MythEclipse/GMW
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

feat(gateway): Jev (System One) as primary text moderator with LLM fallback (#80)

Browse Source 
* feat(gateway): Jev (System One) as primary text moderator with LLM fallback

Add TypeSafe Jev via @typesafe-ai/sdk v0.6.0 as the PRIMARY analyzer for
text-only moderation sub-batches; the existing LLM stays as the fallback
for anything Jev cannot decide confidently (per-message gate rejection or
API failure) and for media batches.

- jevAnalyzer.ts: TypeSafeClient wrapper, declarative state builder
  (System One models MUST get factual state, not chat-XML — chat framing
  made Jev confidently wrong on clean messages at 0.98 confidence),
  message-id-keyed question builder (5 typed questions per message),
  cross-consistency acceptance gate (noul↔status↔severity↔action↔category),
  answer→AnalysisResult mapper, fail-open outcome.
- textBatchProcessor.ts: Jev-first per sub-batch, rejected ids + API
  failures fall back to callModerationLLM; no cross-batch pollution.
- config: AI_LLM_JEV_ENABLED/API_KEY/BASE_URL/MODEL/TIMEOUT_MS/MIN_CONFIDENCE.
- .env.example: documented all 6 Jev env vars.
- tests: unit (question/state builders, gate, mapper) + live smoke
  (gated behind AI_LLM_JEV_SMOKE=1) verified 4/4 accepted vs real 9router.

* fix: auto-fix code quality [skip ci]
This commit is contained in:
asepharyana
2026-09-23 17:51:38 +07:00
committed by GitHub
parent 5ba891ecfb
commit b9bba643bd
8 changed files with 1217 additions and 52 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.env.example
+6
View File
@@ -98,6 +98,12 @@ AI_LLM_IMAGE_MAX_DIMENSION=1024 # Max image dimension in pixels before r
AI_LLM_TEXT_BATCH_SIZE=20 # Max messages per text-only moderation batch (default: 20)
AI_LLM_MEDIA_ANALYSIS_TIMEOUT_MS=60000 # Timeout in ms for media analysis calls (default: 60000)
AI_LLM_TEXT_ANALYSIS_TIMEOUT_MS=30000 # Timeout in ms for text-only analysis calls (default: 30000)
AI_LLM_JEV_ENABLED=false # Use TypeSafe Jev (System One) as PRIMARY text analyzer; LLM is fallback
AI_LLM_JEV_API_KEY= # REQUIRED if AI_LLM_JEV_ENABLED=true. 9router/TypeSafe API key for /v1/systemone
AI_LLM_JEV_BASE_URL=http://127.0.0.1:4014 # 9router base URL (default: local 9router; prod: https://9router.asepharyana.my.id)
AI_LLM_JEV_MODEL=oc/jev-1.13-free # Jev model id (default: oc/jev-1.13-free)
AI_LLM_JEV_TIMEOUT_MS=45000 # Timeout in ms for a Jev systemone batch call (default: 45000)
AI_LLM_JEV_MIN_CONFIDENCE=0.9 # Min status-choice confidence to accept a Jev verdict (default: 0.9)
# === AI Analysis Tuning ===
AI_ANALYSIS_DEBOUNCE_MS=500 # Debounce window for batching messages in ms (default: 500)