "overlay wells on map"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-31T17:01:08.938561Z",
     "start_time": "2018-05-31T17:01:08.925687Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "m = Map(center=(46.6, -119.5), zoom=10, layout = dict(width = '600px', height = '600px'),  basemap=basemaps.Esri.WorldImagery)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-22T17:43:40.750706Z",
     "start_time": "2018-05-22T17:43:40.734967Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "label = ipyw.Label(layout=ipyw.Layout(width='100%'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-31T17:02:36.188944Z",
     "start_time": "2018-05-31T17:02:35.956289Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "with open(fname_GW_NO3_geojson, 'r') as f:\n",
    "    data = json.load(f)\n",
    "# for feature in data['features']:\n",
    "#     feature['properties']['style'] = {\n",
    "#         'color': 'grey',\n",
    "#         'weight': 1,\n",
    "#         'fillColor': 'grey',\n",
    "#         'fillOpacity': 0.5\n",
    "#     }\n",
    "layer = GeoJSON(data=data, name='Nitrate')\n",
    "\n",
    "def hover_handler(event=None, id=None, properties=None):\n",
    "    label.value = properties['WellName']\n",
    "\n",
    "layer.on_hover(hover_handler)\n",
    "\n",
    "m.add_layer(layer)\n",
    "# Adding the control# Addin \n",
    "m.add_control(LayersControl())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-31T17:04:09.120209Z",
     "start_time": "2018-05-31T17:04:09.118345Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# m"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## plot well simu vs obs head & tracer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-08T23:40:28.686786Z",
     "start_time": "2018-06-08T23:40:21.035472Z"
    }
   },
   "outputs": [],
   "source": [
    "# loading objects    \n",
    "with open(fname_well_data_df_pickle, \"rb\") as f:\n",
    "    well_data_df = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(fname_selected_wells_pk, \"rb\") as f:\n",
    "    selected_obs_wells = pickle.load(f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-08T23:41:05.618947Z",
     "start_time": "2018-06-08T23:41:05.449833Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "real_time_6h = []\n",
    "for itime in np.arange(len(well_data_df.index)):\n",
    "    itime_str = list(batch_delta_to_time(date_origin, [float(well_data_df.index[itime])], \"%Y-%m-%d %H:%M:%S\", \"hours\"))\n",
    "    itime_str = ''.join(itime_str)\n",
    "    real_time_6h.append(datetime.datetime.strptime(itime_str, \"%Y-%m-%d %H:%M:%S\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-08T23:30:41.107349Z",
     "start_time": "2018-06-08T23:30:40.542646Z"
    },
    "code_folding": []
   },
   "outputs": [],
   "source": [
    "# read spc data\n",
    "selected_chemical = 'Specific Conductance'\n",
    "iobs_chemical = GW_chem[GW_chem.loc[:,'Chemical'] == selected_chemical].copy()\n",
    "iobs_chemical.rename(columns={'Concentration': 'Spc'}, inplace = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-06-07T00:34:52.345433Z",
     "start_time": "2018-06-07T00:06:57.436784Z"
    },
    "code_folding": [],
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "199-N-27\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/global/homes/p/pshuai/.conda/envs/nersc/lib/python3.6/site-packages/ipykernel/kernelbase.py:399: PerformanceWarning: indexing past lexsort depth may impact performance.\n",
      "  user_expressions, allow_stdin)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "199-N-28\n",
      "199-N-29\n",
      "199-N-31\n",
      "199-N-32\n",
      "199-N-33\n",
      "199-N-34\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/global/homes/p/pshuai/.conda/envs/nersc/lib/python3.6/site-packages/numpy/core/fromnumeric.py:51: FutureWarning: 'argmin' is deprecated. Use 'idxmin' instead. The behavior of 'argmin' will be corrected to return the positional minimum in the future. Use 'series.values.argmin' to get the position of the minimum now.\n",
      "  return getattr(obj, method)(*args, **kwds)\n",
      "/global/homes/p/pshuai/.conda/envs/nersc/lib/python3.6/site-packages/pandas/core/indexing.py:1020: FutureWarning: \n",
      "Passing list-likes to .loc or [] with any missing label will raise\n",
      "KeyError in the future, you can use .reindex() as an alternative.\n",
      "\n",
      "See the documentation here:\n",
      "http://pandas.pydata.org/pandas-docs/stable/indexing.html#deprecate-loc-reindex-listlike\n",
      "  return getattr(section, self.name)[new_key]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "199-N-36\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/global/homes/p/pshuai/.conda/envs/nersc/lib/python3.6/site-packages/ipykernel/kernelbase.py:399: PerformanceWarning: indexing past lexsort depth may impact performance.\n",
      "  user_expressions, allow_stdin)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "199-N-37\n",
      "No spc data for well 199-N-37\n",
      "199-N-39\n",
      "199-N-40\n",
      "No spc data for well 199-N-40\n",
      "199-N-41\n",
      "199-N-42\n",
      "199-N-49\n",
      "199-N-50\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/global/homes/p/pshuai/.conda/envs/nersc/lib/python3.6/site-packages/numpy/core/fromnumeric.py:51: FutureWarning: 'argmin' is deprecated. Use 'idxmin' instead. The behavior of 'argmin' will be corrected to return the positional minimum in the future. Use 'series.values.argmin' to get the position of the minimum now.\n",
      "  return getattr(obj, method)(*args, **kwds)\n",
      "/global/homes/p/pshuai/.conda/envs/nersc/lib/python3.6/site-packages/pandas/core/indexing.py:1020: FutureWarning: \n",
