Skip to content

Multiagent simulated environment of LLM agents in a world simulated by other LLM agents

License

Notifications You must be signed in to change notification settings

PhySim/token_world

 
 

Repository files navigation

token_talkers

codecov CI

Usage

from token_talkers import BaseClass
from token_talkers import base_function

BaseClass().base_method()
base_function()
$ python -m token_talkers
#or
$ token_talkers

Setup

For instructions on setting up the project for development and contributions, see CONTRIBUTING.md

To set up the environment variables, create a .env file in the root directory of your project and add the following lines:

OPENAI_BASE_URL=http://192.168.1.199:11434/v1
OPENAI_API_KEY=your_openai_api_key_here

Alternatively, you can provide the --openai_base_url and --openai_api_key arguments when running the CLI:

$ python -m token_talkers --openai_base_url http://192.168.1.199:11434/v1 --openai_api_key your_openai_api_key_here
#or
$ token_talkers --openai_base_url http://192.168.1.199:11434/v1 --openai_api_key your_openai_api_key_here

File Index

""" Schema Documentation:

The SQLite database schema consists of two tables: hard_files and soft_files.

  1. hard_files Table:
  • path (TEXT PRIMARY KEY): The absolute path to the file.
  • size (INTEGER): The size of the file in bytes.
  • is_binary (BOOLEAN): A flag indicating whether the file is binary.
  • number_of_lines (INTEGER): The number of lines in the file (0 for binary files).
  • processed (BOOLEAN): A flag indicating whether the file has been processed.

This table stores metadata about the actual files present in the file system.

  1. soft_files Table:
  • path (TEXT PRIMARY KEY): The absolute path to the symbolic link.
  • hard_path (TEXT): The absolute path to the actual file that the symbolic link points to.
    • FOREIGN KEY(hard_path) REFERENCES hard_files(path)

This table stores metadata about symbolic links and their corresponding actual files.

The schema is used to index files and symbolic links in a directory, allowing for efficient querying and management of file metadata. """

Running File Index

To run the file_index.py script from the command line, use the following instructions:

  1. Ensure you have Python installed on your system.
  2. Clone an example repository to use as a codebase. You can use the provided setup_demo_repo.sh script to clone the requests repository:
$ ./setup_demo_repo.sh /path/to/destination_folder
  1. Run the file_index.py script to populate the file index database:
$ python file_index.py /path/to/destination_folder /path/to/database.db --wipe

This will index all files in the specified directory and store the information in the SQLite database.

Development

Read the CONTRIBUTING.md file.

About

Multiagent simulated environment of LLM agents in a world simulated by other LLM agents

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 91.9%
  • Makefile 6.8%
  • Shell 1.1%
  • Dockerfile 0.2%