Skip to main content

CLARA: Code Language Assistant & Repository Analyzer

Project description

CLARA: Code Language Assistant & Repository Analyzer 📜🔍🤖

Python 3.10 Code style: black

Clara is a tool to help developers understand and work with a code repository.

Note that creation of the vector database from the code is done only the first time you open the chat in the code repository. Subsequent chats will use the preloaded database, ensuring faster response times."

https://user-images.githubusercontent.com/538203/232823179-586ef7be-370c-4e65-8cf7-913d066ad2c3.mp4

This project is currently in its early stages of development and is considered a work in progress. You may encounter some issues, or incomplete features. We appreciate your understanding and patience as we continue to refine and enhance the project. Your feedback will help us improve and shape this project.

Overview

Clara is an AI-driven solution created to help developers effortlessly explore new or unfamiliar code repositories. It proves especially beneficial during the onboarding phase for new projects or when decoding legacy code.

Moving forward, Clara aims to offer assistance in various tasks, including documentation, auditing, and feature development, among others.

Features

  • Intelligent code and documentation analysis.
  • Integrated Database
    • Utilizes local storage through ChromaDB.
    • Maintains data persistence for individual code repositories.
    • Offers optional in-memory storage without persistence.
  • Context-aware short-term memory: Gathers information from ongoing conversations.

Install

With:

pipx install clara-ai

Or:

pip3 install clara-ai

Usage

Firstly, set an environment variable with your OpenAI API key:

export OPENAI_API_KEY="XXXXXX"

Then, use the command:

$ clara chat [PATH]

If the path is omitted then '.' will be used.

To exit use CTRL-D, or commands /quit or /exit.

All commands:

     ask
       Ask a question about the code from the command-line.

     chat
       Chat about the code.

     clean
       Delete vector DB for a given path.

     config
       Show config for a given path.

Chat commands

During chat you can also use this commands:

/context -- show the context for the last answer

/edit    -- open editor to edit the message

/quit
/exit    -- exit (you can use also CTRL-C or CTRL-D)

/help    -- show this message

Configuration

Run clara config to know from where the program is going to read the configuration. Usually this path is going to be /.config/clara/clara.yaml.

For now, there is only a couple of parameters. This is a sample configuration with the default values:

llm:
  model: gpt-3.5-turbo
index:
  # similarity or mmr
  search: similarity
  k: 5

Change the model for gpt-4 if you have access to it.

Cache

Vector DB and chat history are stored in a cache directory, per code analyzed. Use clara config to know the path to this directory.

You can remove manually this directory, if you want to refresh the data stored, or simply use the command clara clean.

If you want to chat with the code without reading/storing the vector DB (using the DB in memory), use the command clara [PATH] --memory-storage.

Roadmap

  • Short-term history
  • Configurable LLM
  • Agent
    • Access to filesystem
  • Features
    • Work with remote Git repositories
    • Document code with docstrings
    • Test creation
    • Audit code
    • Refactoring

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

clara_ai-0.0.10.tar.gz (11.6 kB view details)

Uploaded Source

Built Distribution

clara_ai-0.0.10-py3-none-any.whl (12.3 kB view details)

Uploaded Python 3

File details

Details for the file clara_ai-0.0.10.tar.gz.

File metadata

  • Download URL: clara_ai-0.0.10.tar.gz
  • Upload date:
  • Size: 11.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.5.1 CPython/3.10.12 Linux/5.15.0-1040-azure

File hashes

Hashes for clara_ai-0.0.10.tar.gz
Algorithm Hash digest
SHA256 50810ae816f5880f8f2b1093950f7c4cff6f0a35302c8340c0dab02b583d3eaa
MD5 4236950590d72421b34812f2bb4d6978
BLAKE2b-256 efa05bdd7e9a2cc1d7c9cc0f815f64ca131bbc920fe3ac6c927c382e90b5a3b9

See more details on using hashes here.

File details

Details for the file clara_ai-0.0.10-py3-none-any.whl.

File metadata

  • Download URL: clara_ai-0.0.10-py3-none-any.whl
  • Upload date:
  • Size: 12.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.5.1 CPython/3.10.12 Linux/5.15.0-1040-azure

File hashes

Hashes for clara_ai-0.0.10-py3-none-any.whl
Algorithm Hash digest
SHA256 49825941aa98980cbd5c67f4190e0883087ee93b1cca9d687072ed6fca2881e4
MD5 a4564890152039fd4488ca6555504df2
BLAKE2b-256 ee3096265cf9ef5f10c903885637f037ffc6f6568aee19acc79fa06c669f67a8

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page