      "Passing list-likes to .loc or [] with any missing label will raise\n",
      "KeyError in the future, you can use .reindex() as an alternative.\n",
      "\n",
      "See the documentation here:\n",
      "http://pandas.pydata.org/pandas-docs/stable/indexing.html#deprecate-loc-reindex-listlike\n",
      "  return getattr(section, self.name)[new_key]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "199-N-51\n",
      "199-N-52\n",
      "199-N-53\n",
      "No spc data for well 199-N-53\n",
      "199-N-54\n",
      "199-N-55\n",
      "199-N-56\n",
      "199-N-57\n",
      "199-N-58\n",
      "199-N-59\n",
      "199-N-60\n",
      "199-N-61\n",
      "199-N-62\n",
      "199-N-63\n",
      "No spc data for well 199-N-63\n",
      "199-N-64\n",
      "199-N-65\n",
      "No spc data for well 199-N-65\n",
      "199-N-66\n",
      "199-N-67\n",
      "199-N-69\n",
      "199-N-70\n",
      "199-N-71\n",
      "199-N-72\n",
      "199-N-73\n",
      "199-N-74\n",
      "199-N-75\n",
      "199-N-76\n",
      "399-1-15\n",
      "399-1-16B\n",
      "399-1-17B\n",
      "399-1-17C\n",
      "399-1-18A\n",
      "399-1-18B\n",
      "399-1-18C\n",
      "399-1-6\n",
      "399-1-8\n",
      "399-1-9\n",
      "399-3-10\n",
      "399-4-10\n",
      "399-4-11\n",
      "399-1-14A\n",
      "399-1-21B\n",
      "399-8-5A\n",
      "399-8-5B\n",
      "399-8-5C\n",
      "199-N-77\n",
      "199-N-81\n",
      "199-N-22\n",
      "199-N-30\n",
      "No spc data for well 199-N-30\n",
      "199-N-43\n",
      "199-N-47\n",
      "199-N-48\n",
      "399-1-10B\n",
      "399-1-13B\n",
      "399-1-14B\n",
      "399-3-11\n",
      "399-5-4B\n",
      "399-6-2\n",
      "699-37-E4\n",
      "199-K-109A\n",
      "199-K-110A\n",
      "199-K-111A\n",
      "199-K-106A\n",
      "199-K-107A\n",
      "199-K-108A\n",
      "199-N-103A\n",
      "199-N-104A\n",
      "199-N-105A\n",
      "199-N-106A\n",
      "199-D8-68\n",
      "199-D8-69\n",
      "199-D8-70\n",
      "199-D8-71\n",
      "199-H4-63\n",
      "199-H4-64\n",
      "199-H3-3\n",
      "199-H3-4\n",
      "199-H3-5\n",
      "199-K-112A\n",
      "199-K-113A\n",
      "199-K-114A\n",
      "199-K-115A\n",
      "199-K-116A\n",
      "199-K-117A\n",
      "199-K-118A\n",
      "199-K-119A\n",
      "199-K-120A\n",
      "199-K-121A\n",
      "199-K-122A\n",
      "199-K-123A\n",
      "199-K-124A\n",
      "199-D4-1\n",
      "199-D4-4\n",
      "199-D4-5\n",
      "199-D4-6\n",
      "199-D4-7\n",
      "199-D4-8\n",
      "199-D4-9\n",
      "199-D4-10\n",
      "199-D4-11\n",
      "199-D4-12\n",
      "199-D4-13\n",
      "199-D4-14\n",
      "199-D4-15\n",
      "199-D3-2\n",
      "199-D4-17\n",
      "No spc data for well 199-D4-17\n",
      "199-D4-18\n",
      "No spc data for well 199-D4-18\n",
      "199-K-125A\n",
      "199-D5-36\n",
      "199-D5-37\n",
      "199-D4-19\n",
      "199-D5-38\n",
      "199-D5-39\n",
      "199-D5-40\n",
      "199-D4-20\n",
      "199-D5-41\n",
      "199-D5-42\n",
      "199-D5-43\n",
      "199-D5-44\n",
      "199-D4-21\n",
      "199-H4-65\n",
      "199-K-126\n",
      "199-D4-22\n",
      "199-D4-23\n",
      "199-D4-25\n",
      "199-D4-26\n",
      "199-D4-27\n",
      "199-D4-31\n",
      "199-D4-32\n",
      "199-D4-34\n",
      "199-D4-35\n",
      "199-D4-36\n",
      "199-D4-38\n",
      "199-D4-39\n",
      "199-D2-8\n",
      "199-D4-48\n",
      "199-D4-62\n",
      "199-D4-68\n",
      "199-D4-78\n",
      "199-D4-80\n",
      "199-D4-81\n",
      "199-D3-3\n",
      "199-D4-82\n",
      "199-D3-4\n",
      "199-D4-83\n",
      "199-D4-84\n",
      "199-D4-85\n",
      "199-D4-86\n",
      "199-D8-72\n",
      "199-K-129\n",
      "199-K-130\n",
      "699-96-44\n",
      "No spc data for well 699-96-44\n",
      "699-96-45\n",
      "No spc data for well 699-96-45\n",
      "199-D5-93\n",
      "199-N-136\n",
      "199-N-138\n",
      "199-N-139\n",
      "199-N-146\n",
      "199-K-137\n",
      "199-K-138\n",
      "199-K-139\n",
      "199-K-140\n",
      "199-N-147\n",
      "199-K-141\n",
      "199-K-142\n",
      "199-K-143\n",
      "199-K-144\n",
      "199-K-145\n",
      "199-K-146\n",
      "199-K-147\n",
      "199-K-148\n",
      "199-K-149\n",
      "199-K-150\n",
      "199-K-151\n",
      "199-K-152\n",
      "199-K-153\n",
      "199-K-154\n",
      "199-K-156\n",
      "199-K-157\n",
      "199-D5-97\n",
      "199-D5-98\n",
      "199-D5-99\n",
      "199-D2-11\n",
      "199-D5-103\n",
      "199-D5-104\n",
      "199-K-158\n",
      "199-D5-107\n",
      "199-D5-108\n",
      "199-D5-109\n",
      "199-D5-110\n",
      "199-D5-114\n",
      "199-D5-115\n",
      "699-100-43B\n",
      "699-99-42B\n",
      "699-99-41\n",
      "699-99-44\n",
      "699-98-43\n",
      "699-97-41\n",
      "699-98-46\n",
      "699-97-45\n",
      "699-95-45\n",
      "699-97-48B\n",
      "699-95-51\n",
      "699-97-43B\n",
      "699-94-41\n",
      "699-101-45\n",
      "699-95-48\n",
      "699-96-52B\n",
      "699-98-51\n",
      "199-B8-7\n",
      "199-B8-8\n",
      "199-D5-119\n",
      "199-D5-120\n",
      "199-D5-121\n",
      "199-D5-122\n",
      "199-K-159\n",
      "199-K-160\n",
      "199-K-161\n",
      "199-K-162\n",
      "199-K-163\n",
      "199-N-159\n",
      "199-K-164\n",
      "199-D5-123\n",
      "199-D5-125\n",
      "199-D5-126\n",
      "199-K-165\n",
      "199-K-166\n",
      "199-K-168\n",
      "199-N-165\n",
      "199-K-169\n",
      "199-K-170\n",
      "199-K-171\n",
      "199-K-172\n",
      "199-F8-7\n",
      "199-K-173\n",
      "199-N-167\n",
      "199-N-169\n",
      "199-N-170\n",
      "199-N-171\n",
      "199-N-172\n",
      "199-N-173\n",
