Merge pull request #33 from inbo/review

Upgrade to version 0.0.4. Ready for review in ropensci/software-review#263

.Rbuildignore

@@ -4,6 +4,7 @@
 ^_pkgdown.yml$
 ^appveyor\.yml$
 ^codemeta\.json$
+^.zenodo\.json$
 ^docs$
 ^man-roxygen$
 ^pkgdown$

.zenodo.json

@@ -0,0 +1,34 @@
+{
+  "creators": [
+    {
+      "affiliation": "Research Institute for Nature and Forest (INBO)",
+      "name": "Onkelinx, Thierry",
+      "orcid": "0000-0001-8804-4216"
+    }
+  ],
+  "contributors": [
+    {
+      "affiliation": "Research Institute for Nature and Forest (INBO)",
+      "name": "Onkelinx, Thiery",
+      "orcid": "0000-0001-8804-4216",
+      "type": ["Contactperson", "ProjectLeader"]
+    },
+    {
+      "affiliation": "Research Institute for Nature and Forest (INBO)",
+      "name": "Vanderhaeghe, Floris",
+      "orcid": "0000-0002-6378-6229",
+      "type": "Projectmember"
+    },
+    {
+      "name": "Research Institute for Nature and Forest (INBO)",
+      "type": "Rightsholder"
+    }
+  ],
+  "keywords": [
+    "r",
+    "version control",
+    "data.frame",
+    "plain text"
+  ],
+  "license": "GPL-3"
+}

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: git2rdata
 Title: Store and Retrieve Data.frames in a Git Repository
-Version: 0.0.3
+Version: 0.0.4
 Authors@R: c(
   person(
     "Thierry", "Onkelinx", role = c("aut", "cre"), 
@@ -14,7 +14,7 @@ Authors@R: c(
     "Research Institute for Nature and Forest",
     role = c("cph", "fnd"), email = "[email protected]"))
 Description: Make versioning of data.frame easy and efficient using git repositories.
-Depends: R (>= 3.4.0)
+Depends: R (>= 3.5.0)
 Imports:
   assertthat,
   git2r (>= 0.23.0),
@@ -36,12 +36,15 @@ BugReports: https://github.com/inbo/git2rdata/issues
 Collate:
     'clean_data_path.R'
     'git2rdata-package.R'
+    'write_vc.R'
+    'is_git2rdata.R'
+    'is_git2rmeta.R'
     'list_data.R'
     'meta.R'
-    'write_vc.R'
     'prune.R'
     'read_vc.R'
     'recent_commit.R'
     'reexport.R'
     'relabel.R'
+    'upgrade_data.R'
 VignetteBuilder: knitr

NAMESPACE

@@ -2,6 +2,12 @@
 
 S3method(format,meta_detail)
 S3method(format,meta_list)
+S3method(is_git2rdata,character)
+S3method(is_git2rdata,default)
+S3method(is_git2rdata,git_repository)
+S3method(is_git2rmeta,character)
+S3method(is_git2rmeta,default)
+S3method(is_git2rmeta,git_repository)
 S3method(list_data,character)
 S3method(list_data,default)
 S3method(list_data,git_repository)
@@ -22,17 +28,23 @@ S3method(prune_meta,git_repository)
 S3method(read_vc,character)
 S3method(read_vc,default)
 S3method(read_vc,git_repository)
+S3method(recent_commit,default)
 S3method(recent_commit,git_repository)
 S3method(relabel,data.frame)
 S3method(relabel,default)
 S3method(relabel,list)
 S3method(rm_data,character)
 S3method(rm_data,default)
 S3method(rm_data,git_repository)
+S3method(upgrade_data,character)
+S3method(upgrade_data,default)
+S3method(upgrade_data,git_repository)
 S3method(write_vc,character)
 S3method(write_vc,default)
 S3method(write_vc,git_repository)
 export(commit)
+export(is_git2rdata)
+export(is_git2rmeta)
 export(list_data)
 export(meta)
 export(prune_meta)
@@ -44,6 +56,7 @@ export(relabel)
 export(repository)
 export(rm_data)
 export(status)
+export(upgrade_data)
 export(write_vc)
 importFrom(assertthat,"on_failure<-")
 importFrom(assertthat,assert_that)
@@ -64,6 +77,7 @@ importFrom(git2r,status)
 importFrom(git2r,workdir)
 importFrom(methods,setOldClass)
 importFrom(stats,setNames)
+importFrom(utils,packageVersion)
 importFrom(utils,read.table)
 importFrom(utils,write.table)
 importFrom(yaml,as.yaml)

