feat(extraction): 컷오버 Phase 1 — mineru-service 를 마크다운 엔진으로 (marker 잔존)

mineru-service profile-gate 해제(상시 기동) + fastapi depends_on 추가 +
MARKER_ENDPOINT 을 mineru-service:3301 로 flip. marker-service 는 롤백 대비
Phase 2 까지 잔존(depends_on 유지, 호출만 안 됨 → idle-unload). 동일 /convert 계약.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This commit is contained in:
hyungi
2026-06-18 16:11:38 +09:00
parent bb929f88d0
commit 28b8afc748
+8 -5
View File
@@ -87,12 +87,11 @@ services:
start_period: 300s
restart: unless-stopped
# MinerU 2.5 VLM PDF→markdown 추출 — marker-service 대체 후보(단일카드 markdown VRAM ~10→~5GB).
# profile-gated: 기본 미기동 = marker 무영향. 활성 = `docker compose --profile mineru up -d mineru-service`.
# 컷오버(A/B 8게이트 PASS) 전까지 fastapi depends_on 에 넣지 않는다(격리). 포트 3301 (marker=3300).
# MinerU 2.5 VLM PDF→markdown 추출 — marker-service 대체(컷오버 2026-06-18, A/B 8/8 PASS).
# 단일카드 markdown VRAM ~10GB(marker)→~5.9GB 고정. fastapi 가 MARKER_ENDPOINT 로 호출.
# 동기 do_parse 버그 회피 위해 server.py 는 async aio_do_parse 사용. 포트 3301.
mineru-service:
build: ./services/mineru
profiles: ["mineru"]
ports:
- "127.0.0.1:3301:3301"
expose:
@@ -221,6 +220,9 @@ services:
condition: service_healthy
kordoc-service:
condition: service_healthy
# 컷오버: mineru-service 가 마크다운 엔진. marker-service 는 Phase 2 에서 제거(롤백 대비 잔존).
mineru-service:
condition: service_healthy
marker-service:
condition: service_healthy
env_file:
@@ -229,7 +231,8 @@ services:
- DATABASE_URL=postgresql+asyncpg://pkm:${POSTGRES_PASSWORD}@postgres:5432/pkm
- KORDOC_ENDPOINT=http://kordoc-service:3100
- OCR_ENDPOINT=http://ocr-service:3200
- MARKER_ENDPOINT=http://marker-service:3300
# ★ 컷오버 2026-06-18: marker-service:3300 → mineru-service:3301 (동일 /convert 계약).
- MARKER_ENDPOINT=http://mineru-service:3301
- MARKER_CONTAINER_PATH_PREFIX=/documents
# 2026-05-08 (D9 Track B revised): GPU stt-service 정식 승격, 내부 DNS 사용.
- STT_ENDPOINT=http://stt-service:3300