      "199-K-174\n",
      "199-K-175\n",
      "199-H3-25\n",
      "199-K-178\n",
      "199-K-179\n",
      "199-K-180\n",
      "199-N-229\n",
      "199-N-211\n",
      "199-N-201\n",
      "199-N-248\n",
      "199-N-268\n",
      "199-N-280\n",
      "199-N-298\n",
      "199-N-316\n",
      "199-N-332\n",
      "199-N-342\n",
      "199-N-349\n",
      "199-N-348\n",
      "199-N-347\n",
      "199-N-346\n",
      "199-N-350\n",
      "199-N-351\n",
      "199-N-352\n",
      "199-N-353\n",
      "199-N-354\n",
      "199-N-355\n",
      "199-N-356\n",
      "199-N-357\n",
      "199-N-358\n",
      "199-N-360\n",
      "199-N-361\n",
      "199-N-362\n",
      "199-N-363\n",
      "199-N-364\n",
      "199-N-365\n",
      "199-N-366\n",
      "199-N-367\n",
      "199-K-181\n",
      "199-K-182\n",
      "199-H1-43\n",
      "199-B5-5\n",
      "199-B3-50\n",
      "199-B5-6\n",
      "199-B8-9\n",
      "199-D3-5\n",
      "199-D5-133\n",
      "199-D5-132\n",
      "199-D6-3\n",
      "199-D5-134\n",
      "199-D5-141\n",
      "199-H3-6\n",
      "199-H3-7\n",
      "199-H6-3\n",
      "199-H6-4\n",
      "199-H1-7\n",
      "199-H2-1\n",
      "199-H3-9\n",
      "199-H3-10\n",
      "399-1-54\n",
      "399-1-55\n",
      "399-1-56\n",
      "399-1-57\n",
      "399-1-58\n",
      "399-6-3\n",
      "399-1-59\n",
      "399-4-15\n",
      "399-3-33\n",
      "199-B2-14\n",
      "199-K-183\n",
      "199-K-184\n",
      "199-K-185\n",
      "199-K-186\n",
      "199-K-187\n",
      "199-K-188\n",
      "199-K-189\n",
      "199-K-190\n",
      "199-K-191\n",
      "199-K-192\n",
      "199-K-193\n",
      "199-K-194\n",
      "199-K-195\n",
      "199-K-196\n",
      "199-K-197\n",
      "199-K-198\n",
      "199-K-199\n",
      "199-B2-15\n",
      "199-B2-16\n",
      "199-B3-51\n",
      "199-B4-14\n",
      "199-F5-52\n",
      "199-F5-53\n",
      "199-F5-54\n",
      "199-K-200\n",
      "199-K-201\n",
      "199-B3-52\n",
      "199-D8-101\n",
      "199-D5-142\n",
      "199-H4-84\n",
      "199-H4-83\n",
      "199-D5-140\n",
      "199-F5-55\n",
      "199-F5-56\n",
      "199-N-182\n",
      "199-N-183\n",
      "199-N-184\n",
      "199-N-186\n",
      "199-N-187\n",
      "199-N-188\n",
      "199-N-189\n",
      "399-6-5\n",
      "199-K-202\n",
      "199-K-203\n",
      "199-K-204\n",
      "199-K-205\n",
      "199-K-206\n",
      "199-K-207\n",
      "199-K-208\n",
      "199-D5-143\n",
      "199-D5-144\n",
      "199-H4-85\n",
      "199-H4-86\n",
      "199-D5-145\n",
      "199-D5-146\n",
      "199-D5-147\n",
      "199-D5-148\n",
      "199-B4-16\n",
      "199-B4-18\n",
      "199-B5-9\n",
      "199-B5-10\n",
      "199-B5-11\n",
      "199-B5-12\n",
      "199-B5-13\n",
      "199-B5-14\n",
      "199-H4-90\n",
      "199-H4-91\n",
      "199-D5-153\n",
      "199-D5-154\n",
      "699-97-60\n",
      "699-97-61\n",
      "199-K-221\n",
      "199-K-222\n",
      "399-1-67\n",
      "199-H6-7\n",
      "199-H6-8\n",
      "199-D5-155\n",
      "199-D5-156\n",
      "199-D5-157\n",
      "199-D5-158\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline \n",
    "\n",
    "RMSE_df = screen_300A_HEIS.copy()\n",
    "RMSE_df['RMSE'] = np.nan\n",
    "\n",
    "for iwell in selected_obs_wells[71:]:\n",
    "#     iwell = '699-99-41'\n",
    "#     iwell = '199-K-117A'\n",
    "    print(iwell)\n",
    "    #------------- read into tracer, spc--------------#\n",
    "    itracer_df = well_data_df[iwell, 'Total Tracer [M]', 'flux_ave_tracer'].copy()\n",
    "    # read obs.head\n",
    "    if iwell in SFA_well['WellName'].unique():\n",
    "        iobs_data_df = SFA_well[SFA_well.loc[:, 'WellName'] == iwell].copy()\n",
    "        iobs_data_df['Spc'] = iobs_data_df['Spc']*1000 # convert ms/cm to us/cm\n",
    "\n",
    "    elif iwell in iobs_chemical['WellName'].unique():\n",
    "        iobs_data_df = iobs_chemical[iobs_chemical.loc[:, 'WellName'] == iwell].copy() \n",
    "\n",
    "    else:\n",
    "        print('No spc data for well ' + iwell)\n",
    "        continue\n",
    "    iobs_data_df['DateTime'] = iobs_data_df['DateTime'].apply(pd.to_datetime)    \n",
    "    iobs_data_df = iobs_data_df.drop_duplicates('DateTime', keep='first') \\\n",
    "                                .sort_values(by = 'DateTime').set_index('DateTime', drop=False).sort_index()\n",
    "    # filter wells outside of dates\n",
    "    iobs_data_df = iobs_data_df[(iobs_data_df['DateTime'] >= '1/1/2011') \\\n",
    "                                & (iobs_data_df['DateTime'] <= '1/1/2016')]   \n",
    "    #------------- read into head --------------#\n",
    "#     iwell_screen_index = np.asarray(np.where(well_screen.index == iwell)).flatten()\n",
    "    iwell_interval_index = np.where((z >= well_screen.loc[iwell, 'SCREEN_ELEV_BOTTOM']) \\\n",
    "                                    & (z <= well_screen.loc[iwell, 'SCREEN_ELEV_TOP']))\n",
    "    iwell_interval_index = np.asarray(iwell_interval_index).flatten()\n",
    "\n",
    "    z_reshape = z[iwell_interval_index].reshape(1, len(iwell_interval_index))\n",
    "\n",
    "    isimu_head = well_data_df[iwell, 'Liquid Pressure'].copy()\n",
    "    isimu_head = (isimu_head - P_atm) / rho / g_const\n",
    "    # apply broadcasting rule\n",
    "    isimu_head_df = pd.DataFrame(isimu_head.values + z_reshape, columns=isimu_head.columns, index=isimu_head.index)\n",
    "    isimu_head_df = isimu_head_df.sort_index()\n",
    "    \n",
    "    selected_well_simu_head = isimu_head_df.iloc[:, 0].to_frame(name='simu_head') # choose the bottom screen interval as simu head\n",
    "    selected_well_simu_head['DateTime'] = real_time_6h\n",
