xcaliber 6c03be54f3 feat(gitea): add diff-based updates and size delta gating to coder role
Enhancements:
- Added apply_diff() method for unified diff-based file updates
- Added commit_changes() with size delta quality gate (default 50%)
- Size delta gate blocks commits that exceed threshold to prevent data loss
- Auto-detects create vs replace operations
- Generates contextual commit messages from diffs
- Added _apply_unified_diff() helper for parsing and applying patches
- Added _generate_diff_commit_message() for auto-generating commit messages

Quality Gates:
- Files exceeding max_delta_percent (default 50%) are rejected
- Returns helpful guidance to use apply_diff() instead
- Prevents LLM accidental full file replacements

Refs: #11
2026-01-17 14:28:56 +00:00
2026-01-15 13:35:29 +00:00
2026-01-17 12:31:59 +00:00

Tools

This monorepo contains a collection of automation tools for Open WebUI, designed to facilitate integrations with Git repositories via Gitea and AI/ML platforms via Venice. These utilities enable enhanced functionality, such as automated repository management, AI-powered chat interactions, and image processing workflows.

Included Tools

gitea/

Python scripts for Git operations and repository management:

  • admin.py: Administrative utilities for managing Gitea repositories, potentially including user management, permissions, and batch operations.
  • dev.py: Development-focused tools for Git workflows, branch handling, and repository interactions tailored for software development processes.
  • coder.py: Development workflow role that reads tickets, creates branches with scope enforcement, generates commit messages with issue references, and creates pull requests.

venice/

Tools for interacting with Venice AI services:

  • chat.py: Scripts for AI chat interactions, enabling conversational interfaces and model-based responses.
  • image.py: Utilities for image processing and generation using AI models.
  • info.py: Information retrieval and data handling tools, possibly for querying AI models or processing related data.

These tools are modular and can be used independently or integrated into larger automation pipelines for Open WebUI.

Description
No description provided
Readme 460 KiB
Languages
Python 100%