NEWS.md

@@ -1,47 +1,86 @@
+git2rdata 0.0.4 (2019-05-16)
+============================
+
+### BREAKING FEATURES
+
+  * `write_vc()` stores the `git2rdata` version number to the metadata. Use `upgrade_data()` to update existing data.
+
+### NEW FEATURES
+
+  * `read_vc()` checks the meta data hash. A mismatch results in an error.
+  * The meta data gains a data hash. A mismatch throws a warning when reading the object. This tolerates updating the data by other software, while informing the user that such change occurred.
+  * `is_git2rmeta()` validates metadata.
+  * `list_data()` lists files with valid metadata. 
+  * `rm_data()` and `prune_meta()` remove files with valid metadata. Other files are untouched.
+  *  Files with invalid metadata yield a warning with `list_data()`, `rm_data()` and `prune_meta()`.
+
+### Bugfixes
+
+  * `write_vc()` and `relabel()` handle empty strings (`''`) in characters and factors (#24).
+  * `read_vc()` no longer treats `#` as a comment character.
+  * `read_vc()` handles non ASCII characters on Windows.
+
+### Other changes
+
+  * Use a faster algorithm to detect duplicates (suggestion by @brodieG). 
+  * Improve documentation.
+  * Fix typo's in documentation, vignettes and README.
+  * Add a ROpenSci review badge to the README.
+  * The README mentions on upper bound on the size of dataframes.
+  * Set lifecycle to "maturing" and repo status to "active".
+  * The functions handle `root` containing regex expressions.
+  * Rework `vignette("workflow", package = "git2rdata")`.
+  * Update timings in `vignette("efficiency", package = "git2rdata")`
+  * Minor tweaks in `vignette("plain_text", package = "git2rdata")`
+
+git2rdata 0.0.3 (2019-03-12)
+============================
+
+  * Fix typo's in documentation, vignettes and README.
+
 git2rdata 0.0.2 (2019-02-26)
 ============================
 
 ### BREAKING CHANGES
 
-  * metadata is added as a list to the objects rather than in YAML format.
-  * the [yaml](https://cran.r-project.org/package=yaml) package is used to store the metadata list in YAML format.
-  * `write_vc()` now uses the 'strict' argument instead of 'override'
-  * the functionality `rm_data()` is split into `rm_data()` and `prune_meta()` (#9)
+  * `meta()` appends the metadata as a list to the objects rather than in YAML format.
+  * `yaml::write_yaml()` writes the metadata list in YAML format.
+  * `write_vc()` now uses the 'strict' argument instead of 'override'.
+  * `rm_data()` removes the data files. Use `prune_meta()` to remove left-over metadata files (#9).
 
 ### NEW FEATURES
 
-  * vignette on [efficiency](../articles/efficiency.html) added (#2)
-  * existing vignette was split over three vignettes
-      * focus on the [plain text format](../articles/plain_text.html)
-      * focus on [version control](../articles/version_control.html)
-      * focus on [workflows](../articles/workflow.html)
-  * S4 methods are replaced by S3 methods (#8)
-  * optimized factors use stable indices, resulting in smaller diffs when levels are added or removed (#13)
-  * use `relabel()` to alter factor levels without changing their index (#13)
-  * the raw data is written and read by base R functions instead of `readr` functions (#7)
-  * `write_vc()` and `read_vc()` use the current working directory as default root (#6, @florisvdh)
-  * the user can specify a string to code missing values (default = `NA`). This allows the storage of the character string `"NA"`.
+  * Vignette on [efficiency](../articles/efficiency.html) added (#2).
+  * Three separate vignettes instead of one large vignette.
+      * Focus on the [plain text format](../arsticles/plain_text.html).
+      * Focus on [version control](../articles/version_control.html).
+      * Focus on [workflows](../articles/workflow.html).
+  * S3 methods replace the old S4 methods (#8).
+  * Optimized factors use stable indices. Adding or removing levels result in smaller diffs (#13).
+  * Use `relabel()` to alter factor levels without changing their index (#13).
+  * `write.table()` stores the raw data instead of `readr::write_tsv()` (#7). This avoids the `readr` dependency.
+  * `write_vc()` and `read_vc()` use the current working directory as default root (#6, @florisvdh).
+  * The user can specify a string to code missing values (default = `NA`). This allows the storage of the character string `"NA"`.
   * `write_vc()` returns a list of issues which potentially result in large diffs.
-  * `list_data()` returns a vector with dataframes in the repository
+  * `list_data()` returns a vector with dataframes in the repository.
 
