b3e8a9a058
Major additions: - Community Next.js app (port 18187) for browsing claims with API docs - stemedb-chaos crate: Fault injection, chaos testing, CRDT properties - Latent ingestion system: Reddit/FDA ingesters with ADK-Go agents - Disputed claims handling: Manual review workflows and validation - Aphoria security scanner: New extractors (SQL injection, command injection, weak crypto, TLS version), policy-based ignores, UAT reports - Docker infrastructure: Dockerfile, docker-compose.yml for full stack - VulnBank demo: Intentionally vulnerable multi-language test corpus SDK & API enhancements: - Source registry handlers for tracking data provenance - Metrics endpoint - Skeptic filtering improvements Code quality: - Split 14 large files (>500 lines) into focused modules - All files now under 500-line limit per project guidelines Documentation: - Chaos testing guide, circuit breakers, observability docs - Phase 7 UAT documentation updates - Martin Kleppmann technical writer agent Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
1.2 KiB
1.2 KiB
Latent: Divergence Engine (Week 3)
This is the core logic of Latent. It implements the Skeptic Lens by comparing the "Official Truth" (Tier 0) against the "Latent Signal" (Tier 5).
Logic (The "Alpha")
The engine calculates a divergence_score (0.0 - 1.0) for every symptom cluster found on Reddit.
Divergence = \begin{cases}
0.1 \times Volume & \text{if present in FDA Label} \\
0.9 \times Volume \times Severity & \text{if absent from FDA Label}
\end{cases}
- High Divergence (>0.6): Means Reddit is screaming about something the FDA label does not mention. This is the Alpha.
- Low Divergence (<0.3): Means Reddit is complaining about nausea, and the FDA label says "Nausea is common." This is Noise.
Usage
- Prerequisites: You must have run
ingest-fdaandingest-redditfirst to generate the.jsonlgraph files. - Install:
pip install -r requirements.txt - Run:
python main.py
Output
Generates divergence_report.json.
[
{
"molecule": "semaglutide",
"signal": "gastroparesis",
"volume": 42,
"divergence_score": 0.85,
"status": "LATENT_SIGNAL",
"regulatory_status": "Silent/Absent"
}
]