    "    selected_well_simu_head.set_index('DateTime', inplace=True)\n",
    "    # read obs.head\n",
    "    if iwell in SFA_well['WellName'].unique():\n",
    "        iobs_head_df = SFA_well[ SFA_well.loc[:, 'WellName'] == iwell].copy()\n",
    "    elif iwell in HEIS_auto['WellName'].unique():\n",
    "        iobs_head_df = HEIS_auto[ HEIS_auto.loc[:, 'WellName'] == iwell].copy()\n",
    "    elif iwell in HEIS_manual['WellName'].unique():\n",
    "        iobs_head_df = HEIS_manual[ HEIS_manual.loc[:, 'WellName'] == iwell].copy()\n",
    "    else:\n",
    "        iobs_head_df = pd.DataFrame()\n",
    "        print('No well obs data available for' + iwell)\n",
    "        continue\n",
    "   \n",
    "    #     iobs_head_df = iobs_head_df.sort_index(axis=1)\n",
    "    iobs_head_df['DateTime'] = iobs_head_df['DateTime'].apply(pd.to_datetime)\n",
    "    iobs_head_df = iobs_head_df.drop_duplicates('DateTime', keep='first') \\\n",
    "                                .sort_values(by = 'DateTime').set_index('DateTime', drop=False)\n",
    "    # filter wells outside of dates\n",
    "    iobs_head_df = iobs_head_df[(iobs_head_df['DateTime'] >= '1/1/2011') \\\n",
    "                                & (iobs_head_df['DateTime'] <= '1/1/2016')]            \n",
    "    \n",
    "    if iobs_head_df['WL'].count() > 1000 and iobs_data_df['Spc'].count() > 0:           \n",
    "        # find closest mass1 location\n",
    "        iwell_easting = well_screen.loc[iwell, 'Easting']\n",
    "        iwell_northing = well_screen.loc[iwell, 'Northing']\n",
    "        dist = np.sqrt((mass1_coord['Easting']- iwell_easting)**2 + (mass1_coord['Northing']- iwell_northing)**2)\n",
    "        mass_id = mass1_coord.loc[np.argmin(dist), 'Mass_ID']\n",
    "\n",
    "        datum_files = glob.glob(fname_river_bc_6h + \"Datum*.txt\")\n",
    "#         datum_files = natsorted(datum_files) \n",
    "        datum_files = natural_sort(datum_files)\n",
    "    \n",
    "        for ifile in datum_files:\n",
    "            if '_' + str(mass_id) in ifile:\n",
    "                river_stage = pd.read_table(ifile, sep=' ', header=None, names=['time', 'x', 'y', 'wl'])\n",
    "            elif mass_id == 333:\n",
    "                river_stage = pd.read_table(datum_files[-1], sep=' ', header=None, names=['time', 'x', 'y', 'wl'])\n",
    "\n",
    "    #     idatum = [ifile if '_' + str(mass_id) in ifile else datum_files[-1] for ifile in datum_files]\n",
    "    #     river_stage = pd.read_table(idatum, sep=' ', header=None, names=['time', 'x', 'y', 'wl'])\n",
    "\n",
    "        river_stage.time = batch_delta_to_time(date_origin, river_stage.time.astype(np.float64), \"%Y-%m-%d %H:%M:%S\", \"seconds\")\n",
    "        river_stage['time'] = river_stage['time'].apply(pd.to_datetime)    \n",
    "        ##================== apply moving-average ========================##\n",
    "        # find the mode of frequecy (in sec)\n",
    "        res = iobs_head_df['DateTime'].diff().dt.seconds.value_counts().index[0]\n",
    "        # apply moving average to well with sampling freq.< 3600 sec\n",
    "        win_size = int(6*3600/res) # 6-hour moving average       \n",
    "        if res <= 3600 :\n",
    "            datetime_ind = pd.date_range(start='1/1/2011', end='1/1/2016', freq=str(res) + 'S')\n",
    "            df = pd.DataFrame(np.nan, index=datetime_ind, columns=['WL_nan'])\n",
    "            iobs_head_df = pd.concat([df, iobs_head_df], axis=1, join='outer', join_axes=[df.index])\n",
    "\n",
    "            iobs_head_df['mv_ave_6h'] = iobs_head_df['WL'].rolling(win_size).mean()\n",
    "            style = '-k'\n",
    "#         elif res == 3600 :\n",
    "#             datetime_ind = pd.date_range(start='1/1/2011', end='1/1/2016', freq=str(res) + 'S')\n",
    "#             df = pd.DataFrame(np.nan, index=datetime_ind, columns=['WL_nan'])\n",
    "#             iobs_head_df = pd.concat([df, iobs_head_df], axis=1, join='outer', join_axes=[df.index])        \n",
    "#             win_size = int(6*3600/res)\n",
    "#             iobs_head_df['mv_ave_6h'] = iobs_head_df['WL'].rolling(win_size).mean()\n",
    "#             style = '-k'\n",
    "            ##================= get RMSE====================##\n",
    "            iobs_head_df_6h = iobs_head_df.loc[real_time_6h, 'mv_ave_6h']\n",
    "\n",
    "            df = pd.concat([selected_well_simu_head, iobs_head_df_6h], axis=1).\\\n",
    "                    rename(columns = {'mv_ave_6h': 'obs_head'}).dropna(how = 'any')\n",
    "\n",
    "            iRMSE = math.sqrt(mean_squared_error(df.simu_head, df.obs_head))\n",
    "            \n",
    "            RMSE_df.loc[iwell, 'RMSE'] = iRMSE\n",
    "                \n",
    "        else :\n",
    "            iobs_head_df['mv_ave_6h'] = iobs_head_df['WL']\n",
    "            style = '-ko'\n",
    "            iRMSE = np.nan\n",
    "        \n",
    "        # find the mode of frequecy (in sec)\n",
    "        res = iobs_data_df['DateTime'].diff().dt.seconds.value_counts().index[0]\n",
    "        val = iobs_data_df['DateTime'].diff().dt.seconds.value_counts().values[0]\n",
    "        if res <= 86400 and val >=100 :    \n",
    "            style_spc = '-'\n",
    "        else:\n",
    "            style_spc = '-o'\n",
    "\n",
    " \n",
    "\n",
    "       ##================= plot simu vs obs head========================##\n",
    "\n",
    "        fig, [ax,ax2] = plt.subplots(2, 1, sharex=True) # short hand\n",
    "\n",
    "        #------------ plot (1,1)---------------#\n",
    "        river_stage.plot(x= 'time', y= 'wl', legend=False, ax=ax, color = 'gray' , lw = 1, rot = 0, alpha = 0.5)\n",
