hyungi/hyungi_document_server
1
1
Fork 0
Code Issues Pull Requests 2 Actions Packages Projects Releases Wiki Activity

refactor(board): 처리 머신 보드 나스+맥미니 2노드 재구성

Browse Source 
2026-07-02 컷오버 반영 — GPU 서버 퇴역, 맥북 night-drain 보류(06-29 결정).

- 레인 2개: 나스(추출/마크다운/청크·임베딩 등 DS 본체 Docker 스테이지),
  맥미니(분류/요약/심층분석 — 단일 생성 LLM 허브 + bge-m3/리랭크)
- summarize 풀 분리(summarize_by_machine·ai_model_version 조인 SQL) 제거
  — FE 유일 소비자 확인 후 응답 스키마에서 정리 (5쿼리 -> 4쿼리)
- 맥북 전제 UI 제거: 요약 오프로드 분담막대·요약 합류 칩·번다운 합류
  변곡점 마커·잠듦 문구·전역 스트립 맥북 칩(맥미니 칩으로 대체)
- deferred_pending = LLM 백오프 신호로 맥미니 카드 귀속 (기능 보존)
- 번다운 차트·정직 ETA·실패 드로어·백그라운드 작업 등 머신 무관 기능 보존
- background_jobs 머신 귀속 기본값 gpu -> nas
- 단위테스트 2노드 기준 재작성 (27 passed)

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
hyungi
2026-07-02 16:51:32 +09:00
parent 304a2b9c0f
commit b91b05e889
8 changed files with 130 additions and 364 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/test_queue_overview.py
+66 -156
View File
@@ -4,6 +4,8 @@ services/queue_overview 의 SQL 수집부와 분리된 순수 판정 함수
(stage_machine_map / build_machines / build_summarize_eta / build_trend /
build_totals / compute_eta_minutes / rows_to_* / display_title) 를
mock 행으로 검증한다. 통합(실 SQL)은 배포 후 라이브 smoke 로 확인.
2026-07-02 컷오버 후 2노드(나스+맥미니) 기준 — 구 3노드 레인은 제거됨.
"""
from datetime import datetime
@@ -18,7 +20,6 @@ from services.queue_overview import (
compute_eta_minutes,
display_title,
rows_to_stage_stats,
rows_to_summarize_split,
stage_machine_map,
)
@@ -36,186 +37,115 @@ def _stage(**kw) -> dict:
return base
def _split(macbook: dict | None = None, macmini: dict | None = None) -> dict:
"""summarize 풀 완료 실적 split — 미지정 0."""
zero = {"done_1h": 0, "done_today": 0, "done_15m": 0}
return {
"macbook": {**zero, **(macbook or {})},
"macmini": {**zero, **(macmini or {})},
}
def _machine(machines: list[dict], key: str) -> dict:
return next(m for m in machines if m["key"] == key)
# ─── stage→machine 귀속 맵 ────────────────────────────────────────────────────
def test_stage_machine_map_deep_enabled():
smap = stage_machine_map(deep_enabled=True)
def test_stage_machine_map_two_nodes():
smap = stage_machine_map()
for s in ("extract", "embed", "chunk", "markdown", "preview", "thumbnail", "fulltext", "stt"):
assert smap[s] == "gpu"
assert smap[s] == "nas"
assert smap["classify"] == "macmini"
assert smap["summarize"] == "macmini"
assert smap["deep_summary"] == "macbook"
def test_stage_machine_map_deep_disabled():
"""deep 슬롯 부재 시 deep_summary 도 macmini 귀속."""
