hyungi
7031439364
'대기만 보이고 성공은 안 보인다' 피드백 반영: - overview 에 stages[] 노출 (stage 별 done_today + oldest_pending_age, SQL 1필드 추가) - 게이지 의미 전환: 단계 간 대기량 비교(amber) → 단계 내 오늘 진척(완료=green 비율, 가득 찬 초록 = 다 끝남) + 처리 중 pulse dot - 움직임 없는 단계는 행 제거, 하단 '비어 있음: ...' 한 줄로 - 라벨 누수 fix: details 가 구 STAGE_LABEL 을 쓰던 것 → queueStageLabel 통일 (deep_summary/markdown/summarize/chunk/fulltext 한글화) - 헤더: 오늘 N 완료(성공 가시화) · 실패(error) · 대기. 데이터 소스 = overview 단일화 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
371 lines
16 KiB
Python
371 lines
16 KiB
Python
"""GET /api/queue/overview 판정부 단위테스트 — DB 불요 (plan ds-processing-ui-6an).
|
|
|
|
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.
|
|
"""
|
|
|
|
from datetime import datetime
|
|
from zoneinfo import ZoneInfo
|
|
|
|
from services.queue_overview import (
|
|
build_machines,
|
|
build_summarize_eta,
|
|
build_totals,
|
|
build_trend,
|
|
compose_overview,
|
|
compute_eta_minutes,
|
|
display_title,
|
|
rows_to_stage_stats,
|
|
rows_to_summarize_split,
|
|
stage_machine_map,
|
|
)
|
|
|
|
KST = ZoneInfo("Asia/Seoul")
|
|
|
|
|
|
def _stage(**kw) -> dict:
|
|
"""stage 통계 1건 — 미지정 필드 0."""
|
|
base = {
|
|
"pending": 0, "processing": 0, "failed": 0,
|
|
"done_1h": 0, "done_today": 0, "done_15m": 0,
|
|
"deferred_pending": 0, "created_1h": 0,
|
|
}
|
|
base.update(kw)
|
|
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)
|
|
for s in ("extract", "embed", "chunk", "markdown", "preview", "thumbnail", "fulltext", "stt"):
|
|
assert smap[s] == "gpu"
|
|
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():
|
|
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"] == [
|
|
"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."""
|
|
stats = {
|
|
"classify": _stage(done_1h=2, done_today=3),
|
|
"summarize": _stage(pending=7, failed=1, done_1h=10, done_today=20),
|
|
}
|
|
split = _split(macbook={"done_1h": 4, "done_today": 8}, macmini={"done_1h": 6, "done_today": 12})
|
|
machines = build_machines(stats, split, [], deep_enabled=True)
|
|
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
|
|
|
|
|
|
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)."""
