Is this simulator still working?

[vmatter]

Hi everyone,

I would really like to use this simulator in my dissertation. However, from what I have tested so far, there are lots of errors. I tried to fix the TOKEN LIMIT EXCEEDED error and the static images, but honestly, no luck from my side.

Does anyone have a stable version of this simulator? Or do you guys know any other context simulator?

Best Regards

[LuisESV]

I run this like a month ago try it out https://github.com/drudilorenzo/generative_agents Luis E Silva Vargas

…

On Tue, Jan 7, 2025 at 9:27 PM vmatter ***@***.***> wrote: Hi everyone, I would really like to use this simulator in my dissertation. However, from what I have tested so far, there are lots of errors. I tried to fix the TOKEN LIMIT EXCEEDED error and the static images, but honestly, no luck from my side. Does anyone have a stable version of this simulator? Or do you guys know any other context simulator? Best Regards — Reply to this email directly, view it on GitHub <#177>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ACGZYDIL5YQAYQCBPEJJ77T2JSEJVAVCNFSM6AAAAABUY5PKC2VHI2DSMVQWIX3LMV43ASLTON2WKOZSG43TIMBZGQYDKOI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[ornery8x]

i dont think so,

i also tryed https://github.com/drudilorenzo/generative_agents
but couldnt get it to work either

please msg me if anyone has a working version !

[vmatter]

Hi @LuisESV and @ornery8x, I was able to reach the option's part after lots of fixes to run it on Windows, but it does not show anything:

And tons of issues when I execute the command run 1:

The steps that I executed so far to reach this part above on Windows:

Step 1 - In Anaconda Prompt:
conda create -n simulacra python=3.9.12 pip

Step 2 - In Anaconda Prompt:
conda activate simulacra

Step 3 - In Anaconda Prompt:
cd "<MY_PATH>\generative_agents-fix-and-improve" && pip install -r requirements.txt

Step 4 - Changed the run_frontend.sh file

#!/bin/bash

FRONTEND_SCRIPT_PATH="environment/frontend_server"
FRONTEND_SCRIPT_FILE="manage.py"
CONDA_ENV="simulacra"

FILE_NAME="Bash-Script-Frontend"
echo "(${FILE_NAME}): Running frontend server"

# Activate the Conda environment (Windows compatible)
source C:/Users/<MY_USER>/anaconda3/etc/profile.d/conda.sh
conda activate ${CONDA_ENV}

cd ${FRONTEND_SCRIPT_PATH}

PORT=8000
if [ -z "$1" ]; then
    echo "(${FILE_NAME}): No port provided. Using default port: ${PORT}"
else
    PORT=$1
fi

python ${FRONTEND_SCRIPT_FILE} runserver ${PORT}

# Pause to see errors
read -p "Press Enter to close..."

Step 5 - In Anaconda Prompt (separated prompt, needed for the front_end):
cd "<MY_PATH>\generative_agents-fix-and-improve" && run_frontend.sh

Step 6 - Changed run_backend.sh

#!/bin/bash

BACKEND_SCRIPT_PATH="reverie/backend_server"
BACKEND_SCRIPT_FILE="reverie.py"
LOGS_PATH="logs"

echo "Running backend server at: http://127.0.0.1:8000/simulator_home"
cd ${BACKEND_SCRIPT_PATH} || exit

timestamp=$(date +"%Y-%m-%d_%H-%M-%S")
echo "Timestamp: ${timestamp}"

# Ensure the logs directory exists
mkdir -p ${LOGS_PATH}

# Run the Python script directly using the Python executable from the simulacra environment
C:/Users/<MY_USER>/anaconda3/envs/simulacra/python ${BACKEND_SCRIPT_FILE} --origin ${1} --target ${2} | tee ${LOGS_PATH}/${2}_${timestamp}.txt

# Pause to see any errors
read -p "Press Enter to close..."

Step 7 - In Anaconda Prompt fixed issue "with init() got an unexpected keyword argument 'proxies'"
pip install "httpx<0.28"