smap = stage_machine_map(deep_enabled=False)
assert smap["deep_summary"] == "macmini"
# ─── 머신 카드 귀속 합산 ──────────────────────────────────────────────────────
def test_gpu_stage_counts_attribution():
def test_nas_stage_counts_attribution():
stats = {
"extract": _stage(pending=3, processing=1, done_1h=5, done_today=9, done_15m=1),
"stt": _stage(failed=2, done_1h=1, done_today=2),
}
machines = build_machines(stats, _split(), [], deep_enabled=True)
gpu = _machine(machines, "gpu")
assert (gpu["pending"], gpu["processing"], gpu["failed"]) == (3, 1, 2)
assert (gpu["done_1h"], gpu["done_today"]) == (6, 11)
# gpu 의 stages 는 정적 8종 전부 (집계 0 이어도 표시)
assert gpu["stages"] == [
machines = build_machines(stats, [])
nas = _machine(machines, "nas")
assert (nas["pending"], nas["processing"], nas["failed"]) == (3, 1, 2)
assert (nas["done_1h"], nas["done_today"]) == (6, 11)
# nas 의 stages 는 정적 8종 전부 (집계 0 이어도 표시)
assert nas["stages"] == [
"extract", "embed", "chunk", "markdown",
"preview", "thumbnail", "fulltext", "stt",
]
def test_summarize_pool_split_attribution():
"""summarize pending/failed = macmini 귀속, 완료 실적은 split 로 분리 —
stage-level summarize done 수치는 카드에 이중 합산되지 않는다."""
def test_macmini_llm_stages_attribution():
"""classify/summarize/deep_summary 전부 macmini 귀속 (단일 생성 LLM 허브)."""
stats = {
"classify": _stage(done_1h=2, done_today=3),
"summarize": _stage(pending=7, failed=1, done_1h=10, done_today=20),
"deep_summary": _stage(pending=2, processing=1, done_1h=3, done_today=4),
}
split = _split(macbook={"done_1h": 4, "done_today": 8}, macmini={"done_1h": 6, "done_today": 12})
machines = build_machines(stats, split, [], deep_enabled=True)
machines = build_machines(stats, [])
macmini = _machine(machines, "macmini")
macbook = _machine(machines, "macbook")
assert macmini["pending"] == 7 and macmini["failed"] == 1
assert macmini["done_1h"] == 2 + 6 # classify + macmini 몫 (10 아님)
assert macmini["done_today"] == 3 + 12
assert macbook["done_1h"] == 4 and macbook["done_today"] == 8
assert macbook["pending"] == 0 # 풀 pending 은 macmini 만
assert macmini["pending"] == 9 and macmini["failed"] == 1
assert macmini["processing"] == 1
assert macmini["done_1h"] == 2 + 10 + 3
assert macmini["done_today"] == 3 + 20 + 4
assert macmini["stages"] == ["classify", "summarize", "deep_summary"]
assert _machine(machines, "nas")["pending"] == 0
def test_summarize_by_machine_projection():
"""build_summarize_by_machine = split 의 done_1h/done_today 를 머신별로 투영
(done_15m 은 제외 — 내부 state 판정 전용)."""
from services.queue_overview import build_summarize_by_machine
split = _split(
macbook={"done_1h": 226, "done_today": 312, "done_15m": 60},
macmini={"done_1h": 37, "done_today": 94, "done_15m": 9},
)
sbm = build_summarize_by_machine(split)
assert sbm == {
"macmini": {"done_1h": 37, "done_today": 94},
"macbook": {"done_1h": 226, "done_today": 312},
}
assert "done_15m" not in sbm["macbook"]
def test_compose_overview_includes_summarize_by_machine():
"""compose_overview 응답 계약에 summarize_by_machine 포함 (FE 레인 분담 재료)."""
now_kst = datetime(2026, 6, 13, 13, 0, tzinfo=KST)
stats = {"summarize": _stage(pending=1317, done_1h=264)}
split = _split(macbook={"done_1h": 226, "done_today": 312}, macmini={"done_1h": 37, "done_today": 94})
ov = compose_overview(stats, split, {}, {}, [], deep_enabled=True, now_kst=now_kst)
assert ov["summarize_by_machine"]["macbook"]["done_1h"] == 226
assert ov["summarize_by_machine"]["macmini"]["done_today"] == 94
def test_deep_disabled_deep_summary_counts_to_macmini():
stats = {"deep_summary": _stage(pending=2, processing=1, done_1h=3, done_today=4)}
machines = build_machines(stats, _split(), [], deep_enabled=False)
macmini = _machine(machines, "macmini")
macbook = _machine(machines, "macbook")
assert macmini["pending"] == 2 and macmini["processing"] == 1
assert macmini["done_1h"] == 3 and macmini["done_today"] == 4
assert macbook["stages"] == [] and macbook["pending"] == 0
assert _machine(machines, "macmini")["stages"] == ["classify", "summarize", "deep_summary"]
def test_deferred_pending_always_on_macbook_card():
"""보류(deferred_until 미래)는 summarize+deep_summary 합산으로 macbook 카드 귀속.
