_api_key() read ONLY `GEMINI_API_KEY`, but the canonical secret (host ~/.env and
Infisical SoT nautilus:/external-apis/gemini) is `GOOGLE_GEMINI_API_KEY`. The key
was present but under the canonical name → `_api_key()` raised "GEMINI_API_KEY אינו
מוגדר" on every call → ALL host precedent-metadata extraction via Gemini failed
silently (186 such errors in the legal-metadata-drain err log, latest 2026-06-14).
Fix: read GEMINI_API_KEY if set, else fall back to GOOGLE_GEMINI_API_KEY. No new
secret, no duplication — aligns the code to the existing SoT name (G1: fix at
source). Verified live: _api_key() resolves (len=53) and a real gemini query_json
call returns {"ok": true}.
Invariants: G1 (fix at source — code reads the canonical secret name, not a
parallel/duplicated env var) · X10 (deploy-env-secrets: single SoT name honored).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The /precedents metadata queue was stuck — 24 rows requested, nothing draining
them — and the agentic claude CLI hit error_max_turns on what is a single
structured text→JSON task (slow + flaky). Metadata extraction is bounded
extraction, the wrong fit for an agentic loop.
- gemini_session.py: query_json drop-in (gemini-2.5-flash, JSON mode, httpx —
no new SDK dep). Reads GEMINI_API_KEY (~/.env; SoT Infisical
nautilus:/external-apis/gemini). Host-side only — no LLM from the container.
- precedent_metadata_extractor: claude_session.query_json → gemini_session.
Validated live: rich, accurate fields (case_name/summary/appeal_subtype/tags).
- process_pending_extractions: kind-aware cooldown — metadata 2s (Gemini, fast),
halacha keeps 30s (Claude rate limits).
- drain_metadata_queue.py + legal-metadata-drain.config.cjs (pm2 cron */15) so
the queue never clogs again. SCRIPTS.md.
- X8 INV-FP5 updated: per-task engine choice (Gemini=bounded metadata,
claude_session=agentic halacha), both host-side, single canonical queue (G2).
Agentic/voice-sensitive work (writing, analysis, halacha) stays on claude_session
(Daphna's subscription). Gemini cost ≈ $0.10/1M tokens — negligible.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
