Function History Retriever

This project is a small starter tool for exploring how a function evolved over time in a GitHub repository.

What it does

1. Clones or reuses a local checkout of a Git repository.
2. Retrieves commit history from that repository.
3. Filters commits to those likely related to a given function name.
4. Prints a compact timeline you can build on later.
5. Optionally asks an LLM to explain how the function evolved across those commits.

Project layout

• src/repo_history/git_tools.py: Git clone and log retrieval helpers.
• src/repo_history/function_history.py: Function-oriented commit filtering logic.
• src/repo_history/cli.py: Command-line entrypoint.

Quick start

python3 -m src.repo_history.cli \
  --repo-url https://github.com/pallets/flask.git \
  --function-name route

AI explanation mode

Set an API key and model first:

export OPENAI_API_KEY=your_api_key
export OPENAI_MODEL=your_model_id

Then run:

python3 -m src.repo_history.cli \
  --repo-url https://github.com/pallets/flask.git \
  --function-name route \
  --explain-with-ai

Useful flags:

• --show-patches: print full matching patches in the terminal
• --ai-model: override OPENAI_MODEL for one run
• --ai-base-url: point at an OpenAI-compatible API endpoint
• --ai-max-commits: limit how many matching commits are sent to the model
• --ai-max-patch-chars: truncate each patch before it is added to the prompt

Notes

This version uses git log -S to find commits where the exact function name text changed. That is a useful baseline, but it is not perfect:

• It may miss refactors where the function moved without the name text changing.
• It may include unrelated uses of the same name.
• It does not yet parse language syntax.
• The AI explanation is only as good as the commits surfaced by that search strategy.