GitHub PR Comments Fetcher

A Python CLI tool to fetch GitHub PR comments and optionally generate Claude Code hooks for processing them.

Table of Contents

• Prerequisites
• Installation
• Usage
  • Basic Usage
  • With Claude Code Hooks
  • Examples
• What's in the box ?
• Testing
• Docker

Prerequisites

• Python >=3.11.0 (tested with 3.11.14 and 3.13.7)
• pre-commit >=3.2.0 <5.0.0 (tested with 4.3.0)
• uv >=0.8.17 (tested with 0.8.24)
• docker (optional)

Installation

1. Clone the git repository

   git clone https://github.com/arthrod/pr_to_task.git
   

2. Go into the project directory

   cd python-boilerplate/
   

3. Checkout working branch

   git checkout <branch>
   

4. Enable pre-commit hooks

   pre-commit install
   

5. Configure Python environment

   uv python install
   uv venv
   source .venv/bin/activate
   uv sync
   

Usage

The get-pr-comments CLI tool fetches GitHub PR comments and generates a JSONL file. It can optionally generate Claude Code hooks for processing comments one at a time.

Basic Usage

# Activate virtual environment
source .venv/bin/activate

# Fetch comments from a public repository
uv run get-pr-comments --project owner/repo --pr 123

# With authentication (for private repos or higher rate limits)
export GITHUB_TOKEN=your_token_here
uv run get-pr-comments --project owner/repo --pr 123

# Or provide token directly
uv run get-pr-comments --project owner/repo --pr 123 --token your_token_here

Command Options

• --user TEXT: GitHub username (default: "arthrod")
• --token TEXT: GitHub personal access token (optional for public repos)
• --project TEXT: GitHub repository in format "owner/repo" (required)
• --pr INTEGER: Pull request number (required)
• --since TEXT: Filter comments since date in ISO 8601 format (optional)
• --quantity INTEGER: Number of comments to fetch from most recent (optional)
• --output PATH: Output JSONL file path (default: "pr_comments.jsonl")
• --hooks: Generate Claude Code hooks configuration (default: false)

With Claude Code Hooks

Generate hooks to process comments automatically in Claude Code:

uv run get-pr-comments --project owner/repo --pr 123 --hooks

This creates:

• .claude/settings.json: Claude Code hooks configuration
• process_comments.py: Script to process comments one at a time

The SessionStart hook will automatically present comments to Claude one line at a time.

Examples

# Fetch last 10 comments from a PR
uv run get-pr-comments --project arthrod/my-repo --pr 42 --quantity 10

# Fetch comments since a specific date
uv run get-pr-comments --project arthrod/my-repo --pr 42 --since "2024-01-01T00:00:00Z"

# Generate hooks and custom output file
uv run get-pr-comments --project arthrod/my-repo --pr 42 --output my_comments.jsonl --hooks

# Get help
uv run get-pr-comments --help

Output Format

The tool generates a JSONL file where each line is a JSON object representing one complete comment:

{
  "comment_id": 123456,
  "comment_url": "https://github.com/owner/repo/pull/42#discussion_r123456",
  "author": "username",
  "created_at": "2024-01-01T12:00:00Z",
  "updated_at": "2024-01-01T12:00:00Z",
  "comment_type": "review",
  "path": "src/file.py",
  "position": 10,
  "line": 42,
  "text": "This is the full comment text.\nIt can contain multiple lines.\nAll preserved in a single JSON object."
}

What's in the box ?

uv

uv is an extremely fast Python package and project manager, written in Rust.

pyproject.toml file (pyproject.toml): orchestrate your project and its dependencies uv.lock file (uv.lock): ensure that the package versions are consistent for everyone working on your project

For more configuration options and details, see the configuration docs.

pre-commit

pre-commit is a framework for managing and maintaining multi-language pre-commit hooks.

.pre-commit-config.yaml file (.pre-commit-config.yaml): describes what repositories and hooks are installed

For more configuration options and details, see the configuration docs.

ruff

ruff is an extremely fast Python linter, written in Rust.

Rules are defined in the pyproject.toml.

For more configuration options and details, see the configuration docs.

mypy

mypy is an optional static type checker for Python that aims to combine the benefits of dynamic (or "duck") typing and static typing.

Rules are defined in the pyproject.toml.

For more configuration options and details, see the configuration docs.

bandit

bandit is a tool designed to find common security issues in Python code.

Rules are defined in the pyproject.toml.

For more configuration options and details, see the configuration docs.

docformatter

docformatter is a tool designed to format docstrings to follow PEP 257.

Options are defined in the .pre-commit-config.yaml.

Testing

We are using pytest & pytest-cov to write tests.

To run tests:

uv run pytest tests

collected 1 item

tests/test_python_boilerplate.py::test_hello_world PASSED

To run tests with coverage:

uv run pytest tests --cov=src

collected 1 item

tests/test_python_boilerplate.py::test_hello_world PASSED

---------- coverage: platform linux, python 3.13.7-final-0 -----------
Name                                 Stmts   Miss  Cover
--------------------------------------------------------
src/python_boilerplate/__init__.py       1      0   100%
src/python_boilerplate/main.py           6      2    67%
--------------------------------------------------------
TOTAL                                    7      2    71%

Docker

Build

To build the docker production image using Dockerfile:

docker build . -t my-python-application:latest

To build the docker development image using Dockerfile:

docker build . --target development -t my-python-application:dev

Run

To run the python app example inside Docker:

docker run -it --rm my-python-application:latest # or :dev for development

Hello World

Execute command inside container

docker run -it --rm my-python-application:latest bash