|
|
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
|
|
|
|
|
|
# ─── state 판정 ───────────────────────────────────────────────────────────────
|
|
|
|
def test_macbook_state_deferred_wins_over_active():
|
|
stats = {"summarize": _stage(pending=1, deferred_pending=1)}
|
|
split = _split(macbook={"done_15m": 3}) # 최근 완료가 있어도 deferred 우선
|
|
machines = build_machines(stats, split, [], deep_enabled=True)
|
|
assert _machine(machines, "macbook")["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)
|
|
assert _machine(machines, "macmini")["state"] == "idle"
|
|
|
|
|
|
def test_macmini_state_active_on_summarize_processing():
|
|
stats = {"summarize": _stage(processing=1)}
|
|
machines = build_machines(stats, _split(), [], deep_enabled=True)
|
|
assert _machine(machines, "macmini")["state"] == "active"
|
|
|
|
|
|
# ─── current 귀속 ─────────────────────────────────────────────────────────────
|
|
|
|
def test_current_summarize_to_macmini_max_two():
|
|
rows = [
|
|
{"stage": "summarize", "document_id": 1, "title": "문서A", "original_filename": None, "file_path": None},
|
|
{"stage": "summarize", "document_id": 2, "title": "문서B", "original_filename": None, "file_path": None},
|
|
{"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)
|
|
macmini = _machine(machines, "macmini")
|
|
gpu = _machine(machines, "gpu")
|
|
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"] == []
|
|
|
|
|
|
def test_current_deep_summary_follows_deep_slot():
|
|
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
|
|
|
|
|
|
def test_display_title_fallback_chain():
|
|
assert display_title({"document_id": 1, "title": "제목"}) == "제목"
|
|
assert display_title({"document_id": 1, "title": None, "original_filename": "a.pdf"}) == "a.pdf"
|
|
assert display_title(
|
|
{"document_id": 1, "title": None, "original_filename": None, "file_path": "/documents/PKM/Inbox/b.hwp"}
|
|
) == "b.hwp"
|
|
assert display_title(
|
|
{"document_id": 7, "title": None, "original_filename": None, "file_path": None}
|
|
) == "문서 #7"
|
|
|
|
|
|
# ─── summarize ETA ────────────────────────────────────────────────────────────
|
|
|
|
def test_eta_minutes_positive_drain():
|
|
# 순소화 6건/h, 잔량 30건 → 300분
|
|
assert compute_eta_minutes(30, 10, 4) == 300
|
|
|
|
|
|
def test_eta_minutes_null_when_not_draining():
|
|
assert compute_eta_minutes(30, 4, 10) is None # 유입 > 소화
|
|
assert compute_eta_minutes(30, 5, 5) is None # 동률도 null
|
|
assert compute_eta_minutes(30, 0, 0) is None
|
|
|
|
|
|
def test_eta_minutes_zero_pending():
|
|
assert compute_eta_minutes(0, 10, 4) == 0
|
|
|
|
|
|
def test_build_summarize_eta_pending_includes_deferred():
|
|
stats = {"summarize": _stage(pending=12, deferred_pending=5, done_1h=8, created_1h=2)}
|
|
eta = build_summarize_eta(stats)
|
|
assert eta == {
|
|
"pending": 12, # 보류 포함 총수 (pending 자체에 deferred 포함)
|
|
"done_rate_1h": 8,
|
|
"inflow_rate_1h": 2,
|
|
"eta_minutes": round(12 / 6 * 60),
|
|
}
|
|
|
|
|
|
def test_build_summarize_eta_empty_stats():
|
|
eta = build_summarize_eta({})
|
|
assert eta == {"pending": 0, "done_rate_1h": 0, "inflow_rate_1h": 0, "eta_minutes": None}
|
|
|
|
|
|
# ─── trend 24h ────────────────────────────────────────────────────────────────
|
|
|
|
def test_trend_24_buckets_oldest_first_with_gaps():
|
|
now_kst = datetime(2026, 6, 11, 14, 30, tzinfo=KST)
|
|
inflow = {"2026-06-11 13:00": 3, "2026-06-10 15:00": 1} # 15:00 어제 = 최고령 버킷
|
|
done = {"2026-06-11 14:00": 2}
|
|
trend = build_trend(inflow, done, now_kst)
|
|
|
|
assert len(trend) == 24
|
|
assert trend[0] == {"hour": "15:00", "inflow": 1, "done": 0} # 오래된 것부터
|
|
assert trend[-1] == {"hour": "14:00", "inflow": 0, "done": 2} # 현재 시각 버킷
|
|
assert trend[-2] == {"hour": "13:00", "inflow": 3, "done": 0}
|
|
# 빈 버킷은 0
|
|
assert sum(b["inflow"] for b in trend) == 4
|
|
assert sum(b["done"] for b in trend) == 2
|
|
|
|
|
|
def test_trend_ignores_out_of_window_bucket():
|
|
"""창 밖(24버킷 미포함) key 는 무시 — cutoff 경계 행이 섞여도 안전."""