    "#         ax.plot(real_time_6h, selected_well_simu_head, 'r-', lw= 1)\n",
    "        selected_well_simu_head.plot(y='simu_head', ax=ax, style='-r', lw=1)\n",
    "        iobs_head_df.plot(x= 'DateTime', y= 'mv_ave_6h', legend=False, ax=ax, style=style ,\\\n",
    "                          markerfacecolor = \"None\", ms = 3, lw = 1, rot = 0)\n",
    "\n",
    "        ax.set_title('Well ' + iwell,  fontweight = 'bold', fontsize = 14)\n",
    "        # format the ticks\n",
    "        ax.xaxis.set_major_locator(mdates.YearLocator())\n",
    "        ax.xaxis.set_major_formatter(mdates.DateFormatter('%Y'))\n",
    "        ax.xaxis.set_tick_params('labelbottom')\n",
    "        plt.setp(ax.xaxis.get_majorticklabels(), rotation=0, ha = 'center' )\n",
    "        \n",
    "        ax.set_xlim('01/01/2011', '01/01/2016')\n",
    "#         ax.set_ylim([104.5, 108.5])\n",
    "        ax.set_ylabel('Water Level (m)', fontsize = 14)\n",
    "        ax.set_xlabel('')\n",
    "        ax.legend(['river stage', 'simu.head', 'obs.head'], frameon = False)\n",
    "        \n",
    "        if iRMSE != np.nan:\n",
    "            ax.text(0.5,0.9,'RMSE = {:.2}'.format(iRMSE), transform = ax.transAxes, fontsize = 14)\n",
    "#         fig.autofmt_xdate()\n",
    "\n",
    "        #------------ plot (2,1)---------------#\n",
    "        # plot simu. tracer vs spc\n",
    "        ax2.plot(iobs_data_df.DateTime, iobs_data_df.Spc, style_spc, lw = 1, ms = 5, \\\n",
    "                                 mfc='none' , markeredgewidth= 0.5, color='b')\n",
    "        \n",
    "#         ax2.set_xlim([real_time_6h[0], real_time_6h[-1]])\n",
    "\n",
    "        ax2.set_ylabel(r'SpC ($\\mu$s/cm)', color='b', fontsize = 14)\n",
    "        ax2.set_xlabel('')\n",
    "        ax2.tick_params('y', colors='b')\n",
    "        \n",
    "         # create another y-scale using the same x axis\n",
    "        ax3 = ax2.twinx()\n",
    "        ax3.plot(real_time_6h, itracer_df.values, '-', color = 'darkorange', lw = 1)\n",
    "        \n",
    "        ax3.set_ylim([0, 1])\n",
    "        ax2.set_xlim(['01/01/2011', '01/01/2016'])\n",
    "        ax3.set_ylabel('Tracer (-)', color='darkorange' , fontsize = 14)\n",
    "        ax3.tick_params('y', colors='darkorange')\n",
    "        \n",
    "#         plt.setp(ax1.get_xticklabels(), visible=False)\n",
    "        plt.minorticks_off()\n",
    "\n",
    "        fig.tight_layout()\n",
    "        fig.set_size_inches(8, 8)\n",
    "        \n",
    "\n",
    "        fname = out_dir + 'simu_obs_tracer_wl/Well_' + iwell + '_sup.jpg' \n",
    "        fig.savefig(fname , dpi=300)\n",
    "        plt.close(fig)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# calculate simulated head using H5 output\n",
    "We can calculate total head by $$H = h + z, (h = (P-P_{atm})/\\gamma/g)$$  \n",
    "which $h$ has shape of **(100, 300, 300)**, and $z$ has shape of **(100,)**. To use broadcasting rule, the dimension of two arrays must match or either one has $dim = 1$. In this case, $z$ is reshaped to **(100, 1, 1)**. that is $$ H(100, 300, 300) = h(100,300,300) + z(100, 1, 1) $$\n",
    "\n",
    "\n",
    "note: when importing hdf5 dataset of dimension **(nx, ny, nz)**, the shape of numpy array will be **(nx, ny, nz)**. The correct logical shape in numpy would need to change to **(nz, ny, nx)** by using `np.swapaxes(0, 2)`."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Using interpolation to get well simu data at given location\n",
    "notes about using scipy.interpolate:\n",
    "* **`interpolate.Rbf`** performs best for irregular dataset compared to `interpolate.interp2d` and `interpolate.griddata` and can do both interpolation and extrapolation, however, it is **extremely slow** for large dataset. Note: __`.Rbf` method could not have duplicate in points, otherwise it will return sigular matrix__ (see this [post](https://stackoverflow.com/questions/37872171/how-can-i-perform-two-dimensional-interpolation-using-scipy))\n",
    "* **`interpolate.RectBivariateSpline`** is fast for structured grid, however, it will return NaN if there is missing data in the inputs. A work around is to replace those missing data with 0 or -9999. After interpolation, set those missing data to NaN again. (see this [post](https://stackoverflow.com/questions/15485343/bivariate-structured-interpolation-of-large-array-with-nan-values-or-mask)) (_note: this method can create some erratic numbers (negative concentration)_)\n",
    "* **`interpolate.griddata`** interpolate scatter points to a regular grid. It does not perform extrapolation. Need to test its performance against large dataset.\n",
    "\n",
    "**note:** this method won't work where there are inactive cells (NA) surrounding wells. In this case, river shore wells yield underestimated water level using interpolation."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Initialize arrays...\n",
    "Be careful using `np.empty()`, it will generate random number in the array."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-18T23:35:34.362804Z",
     "start_time": "2018-05-18T23:35:34.358512Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "well_simu_head = np.zeros((len(time_index), nz, len(well_screen.index))) # initialize array to store simu.head at obs.wells\n",
    "well_xflux = np.zeros((len(time_index), nz, len(well_screen.index))) # initialize array to store abs.flux at obs.wells\n",
    "well_yflux = np.zeros((len(time_index), nz, len(well_screen.index))) # initialize array to store abs.flux at obs.wells\n",
    "well_ave_tracer = np.zeros((len(time_index), len(well_screen.index))) # store flux-average tracer conc.\n",
