flatai: A Rust repository from christianramsey

flatai is a Rust-based tool designed to equip a language model (LLM) with "just enough codebase awareness" to assist you in problem-solving. Currently, it's a brute force operation that works exclusively with Rust projects, but optimization are planned weekly.

flatai prepares your codebase for AI analysis by traversing a given directory and consolidating all .rs, and Cargo.toml files (py and js/ts coming soon) into a single XML-formatted text file. This approach simplifies the parsing and analysis process for AI tools, as they can process a single, comprehensive file instead of numerous smaller ones.

Please note that flatai is a work in progress. Future update will include ability to use flatai as an agent that you pass a question to and it regularly updates your contect file for use with LLMs (currently supporting anthropic's claude and openai's gpt).

Features to Add 🚀

Here are some features we're planning to add in the future:

Compress Your Repo: We're planning to add a feature that will compress your repository for easier handling and storage.
Optimize Based on Your Question: We're working on a feature that will take a question or problem as an argument and optimize the code based on your question and all files in the repository.
Support for JavaScript/TypeScript and Python: We're planning to add support for more languages, including JavaScript/TypeScript and Python.
Provide Just Enough Codebase Context: The output will be improved to provide just enough codebase context to answer your question.
Vector Index of Packages: We're planning to pull in the packages you are using and create a vector index of packages. This will allow you to reference specific parts of the question more easily.

Installation ⚙️

To install FLAT.AI, clone the repository and run cargo build --release to compile the Rust code. This will create a binary named flatai.

Use Executable

To use the included executable, simply download the binary file from the repository and move it to your desired directory.

Download the flatai binary from the repository.
Open your terminal and navigate to the directory where the binary is located.
Use the mv command to move the binary to your desired directory. For example:
```
mv flatai /desired/directory/
```
Replace /desired/directory/ with the actual path to your desired directory.

To add the flatai binary to your path, add the following line to your ~/.bashrc or ~/.zshrc file:

export PATH="$PATH:/path/to/your/project/target/release"

Replace /path/to/your/project with the actual path to your project. After adding this line, run source ~/.bashrc or source ~/.zshrc to apply the changes.

Now, you can run the flatai command from any directory in your terminal.

flatai .

flatai target/dir

Output File 📄

After running FLAT.AI, a file named flatten_files.txt will be created in the current directory. This file contains the consolidated code from all the .rs, .py, .js, and Cargo.toml files in the directory.

Here is an example screenshot of the flatten_files.txt file:

You can use this file for AI analysis and RAG (retrieval augmented generation).

Contact 📞