v1.3.0

What's New

Ollama Support (Local LLMs)

Run PR reviews completely offline with local models via Ollama. No API keys required - your code never leaves your machine.

wtp --switch-engine   # Select "ollama"
wtp --review <PR_URL>

Modular Architecture

Split the monolithic whatthepatch.py (2928 lines) into focused, maintainable modules:

Module	Purpose
whatthepatch.py	Main entry point
url_context.py	URL/file context handling
pr_providers.py	GitHub/Bitbucket PR fetching
commands.py	CLI command implementations
output.py	Output formatting (HTML/MD/TXT)
update.py	Self-update system

One-Line Installation

Install:

curl -sSL https://raw.githubusercontent.com/aaronmedina-dev/WhatThePatch/main/install.sh | bash

Uninstall:

curl -sSL https://raw.githubusercontent.com/aaronmedina-dev/WhatThePatch/main/uninstall.sh | bash

New Features

• Ollama engine - Run reviews with local LLMs (codellama, llama3, etc.)
• --edit-config command to open config.yaml in your editor
• manifest.json for self-update file management
• Manifest-based setup.py for cleaner installations

Bug Fixes

• Fixed handling of large PRs with >300 files (GitHub API pagination)

Improvements

• Removed unused pexpect dependency
• Cleaned up unused imports
• Better error handling in install scripts

Upgrading

wtp --update

Or reinstall:

curl -sSL https://raw.githubusercontent.com/aaronmedina-dev/WhatThePatch/main/install.sh | bash