Step 8 - Updated the strftime in "C:\Users**<MY_USER>\anaconda3\envs\simulacra\Lib\site-packages\openai_cost_logger\openai_cost_logger.py" since Windows doesn't allow colons (:) in file or folder names. You need to replace the colons in the timestamp used in filenames.**

Replace:
self.creation_datetime = strftime("%Y-%m-%d_%H:%M:%S")
With:
self.creation_datetime = strftime("%Y-%m-%d_%H-%M-%S")

Replace:
"creation_datetime": strftime("%Y-%m-%d %H:%M:%S"),
With:
self.creation_datetime = strftime("%Y-%m-%d_%H-%M-%S")

Replace:
"datetime": strftime("%Y-%m-%d %H:%M:%S"),
With:
"datetime": strftime("%Y-%m-%d %H-%M-%S"),

Step 9 - Changed run_backend_automatic.sh

#!/bin/bash

# Paths and file names
BACKEND_SCRIPT_PATH="reverie/backend_server"
BACKEND_SCRIPT_FILE="automatic_execution.py"
CONDA_ENV="simulacra"
LOGS_PATH="../../logs"

FILE_NAME="Bash-Script"

# Navigate to the backend script directory
cd ${BACKEND_SCRIPT_PATH} || { echo "Error: Cannot find ${BACKEND_SCRIPT_PATH}"; exit 1; }

# Activate the Conda environment (Windows version)
C:/Users/<MY_USER>/anaconda3/Scripts/activate ${CONDA_ENV}

# Initialize arguments
ARGS=""
while [[ $# -gt 0 ]]; do
    case "$1" in
        --origin|-o)
            ARGS="${ARGS} --origin ${2}"
            shift 2
            ;;
        --target|-t)
            ARGS="${ARGS} --target ${2}"
            TARGET=${2}
            shift 2
            ;;
        --steps|-s)
            ARGS="${ARGS} --steps ${2}"
            shift 2
            ;;
        --ui)
            ARGS="${ARGS} --ui ${2}"
            shift 2
            ;;
        --browser_path|-bp)
            ARGS="${ARGS} --browser_path ${2}"
            shift 2
            ;;
        --port|-p)
            ARGS="${ARGS} --port ${2}"
            echo "(${FILE_NAME}): Running backend server at: http://127.0.0.1:${2}/simulator_home"
            shift 2
            ;;
        *)
            echo "Unknown argument: $1"
            exit 1
            ;;
    esac
done

# Display arguments
echo "(${FILE_NAME}): Arguments: ${ARGS}"

# Generate a timestamp
timestamp=$(date +"%Y-%m-%d_%H-%M-%S")
echo "(${FILE_NAME}): Timestamp: ${timestamp}"

# Create the logs directory if it doesn't exist
mkdir -p ${LOGS_PATH}

# Run the backend script and save the output to a log file
python ${BACKEND_SCRIPT_FILE} ${ARGS} | tee ${LOGS_PATH}/${TARGET}_${timestamp}.txt

# Pause for the user to see any errors
read -p "Press Enter to close..."

Step 10 - In Anaconda Prompt:
cd "<MY_PATH>\generative_agents-fix-and-improve" && run_backend.sh base_the_ville_isabella_maria_klaus simulation-test

Now I am trying to find out why the rest is not working...

[vmatter]

Ok, I think I solved the issue below:
Error: Error code: 404 - {'error': {'message': 'The model gpt-35-turbo-0125 does not exist or you do not have access to it.', 'type': 'invalid_request_error', 'param': None, 'code': 'model_not_found'}}

I replaced every reference of "gpt-35-turbo-0125" with "gpt-3.5-turbo".
Just search all (ctrl+shift+f) in VS Code and replace them in the classes.

Now at least I cannot see any API errors anymore:

Now I just need to find out why the screen is blank:

[vmatter]

I found out that if you execute only once the url "http://localhost:8000/simulator_home" and never refresh it the image appears:

But if you refresh the page you face the issue "Please start the backend first.":

From what I checked, even with the code "running without errors" the sprites do not move, but the different contexts appear below. I must say that I must still check what is causing this issue.

Well, huge progress for today, at least the context generation of the simulator seems to be working.