PlantAI-rag-new

Files

T

History

adminandClaude Fable 5 a9a0ece4b4 feat(aqua): aqua extraction chain v16_aqua_1_0 (copy of v16_1_2, aquatized)

- entity validator: 66 aquatic entity types / 45 aquatic relation categories
  (fish/mollusk/crustacean/algae/water_quality/feed_ingredient/...)
- V16_SYSTEM_PROMPT rewritten for aquaculture; snake_case predicate rules
- rule seed vocab aquatized (60+ species incl. Litopenaeus vannamei,
  Oreochromis niloticus, Oncorhynchus mykiss, Salmo salar)
- cost hook on all 4 chat.completions.create (no-op by default)
- run_v16_aqua_one_paper.py: --md entry (skip MinerU) + --pdf (MinerU)
- run_v16_aqua_batch.py: workers + batch_progress.json resume
- pilot: 500 MD + 30 PDF extracted, 0 failures

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

2026-09-12 22:19:32 +08:00

api

chore(r4): 恢复 methodology_routes.py 执行位 755(backend2 合并引入的 mode 变更)

2026-09-11 14:23:29 +08:00

configs/rag

P4/M5: agent/legacy-rag 源码迁移(449 模块 191K 行)+ 影子 7911/7912 冒烟

2026-08-31 14:20:41 +08:00

env

fix(backend): r3 七项缺陷修复——图片预算/neo4j normalized/超时分流/mineru 加固/检索降级/DOCX 同源/docgen 异步任务化

2026-09-10 16:46:48 +08:00

prompts

P4/M5: agent/legacy-rag 源码迁移(449 模块 191K 行)+ 影子 7911/7912 冒烟