|
|
now_kst = datetime(2026, 6, 11, 14, 30, tzinfo=KST)
|
|
inflow = {"2026-06-10 14:00": 99} # 14:00 어제 — 창의 최고령(15:00 어제) 이전
|
|
trend = build_trend(inflow, {}, now_kst)
|
|
assert sum(b["inflow"] for b in trend) == 0
|
|
|
|
|
|
def test_trend_kst_midnight_crossing_labels():
|
|
now_kst = datetime(2026, 6, 11, 2, 5, tzinfo=KST)
|
|
trend = build_trend({}, {}, now_kst)
|
|
assert trend[-1]["hour"] == "02:00"
|
|
assert trend[0]["hour"] == "03:00" # 전날 03:00 (라벨은 HH:00 만)
|
|
assert [b["hour"] for b in trend[-3:]] == ["00:00", "01:00", "02:00"]
|
|
|
|
|
|
# ─── totals / row 변환 / 전체 조립 ───────────────────────────────────────────
|
|
|
|
def test_totals_sum_all_stages():
|
|
stats = {
|
|
"extract": _stage(pending=1, processing=2, failed=3),
|
|
"summarize": _stage(pending=4, failed=1),
|
|
"deep_summary": _stage(pending=2),
|
|
}
|
|
assert build_totals(stats) == {"pending": 7, "processing": 2, "failed": 4}
|
|
|
|
|
|
def test_rows_to_stage_stats_conversion():
|
|
rows = [
|
|
("extract", 3, 1, 0, 5, 9, 1, 0, 2),
|
|
("summarize", 7, None, 1, 10, 20, 0, 2, 4), # None 방어
|
|
]
|
|
stats = rows_to_stage_stats(rows)
|
|
assert stats["extract"]["pending"] == 3 and stats["extract"]["created_1h"] == 2
|
|
assert stats["summarize"]["processing"] == 0
|
|
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"]
|
|
for m in out["machines"]:
|
|
assert set(m.keys()) == {
|
|
"key", "label", "state", "stages", "pending", "processing", "failed",
|
|
"done_1h", "done_today", "deferred_pending", "current",
|
|
}
|
|
assert m["state"] in ("active", "deferred", "idle")
|
|
assert set(out["summarize_eta"].keys()) == {"pending", "done_rate_1h", "inflow_rate_1h", "eta_minutes"}
|
|
assert len(out["trend_24h"]) == 24
|
|
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"]
|
|
|
|
|
|
# ─── build_stages (단계별 현황 — 2026-06-11 사용자 피드백: 완료 가시화) ──────
|
|
|
|
def test_build_stages_order_fields_and_age():
|
|
from datetime import timedelta, timezone
|
|
from services.queue_overview import build_stages
|
|
now = datetime(2026, 6, 11, 14, 0, tzinfo=timezone.utc)
|
|
stats = {
|
|
"summarize": {**_stage(pending=5, done_today=12),
|
|
"oldest_pending_at": now - timedelta(hours=4)},
|
|
"extract": _stage(failed=2),
|
|
}
|
|
rows = build_stages(stats, now=now)
|
|
by = {r["stage"]: r for r in rows}
|
|
# 파이프라인 순서: extract 가 summarize 보다 앞
|
|
assert rows[0]["stage"] == "extract"
|
|
assert by["summarize"]["pending"] == 5
|
|
assert by["summarize"]["done_today"] == 12
|
|
assert by["summarize"]["oldest_pending_age_sec"] == 4 * 3600
|
|
assert by["extract"]["failed"] == 2
|
|
assert by["extract"]["oldest_pending_age_sec"] is None
|
|
# 전 stage 행 존재 (빈 단계 숨김은 FE 몫)
|
|
assert {"stage", "pending", "processing", "failed", "done_today",
|
|
"oldest_pending_age_sec"} == set(rows[0].keys())
|