OSS Sentinel v0.2.0

Functional portfolio release with a reproducible CLI pipeline: ingest, process, enrich, analyze, and run. Includes fixture-driven offline execution, resilient GitHub ingestion, deterministic fake-provider quickstart, packaging, CI, SBOM, and verified checksums. Pain Index uses a 0–3 scale where higher means worse; no historical result values are claimed under the new formula.

v0.1.0 - MVP Release

v0.1.0: MVP - AI-Driven OSS Health Monitor
This release marks the initial MVP of OSS Sentinel, an automated engine designed to evaluate the operational health of open source ecosystems beyond vanity metrics.

What's Changed
Core Data Pipeline
Ingestion: Implemented automated fetching of issue data via GitHub API (PyGithub).
Processing: Robust normalization and cleaning of raw JSON data using Pandas.
Enrichment: Integration of OpenAI's GPT-4o-mini to perform semantic classification (Sentiment, Category, Urgency) on every issue.
Deep Diagnostic Analytics
Pain Index: Introduction of a proprietary metric (Sentiment × Urgency) to quantify "pain" in repositories.
Comparative Analysis: Initial benchmark analyzing the health of Apache Superset, Grafana, and Metabase.
Visualization: Automated generation of Heatmaps and Health Comparison bar charts.
Project Infrastructure
Architecture: Modular structure segregating Ingestion, Processing, Enrichment, and Analytics.
Assets: All diagnostic plots and reports generated and stored in assets/plots/.
Security: Proper .gitignore configuration to protect API keys and raw data volumes.
Installation
git clone https://github.com/cesaremcasa/oss-sentinel.gitcd oss-sentinelpip install -r requirements.txt
Usage
Set your API keys and run the orchestrator:
export GITHUB_TOKEN="..."
export OPENAI_API_KEY="..."
python main.py