    "well_tracer = np.zeros((len(time_index), nz, len(well_screen.index))) # initialize array to store tracer conc. at obs.wells\n",
    "\n",
    "# create index for well screen intervals range within model domain\n",
    "interval_index = np.where((z >= np.min(well_screen['SCREEN_ELEV_BOTTOM'])) \\\n",
    "                          & (z <= np.max(well_screen['SCREEN_ELEV_TOP'])))\n",
    "interval_index = np.asarray(interval_index).flatten()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-19T01:04:34.449611Z",
     "start_time": "2018-05-19T00:07:21.946088Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# loop over time step\n",
    "for itime in np.arange(len(time_index))[:]:\n",
    "#         ipdb.set_trace()\n",
    "    print(time_index[itime])\n",
    "    real_itime = batch_delta_to_time(date_origin, [float(time_index[itime][7:18])], \"%Y-%m-%d %H:%M:%S\", \"hours\")\n",
    "    real_itime = str(real_itime)\n",
    "    print(real_itime, type(real_itime))\n",
    "    # read pressure\n",
    "    temp_pressure = np.asarray(list(input_h5[time_index[itime]][\"Liquid_Pressure [Pa]\"])).swapaxes(0,2)\n",
    "    temp_pressure[temp_pressure == 0] = np.nan # replace 0 (inactive cell pressure) with nan to avoid calculations\n",
    "\n",
    "    # convert liquid_pressure to pressure head\n",
    "    temp_head = (temp_pressure - P_atm) / rho / g_const\n",
    "    # reshape z to use broadcasting rule\n",
    "    z_reshape = z.reshape(nz, 1, 1)\n",
    "    # total head  = presure head + elevation head\n",
    "    temp_total_head = temp_head + z_reshape\n",
    "\n",
    "    # create a copy of total_head and replace nan with 0\n",
    "    total_head_copy = temp_total_head.copy()\n",
    "    total_head_copy = np.nan_to_num(total_head_copy) # replace nan with 0 for interpolation\n",
    "    \n",
    "    # calculate ave. tracer conc. at wells\n",
    "    xflux = np.asarray(list(input_h5[time_index[itime]]['Liquid X-Flux Velocities'])).swapaxes(0,2) \n",
    "    yflux = np.asarray(list(input_h5[time_index[itime]]['Liquid Y-Flux Velocities'])).swapaxes(0,2)\n",
    "\n",
    "    \n",
    "    temp_tracer = np.asarray(list(input_h5[time_index[itime]]['Total_Tracer [M]'])).swapaxes(0,2)\n",
    "#     temp_tracer[temp_tracer == 0] = np.nan\n",
    "    # interpolate head at each iz layer\n",
    "    for iz in interval_index:\n",
    "        f_head = interpolate.RectBivariateSpline(y, x, total_head_copy[iz, :,:])\n",
    "        f_xflux = interpolate.RectBivariateSpline(y, x_grids[1:-1], xflux[iz, :,:]) #x-flux shape is different from y-flux shape\n",
    "        f_yflux = interpolate.RectBivariateSpline(y_grids[1:-1], x, yflux[iz, :,:])        \n",
    "        f_tracer = interpolate.RectBivariateSpline(y, x, temp_tracer[iz, :,:])\n",
    "#         f_tracer = interpolate.SmoothBivariateSpline(yy.flatten(), xx.flatten(), temp_tracer[iz, :,:].flatten())        \n",
    "        # evaluate at well points\n",
    "        well_simu_head[itime, iz, :] = f_head.ev(well_screen['Northing'], well_screen['Easting']) \n",
    "\n",
    "        well_xflux[itime, iz, :] = f_xflux.ev(well_screen['Northing'], well_screen['Easting'])\n",
    "        well_yflux[itime, iz, :] = f_yflux.ev(well_screen['Northing'], well_screen['Easting'])\n",
    "        \n",
    "        well_tracer[itime, iz, :] = f_tracer.ev(well_screen['Northing'], well_screen['Easting']) # has negative value?\n",
    "    \n",
    "    well_tracer[well_tracer < 0] = np.nan     \n",
    "    # calculate flux averageing tracer at each well\n",
    "    for iwell in well_screen.index[:]:\n",
    "    #     iwell = '399-1-10A'\n",
    "#         print(iwell)\n",
    "        iwell_abs_flux = []\n",
    "        iwell_screen_index = np.asarray(np.where(well_screen.index == iwell)).flatten()\n",
    "        iwell_interval_index = np.where((z > well_screen.loc[iwell, 'SCREEN_ELEV_BOTTOM']) \\\n",
    "                                        & (z < well_screen.loc[iwell, 'SCREEN_ELEV_TOP']))\n",
    "        iwell_interval_index = np.asarray(iwell_interval_index).flatten()\n",
    "        \n",
    "        iwell_abs_flux = np.sqrt(well_xflux[itime, iwell_interval_index, iwell_screen_index]**2 + \\\n",
    "                                well_yflux[itime, iwell_interval_index, iwell_screen_index]**2)\n",
    "        iwell_abs_flux = np.asarray(iwell_abs_flux).flatten()\n",
    "        well_ave_tracer[itime, iwell_screen_index] = np.sum(iwell_abs_flux*well_tracer[itime, iwell_interval_index, \\\n",
    "                                                                                       iwell_screen_index])/np.sum(iwell_abs_flux)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "dump files using `pickle.dump` and load files using `pickle.load`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-19T05:33:58.745198Z",
     "start_time": "2018-05-19T05:33:13.200713Z"
    },
    "code_folding": [],
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# saving objects\n",
    "with open(fname_well_simu_data_pk, \"wb\") as f:\n",
    "    pickle.dump((well_simu_head, well_xflux, well_yflux, well_tracer, well_ave_tracer), f)\n",
    "# # loading objects    \n",
    "# with open(fname_well_simu_data_pk, \"rb\") as f:\n",
    "#     well_simu_head, well_xflux, well_yflux, well_tracer, well_ave_tracer = pickle.load(f)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## plot well simu vs obs head"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-20T05:26:34.512921Z",
     "start_time": "2018-05-20T04:57:30.772851Z"
    },
    "code_folding": [],
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline \n",
    "for iwell in well_screen.index[17:]:\n",
    "#     iwell = '399-2-10'\n",
    "    print(iwell)\n",