deep 슬롯 유무와 무관 (보류 = 맥북 불가 신호)."""
def test_deferred_pending_on_macmini_card():
"""보류(deferred_until 미래)는 summarize+deep_summary 합산으로 macmini 카드 귀속
(보류 = LLM 백오프 신호)."""
stats = {
"summarize": _stage(pending=5, deferred_pending=2),
"deep_summary": _stage(pending=1, deferred_pending=1),
}
for deep_enabled in (True, False):
machines = build_machines(stats, _split(), [], deep_enabled=deep_enabled)
assert _machine(machines, "macbook")["deferred_pending"] == 3
assert _machine(machines, "gpu")["deferred_pending"] == 0
assert _machine(machines, "macmini")["deferred_pending"] == 0
machines = build_machines(stats, [])
assert _machine(machines, "macmini")["deferred_pending"] == 3
assert _machine(machines, "nas")["deferred_pending"] == 0
# ─── state 판정 ───────────────────────────────────────────────────────────────
def test_macbook_state_active_wins_over_deferred_while_working():
def test_macmini_state_active_wins_over_deferred_while_working():
"""가동 > 보류 (사용자 피드백 2026-06-11): 일하고 있으면 백오프 잔여가 있어도 '가동'.
보류 건수는 deferred_pending 필드가 별도로 전달 — 카드 라인이 표시.
"""
stats = {"summarize": _stage(pending=1, deferred_pending=1)}
split = _split(macbook={"done_15m": 3})
machines = build_machines(stats, split, [], deep_enabled=True)
mb = _machine(machines, "macbook")
assert mb["state"] == "active"
assert mb["deferred_pending"] == 1
stats = {"summarize": _stage(pending=1, deferred_pending=1, done_15m=3)}
machines = build_machines(stats, [])
mm = _machine(machines, "macmini")
assert mm["state"] == "active"
assert mm["deferred_pending"] == 1
def test_macbook_state_deferred_only_when_not_working():
def test_macmini_state_deferred_only_when_not_working():
"""일이 멈춰 있고(처리 0·최근 완료 0) 백오프만 쌓인 상태에서만 '보류'."""
stats = {"summarize": _stage(pending=1, deferred_pending=1)}
machines = build_machines(stats, _split(), [], deep_enabled=True)
assert _machine(machines, "macbook")["state"] == "deferred"
machines = build_machines(stats, [])
assert _machine(machines, "macmini")["state"] == "deferred"
def test_macbook_state_active_on_recent_qwen_done():
split = _split(macbook={"done_15m": 1})
machines = build_machines({}, split, [], deep_enabled=True)
assert _machine(machines, "macbook")["state"] == "active"
def test_macbook_state_idle():
machines = build_machines({}, _split(), [], deep_enabled=True)
assert _machine(machines, "macbook")["state"] == "idle"
def test_gpu_state_active_on_processing():
stats = {"extract": _stage(processing=1)}
machines = build_machines(stats, _split(), [], deep_enabled=True)
assert _machine(machines, "gpu")["state"] == "active"
def test_gpu_state_active_on_recent_done():
stats = {"embed": _stage(done_15m=2)}
machines = build_machines(stats, _split(), [], deep_enabled=True)
assert _machine(machines, "gpu")["state"] == "active"
def test_gpu_state_idle_when_old_done_only():
stats = {"embed": _stage(done_1h=5, done_today=9)} # 15분 내 완료 없음
machines = build_machines(stats, _split(), [], deep_enabled=True)
assert _machine(machines, "gpu")["state"] == "idle"
def test_macmini_state_not_active_on_macbook_pool_done():
"""summarize 풀 완료가 전부 macbook 몫이면 macmini 는 active 아님 (귀속 기준)."""
