Triage is a Python framework for end-to-end machine learning workflows focused on temporal prediction problems, particularly in social good domains. It provides modular components for feature generation (architect), model training/evaluation (catwalk), experiment orchestration, model selection (audition), and post-modeling analysis. The system uses PostgreSQL for storing experiment metadata, feature matrices, and model predictions, with SQLAlchemy ORM and Alembic migrations managing the schema. Key entry points include the CLI (`triage experiment`, `triage audition`) and `manage.py` for development tasks. The architecture emphasizes reproducibility through configuration-driven experiments, temporal validation to prevent lookahead bias, and metadata tracking. While powerful, the system has complexity from multiple interdependent components and external service dependencies (PostgreSQL, Redis for RQ).
Octobox is a self-hosted, open-source web application for managing GitHub notifications. Built as a Ruby on Rails monolith (Rails 8.1, Ruby 3.x), it syncs notifications from the GitHub API into a PostgreSQL database, providing a rich UI for filtering, archiving, muting, starring, and searching notifications. Background jobs via Sidekiq handle syncing subjects, labels, comments, and performing actions like marking notifications as read on GitHub. Real-time updates are pushed via ActionCable. The application also exposes a REST API for programmatic access and supports GitHub App and OAuth authentication. It is designed for both hosted (octobox.io) and self-hosted deployments, with Docker, Heroku, and OpenShift support. The codebase is well-tested with Minitest and system tests, and follows standard Rails conventions with some custom patterns for search parsing and API client pagination.
This repository is an automated maintenance and triage system called 'Clownfish' (ProjectClownfish) that manages the large open-source `openclaw/openclaw` repository. It is a Node.js application driven by GitHub Actions workflows that continuously processes GitHub issues and pull requests: performing deduplication, canonical selection, security-sensitive detection, automated repair of contributor branches, exact-head merge preflight validation, and closeout of superseded/fixed-by-candidate items. The system operates in plan (read-only), autonomous (lightweight actions), and execute (mutating) modes, with strict permission boundaries defined in YAML frontmatter job files. It relies on the GitHub API, a Codex/AI worker for classification, and a suite of deterministic safety gates (re-fetch live state, check updated_at, block unsafe closeouts) before any mutation. The codebase is well-structured with clear separation of concerns (scripts, job definitions, results), but it is tightly coupled to the specific workflow of a single large repository, limiting direct reuse. Notable patterns include a robust job frontmatter schema, deterministic preflight validation, and a centralized security routing mechanism.