    "    iwell_screen_index = np.asarray(np.where(well_screen.index == iwell)).flatten()\n",
    "    iwell_interval_index = np.where((z >= well_screen.loc[iwell, 'SCREEN_ELEV_BOTTOM']) \\\n",
    "                                    & (z <= well_screen.loc[iwell, 'SCREEN_ELEV_TOP']) \\\n",
    "                                   )\n",
    "    iwell_interval_index = np.asarray(iwell_interval_index).flatten()\n",
    "    # well head is taken at the lowest screen interval--iwell_interval_index[0]\n",
    "    \n",
    "    if iwell_interval_index.size > 0:\n",
    "        isimu_head = well_simu_head[:, iwell_interval_index[0], iwell_screen_index]\n",
    "        isimu_head = isimu_head.reshape(isimu_head.shape[0])\n",
    "    \n",
    "    # read obs.head\n",
    "    if iwell in SFA_well['WellName'].unique():\n",
    "        iobs_head_df = SFA_well[ SFA_well.loc[:, 'WellName'] == iwell].copy()\n",
    "        iobs_head_df['DateTime'] = iobs_head_df['DateTime'].apply(pd.to_datetime)\n",
    "    elif iwell in HEIS_auto['WellName'].unique():\n",
    "        iobs_head_df = HEIS_auto[ HEIS_auto.loc[:, 'WellName'] == iwell].copy()\n",
    "        iobs_head_df['DateTime'] = iobs_head_df['DateTime'].apply(pd.to_datetime)\n",
    "    elif iwell in HEIS_manual['WellName'].unique():\n",
    "        iobs_head_df = HEIS_manual[ HEIS_manual.loc[:, 'WellName'] == iwell].copy()\n",
    "        iobs_head_df['DateTime'] = iobs_head_df['DateTime'].apply(pd.to_datetime)\n",
    "    else:\n",
    "        iobs_head_df = pd.DataFrame()\n",
    "        print('No well obs data available for' + iwell)\n",
    "   \n",
    "    if iobs_head_df.size > 0:\n",
    "        ##================= plot simu vs obs head========================##\n",
    "        # fig = plt.figure() # open a canvas\n",
    "        # ax = fig.add_subplot(1, 1, 1) # create 1st subplot on a 1x1 grid\n",
    "        fig, ax = plt.subplots(1, 1) # short hand\n",
    "\n",
    "        # plt.plot(iobs_datetime.iloc[:], iobs_head_df['WL'], 'k-')\n",
    "        # plt.plot(iobs_head_df['DateTime'], iobs_head_df['WL'])\n",
    "\n",
    "\n",
    "        iobs_head_df.plot(x= 'DateTime', y= 'WL', color='k', legend=False, ax=ax, \n",
    "                         figsize = (8, 5), title = 'Well ' + iwell,\n",
    "                         rot = 0)\n",
    "        plt.plot(real_time_index, isimu_head, 'r-')\n",
    "\n",
    "        ax.set_xlim([real_time_index[0], real_time_index[-1]])\n",
    "#         ax.set_ylim([104.5, 108.5])\n",
    "        ax.set_ylabel('Water Level (m)')\n",
    "        ax.set_xlabel('')\n",
    "        ax.legend(['obs.head', 'simu.head'], frameon = False)\n",
    "\n",
    "        fig.tight_layout()\n",
    "        fname = fig_simu_obs_wl + 'Well ' + iwell \n",
    "        fig.savefig(fname , dpi=300)\n",
    "        plt.close(fig)\n",
    " "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## plot well tracer vs SpC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-20T07:13:27.042970Z",
     "start_time": "2018-05-20T06:54:26.077281Z"
    },
    "code_folding": [],
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline \n",
    "\n",
    "selected_chemical = 'Specific Conductance'\n",
    "iobs_chemical = GW_chem[GW_chem.loc[:,'Chemical'] == selected_chemical].copy()\n",
    "iobs_chemical.rename(columns={'Chemical': 'Spc'}, inplace = True)\n",
    "\n",
    "for iwell in well_screen.index[:]:\n",
    "#     iwell = '399-1-10A'\n",
    "    print(iwell)\n",
    "    iobs_data_df = pd.DataFrame()\n",
    "    iwell_screen_index = np.asarray(np.where(well_screen.index == iwell)).flatten()\n",
    "    \n",
    "    itracer = well_ave_tracer[:, iwell_screen_index]\n",
    "    \n",
    "    \n",
    "    # read obs.head\n",
    "#     iobs_data_df = spc_u_data[ spc_u_data.loc[:, 'WellName'] == iwell].copy()\n",
    "    if iwell in SFA_well['WellName'].unique():\n",
    "        iobs_data_df = SFA_well[SFA_well.loc[:, 'WellName'] == iwell].copy()\n",
    "        iobs_data_df['Spc'] = iobs_data_df['Spc']*1000\n",
    "        iobs_data_df['DateTime'] = iobs_data_df['DateTime'].apply(pd.to_datetime)\n",
    "    elif iwell in iobs_chemical['WellName'].unique():\n",
    "        iobs_data_df = iobs_chemical[iobs_chemical.loc[:, 'WellName'] == iwell].copy()        \n",
    "    else:\n",
    "        print('No spc data for well ' + iwell)\n",
    "\n",
    "#     iobs_data_df['Spc'] = (iobs_data_df['Spc'] - spc_u_data['Spc'].min()) / (spc_u_data['Spc'].max() - spc_u_data['Spc'].min())\n",
    "#     iobs_data_df['Spc'] = (iobs_data_df['Spc'] - SFA_well['Spc'].min()) / (SFA_well['Spc'].max() - SFA_well['Spc'].min())\n",
    "    if iobs_data_df.size > 0 and iobs_data_df['Spc'].isnull().all() == False:\n",
    "\n",
    "        ## ============= plot tracer vs SpC===================##\n",
    "        fig, ax1 = plt.subplots(1, 1)  # short hand\n",
    "\n",
    "        plt1 = iobs_data_df.plot(x='DateTime', y='Spc', linestyle='-',marker='o', \\\n",
    "                                 mfc='none' , markeredgewidth= 0.5, color='k', legend=False, ax=ax1, rot=0)\n",
    "\n",
    "    #         ax1.title('Well ' + iwell)\n",
    "        ax1.set_xlim([real_time_index[0], real_time_index[-1]])\n",
    "    #         ax1.set_ylim([0, 1])\n",
    "    #         ax.set_ylim([104.5, 108.5])\n",
    "        ax1.set_ylabel(r'SpC ($\\mu$s/cm)', color='k')\n",
    "        ax1.set_xlabel('')\n",
    "        ax1.tick_params('y', colors='k')\n",
    "        ax1.legend(['SpC'], frameon=False,\n",
    "                   loc='lower left', bbox_to_anchor=(0, 1))\n",
    "        plt.suptitle('Well' + iwell)\n",
    "\n",
    "        # create another y-scale using the same x axis\n",