 ### Other changes
 
-  * `write_vc()` allows to use a custom NA string
-  * each helpfile contains a working example (#11)
-  * README updated (#12)
-      * Updated the rationale with links to the vignettes
-      * `git2rdata` has a hexsticker logo
-      * A DOI is added
-      * The installation instructions use `remotes` and build the vignettes
-  * `auto_commit()` was removed because of limited extra functionality over `git2r::commit()`
-  * dataframes are read and written by base R functions instead of `readr` functions
+  * `write_vc()` allows to use a custom `NA` string.
+  * Each helpfile contains a working example (#11).
+  * README updated (#12).
+      * Updated the rationale with links to the vignettes.
+      * `git2rdata` has a hexsticker logo.
+      * Add the [![DOI](https://zenodo.org/badge/147685405.svg)](https://zenodo.org/badge/latestdoi/147685405).
+      * The installation instructions use `remotes` and build the vignettes.
+  * We removed `auto_commit()` because of limited extra functionality over `git2r::commit()`.
 
 git2rdata 0.0.1 (2018-11-12)
 ============================
 
 ### NEW FEATURES
 
-  * use readr to write and read plain text files
-  * allows storage of strings with "NA" or special characters
-  * handle ordered factors
-  * stop handling complex numbers
+  * Use `readr` to write and read plain text files.
+  * Allow storage of strings with "NA" or special characters.
+  * Handle ordered factors.
+  * Stop handling complex numbers.

R/clean_data_path.R

@@ -1,21 +1,23 @@
-#' Clean the data path
+#' Clean the Data Path
 #' Strips any file extension from the path and adds the `".tsv"` and `".yml"`
 #' file extensions
 #' @inheritParams write_vc
-#' @param normalize normalize the path? Defaults to TRUE
-#' @return a named vector with "raw_file" and "meta_file", refering to the
-#' `".tsv"` and `".yml"` files
+#' @param normalize Normalize the path? Defaults to TRUE
+#' @return A named vector with "raw_file" and "meta_file", refering to the
+#' `".tsv"` and `".yml"` files.
 #' @noRd
 #' @family internal
+#' @importFrom assertthat assert_that is.flag noNA
 clean_data_path <- function(root, file, normalize = TRUE) {
+  assert_that(is.flag(normalize), noNA(normalize))
   dir_name <- dirname(file)
   file <- gsub("\\..*$", "", basename(file))
   if (dir_name == ".") {
     path <- file.path(root, file)
   } else {
     path <- file.path(root, dir_name, file)
   }
-  if (isTRUE(normalize)) {
+  if (normalize) {
     path <- normalizePath(path, winslash = "/", mustWork = FALSE)
   }
   c(raw_file = paste0(path, ".tsv"), meta_file = paste0(path, ".yml"))

R/is_git2rdata.R

@@ -0,0 +1,72 @@
+#' Check Whether a Git2rdata Object is Valid.
+#'
+#' A valid git2rdata object has valid metadata. The data hash must match the
+#' data hash stored in the metadata.
+#' @inheritParams write_vc
+#' @inheritParams is_git2rmeta
+#' @return A logical value. `TRUE` in case of a valid git2rdata object.
+#' Otherwise `FALSE`.
+#' @rdname is_git2rdata
+#' @export
+#' @family internal
+#' @template example-isgit2r
+is_git2rdata <- function(file, root = ".",
+                         message = c("none", "warning", "error")) {
+  UseMethod("is_git2rdata", root)
+}
+
+#' @export
+is_git2rdata.default <- function(file, root, message) {
+  stop("a 'root' of class ", class(root), " is not supported")
+}
+
+#' @export
+#' @importFrom assertthat assert_that is.string
+#' @importFrom yaml read_yaml as.yaml
+#' @importFrom utils packageVersion
+#' @importFrom git2r hash
+is_git2rdata.character <- function(file, root = ".",
+                                   message = c("none", "warning", "error")) {
+  assert_that(is.string(file), is.string(root))
+  message <- match.arg(message)
+  root <- normalizePath(root, winslash = "/", mustWork = TRUE)
+  check_meta <- is_git2rmeta(file = file, root = root, message = message)
+  if (!check_meta) {
+    return(FALSE)
+  }
+  file <- clean_data_path(root = root, file = file)
+
+  if (!file.exists(file["raw_file"])) {
+    msg <- "Data file missing."
+    switch(message, error = stop(msg), warning = warning(msg))
+    return(FALSE)
+  }
+
+  # read the metadata
+  meta_data <- read_yaml(file["meta_file"])
+
+  correct <- names(meta_data)
+  correct <- paste(correct[correct != "..generic"], collapse = "\t")
+  header <- readLines(file["raw_file"], n = 1, encoding = "UTF-8")
+  if (correct != header) {
+    msg <- paste("Corrupt data, incorrect header. Expecting:", correct)
+    switch(message, error = stop(msg), warning = warning(msg))
+    return(FALSE)
+  }
+
+  if (meta_data[["..generic"]][["data_hash"]] != hashfile(file[["raw_file"]])) {
+    msg <- "Corrupt data, mismatching data hash."
+    switch(message, error = stop(msg), warning = warning(msg))
+    return(FALSE)
+  }
+
+  return(TRUE)
+}
+
+#' @export
+#' @importFrom git2r workdir
+#' @include write_vc.R
+is_git2rdata.git_repository <- function(
+  file, root, message = c("none", "warning", "error")) {
+  is_git2rdata(file = file, root = workdir(root), message = message)
+}