stats = {"summarize": _stage(done_15m=1)}
split = _split(macbook={"done_15m": 1})
machines = build_machines(stats, split, [], deep_enabled=True)
def test_macmini_state_idle():
machines = build_machines({}, [])
assert _machine(machines, "macmini")["state"] == "idle"
def test_nas_state_active_on_processing():
stats = {"extract": _stage(processing=1)}
machines = build_machines(stats, [])
assert _machine(machines, "nas")["state"] == "active"
def test_nas_state_active_on_recent_done():
stats = {"embed": _stage(done_15m=2)}
machines = build_machines(stats, [])
assert _machine(machines, "nas")["state"] == "active"
def test_nas_state_idle_when_old_done_only():
stats = {"embed": _stage(done_1h=5, done_today=9)} # 15분 내 완료 없음
machines = build_machines(stats, [])
assert _machine(machines, "nas")["state"] == "idle"
def test_macmini_state_active_on_summarize_processing():
stats = {"summarize": _stage(processing=1)}
machines = build_machines(stats, _split(), [], deep_enabled=True)
machines = build_machines(stats, [])
assert _machine(machines, "macmini")["state"] == "active"
@@ -228,21 +158,18 @@ def test_current_summarize_to_macmini_max_two():
{"stage": "summarize", "document_id": 3, "title": "문서C", "original_filename": None, "file_path": None},
{"stage": "extract", "document_id": 4, "title": "문서D", "original_filename": None, "file_path": None},
]
machines = build_machines({}, _split(), rows, deep_enabled=True)
machines = build_machines({}, rows)
macmini = _machine(machines, "macmini")
gpu = _machine(machines, "gpu")
nas = _machine(machines, "nas")
assert [c["document_id"] for c in macmini["current"]] == [1, 2] # 최대 2건
assert macmini["current"][0] == {"document_id": 1, "title": "문서A", "stage": "summarize"}
assert [c["document_id"] for c in gpu["current"]] == [4]
assert _machine(machines, "macbook")["current"] == []
assert [c["document_id"] for c in nas["current"]] == [4]
def test_current_deep_summary_follows_deep_slot():
def test_current_deep_summary_to_macmini():
rows = [{"stage": "deep_summary", "document_id": 9, "title": "심층", "original_filename": None, "file_path": None}]
enabled = build_machines({}, _split(), rows, deep_enabled=True)
disabled = build_machines({}, _split(), rows, deep_enabled=False)
assert _machine(enabled, "macbook")["current"][0]["document_id"] == 9
assert _machine(disabled, "macmini")["current"][0]["document_id"] == 9
machines = build_machines({}, rows)
assert _machine(machines, "macmini")["current"][0]["document_id"] == 9
def test_display_title_fallback_chain():
@@ -344,32 +271,15 @@ def test_rows_to_stage_stats_conversion():
assert stats["summarize"]["deferred_pending"] == 2
def test_rows_to_summarize_split_conversion():
rows = [
(True, 4, 8, 1), # is_macbook
(False, 6, 12, 0),
]
split = rows_to_summarize_split(rows)
assert split["macbook"] == {"done_1h": 4, "done_today": 8, "done_15m": 1}
assert split["macmini"] == {"done_1h": 6, "done_today": 12, "done_15m": 0}
def test_rows_to_summarize_split_empty():
split = rows_to_summarize_split([])
assert split["macbook"]["done_1h"] == 0 and split["macmini"]["done_today"] == 0
def test_compose_overview_contract_shape():
"""응답 dict 의 키가 FE 계약 shape 과 정확히 일치하는지 고정."""
out = compose_overview(
{"summarize": _stage(pending=1)},
_split(),
{}, {}, [],
deep_enabled=True,
now_kst=datetime(2026, 6, 11, 14, 30, tzinfo=KST),
)
assert set(out.keys()) == {"machines", "stages", "summarize_eta", "trend_24h", "totals"}
assert [m["key"] for m in out["machines"]] == ["gpu", "macmini", "macbook"]
assert [m["key"] for m in out["machines"]] == ["nas", "macmini"]
for m in out["machines"]:
assert set(m.keys()) == {
"key", "label", "state", "stages", "pending", "processing", "failed",
@@ -381,7 +291,7 @@ def test_compose_overview_contract_shape():
assert set(out["trend_24h"][0].keys()) == {"hour", "inflow", "done"}
assert set(out["totals"].keys()) == {"pending", "processing", "failed"}
# 머신 label 고정 (raw 모델명 노출 금지 — label 만)
assert [m["label"] for m in out["machines"]] == ["GPU 서버", "맥미니", "맥북 M5 Max"]
assert [m["label"] for m in out["machines"]] == ["나스", "맥미니"]
# ─── build_stages (단계별 현황 — 2026-06-11 사용자 피드백: 완료 가시화) ──────