    "        ax2 = ax1.twinx()\n",
    "        plt2 = ax2.plot(real_time_index, itracer, 'r-')\n",
    "        ax2.set_ylim([0, 1])\n",
    "        ax2.set_ylabel('tracer (-)', color='r')\n",
    "        ax2.tick_params('y', colors='r')\n",
    "        ax2.legend(['tracer'], frameon=False,\n",
    "                   loc='lower right', bbox_to_anchor=(1, 1))\n",
    "\n",
    "\n",
    "        fig.tight_layout()\n",
    "        fname = fig_tracer_spc + 'Well ' + iwell + '_spc'\n",
    "        fig.set_size_inches(8, 5)\n",
    "        fig.savefig(fname, dpi=300)\n",
    "        plt.close(fig)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## plot well tracer vs NO3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-21T22:43:13.762750Z",
     "start_time": "2018-05-21T22:43:12.869924Z"
    },
    "code_folding": [],
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "%matplotlib inline \n",
    "\n",
    "selected_chemical = 'Nitrate'\n",
    "iobs_chemical = GW_chem[GW_chem.loc[:,'Chemical'] == selected_chemical].copy()\n",
    "\n",
    "box_x = [577000, 581000]\n",
    "box_y = [149000, 153000]\n",
    "\n",
    "for iwell in iobs_chemical['WellName'].unique()[:]:\n",
    "#     iwell = '52-D'\n",
    "    print(iwell)\n",
    "    iobs_data_df = iobs_chemical[iobs_chemical.loc[:, 'WellName'] == iwell].copy()\n",
    "    if (iobs_data_df['EASTING'].unique() >= box_x[0] and \\\n",
    "                        iobs_data_df['EASTING'].unique() <= box_x[1] and \\\n",
    "                        iobs_data_df['NORTHING'].unique() >= box_y[0] and \\\n",
    "                        iobs_data_df['NORTHING'].unique() <= box_y[1]):\n",
    "        if iobs_data_df.size >0:\n",
    "            fig, ax1 = plt.subplots(1, 1)  # short hand\n",
    "\n",
    "            plt1 = iobs_data_df.plot(x='DateTime', y='Concentration', linestyle='-',marker='o', \\\n",
    "                                     mfc='none' , markeredgewidth= 0.5, color='k', legend=False, ax=ax1, rot=0)\n",
    "\n",
    "        #         ax1.title('Well ' + iwell)\n",
    "    #         ax1.set_xlim([real_time_index[0], real_time_index[-1]])\n",
    "        #         ax1.set_ylim([0, 1])\n",
    "        #         ax.set_ylim([104.5, 108.5])\n",
    "            unit = iobs_data_df['Unit'].unique()\n",
    "            ax1.set_ylabel('Nitrate ' + '(' + unit[0] + ')', color='k')\n",
    "            ax1.set_xlabel('')\n",
    "            ax1.tick_params('y', colors='k')\n",
    "            ax1.legend(['Nitrate'], frameon=False,\n",
    "                       loc='lower left', bbox_to_anchor=(0, 1))\n",
    "            plt.suptitle('Well ' + iwell)\n",
    "\n",
    "    #         # create another y-scale using the same x axis\n",
    "    #         ax2 = ax1.twinx()\n",
    "    #         plt2 = ax2.plot(real_time_index, itracer, 'r-')\n",
    "    #         ax2.set_ylim([0, 1])\n",
    "    #         ax2.set_ylabel('tracer (-)', color='r')\n",
    "    #         ax2.tick_params('y', colors='r')\n",
    "    #         ax2.legend(['tracer'], frameon=False,\n",
    "    #                    loc='lower right', bbox_to_anchor=(1, 1))\n",
    "\n",
    "\n",
    "            fig.tight_layout()\n",
    "            fname = fig_nitrate + 'Well ' + iwell + '_NO3'\n",
    "            fig.set_size_inches(8, 5)\n",
    "            fig.savefig(fname, dpi=300)\n",
    "            plt.close(fig)\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# plot groundwater level contour "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**note: for plotting contour surface using `plt.contour(X,Y,Z)`, Z must be shape of (ny, nx)**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-14T04:20:45.728829Z",
     "start_time": "2018-05-14T04:20:45.607686Z"
    }
   },
   "outputs": [],
   "source": [
    "iz = interval_index[-7]\n",
    "fig = plt.figure()\n",
    "gs = gridspec.GridSpec(1, 1)\n",
    "ax1 = fig.add_subplot(gs[0, 0])\n",
    "# plot contour\n",
    "cf1 = ax1.contourf(x, y, temp_total_head[iz, :,:],\n",
    "                   cmap=plt.cm.jet,\n",
    "                   vmin=100,\n",
    "                   vmax=130,\n",
    "                   extend=\"both\",\n",
    "                   levels=np.arange(100, 130.1, 1)\n",
    "                   )\n",
    "# plot contour lines\n",
    "cf2 = plt.contour(cf1, colors='grey', levels=np.arange(100, 130.1, 1))\n",
    "ax1.set_aspect(\"equal\", \"datalim\")\n",
    "\n",
    "plt.plot(well_screen['Easting'], well_screen['Northing'], 'ko')\n",
    "plt.clabel(cf2, inline = True, fmt = '%3.0d', fontsize = 10)\n",
    "# plt.xlim(42000, 43000)\n",
    "# plt.ylim(9000, 14000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2018-05-04T00:49:31.772431Z",
     "start_time": "2018-05-04T00:49:29.964670Z"
    },
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "        fig.savefig(fig_name, dpi=600, transparent=True)\n",
    "        plt.close(fig)\n",
    "input_h5.close()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python (widget)",
   "language": "python",
   "name": "widget"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.11"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {
    "height": "calc(100% - 180px)",
    "left": "10px",
    "top": "150px",
    "width": "290px"
   },
   "toc_section_display": true,
   "toc_window_display": true
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "oldHeight": 347,
   "position": {
    "height": "369px",
    "left": "3px",
    "right": "20px",
    "top": "691px",
    "width": "309px"
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "varInspector_section_display": "block",
   "window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}