fix: correct time_detected logic, immutable sort pattern, add diagnose() test

This commit is contained in:
pyr0ball 2026-05-11 09:08:24 -07:00
parent abd142addf
commit ca0cb1361e
2 changed files with 29 additions and 11 deletions

View file

@ -1,12 +1,16 @@
"""Frictionless diagnose service — NL time extraction + layered log search."""
from __future__ import annotations
import logging
import re
from datetime import datetime, timedelta, timezone
from pathlib import Path
from typing import Any
from app.services.search import SearchResult, entries_in_window, search
logger = logging.getLogger(__name__)
try:
from dateparser.search import search_dates as _search_dates # type: ignore[import]
_HAS_DATEPARSER = True
@ -22,7 +26,11 @@ def parse_time_window(query: str) -> tuple[str | None, str | None, str]:
the matched time phrase stripped. Falls back to last-60-min window.
"""
if _HAS_DATEPARSER and _search_dates is not None:
results = _search_dates(query, languages=["en"], settings={"PREFER_DATES_FROM": "past"})
try:
results = _search_dates(query, languages=["en"], settings={"PREFER_DATES_FROM": "past"})
except Exception:
logger.warning("dateparser failed on query %r — falling back to 60-min window", query)
results = None
if results:
phrase, dt = results[0]
if dt.tzinfo is None:
@ -40,10 +48,10 @@ def diagnose(
query: str,
since: str | None = None,
until: str | None = None,
) -> dict:
) -> dict[str, Any]:
"""Run layered log search with NL time extraction. Returns summary + entries."""
time_detected = since is not None or until is not None
if since is None or until is None:
time_detected = since is not None and until is not None
if not time_detected:
parsed_since, parsed_until, keywords = parse_time_window(query)
since = since or parsed_since
until = until or parsed_until
@ -55,14 +63,13 @@ def diagnose(
window_hits = entries_in_window(db_path, since=since, until=until, limit=50)
seen: set[str] = set()
combined: list[SearchResult] = []
merged: list[SearchResult] = []
for r in keyword_hits + window_hits:
if r.entry_id not in seen:
seen.add(r.entry_id)
combined.append(r)
merged.append(r)
combined.sort(key=lambda r: (r.timestamp_iso or "\xff", r.sequence))
combined = combined[:200]
combined = sorted(merged, key=lambda r: (r.timestamp_iso or "\xff", r.sequence))[:200]
by_severity: dict[str, int] = {"CRITICAL": 0, "ERROR": 0, "WARN": 0, "INFO": 0}
by_source: dict[str, int] = {}

View file

@ -5,7 +5,7 @@ from datetime import datetime, timezone
from unittest.mock import patch
from app.services.diagnose import parse_time_window
from app.services.diagnose import diagnose, parse_time_window
def test_no_time_phrase_falls_back_to_last_60_min():
@ -18,7 +18,7 @@ def test_no_time_phrase_falls_back_to_last_60_min():
assert abs(diff - 3600) < 5
def test_time_phrase_detected_produces_30min_window():
def test_time_phrase_detected_produces_60min_window():
dt = datetime(2026, 5, 11, 14, 0, tzinfo=timezone.utc)
with patch("app.services.diagnose._HAS_DATEPARSER", True), \
patch("app.services.diagnose._search_dates", return_value=[("around 2pm", dt)]):
@ -37,7 +37,8 @@ def test_time_phrase_stripped_from_keywords():
def test_no_dateparser_falls_back_to_60min():
with patch("app.services.diagnose._HAS_DATEPARSER", False):
with patch("app.services.diagnose._HAS_DATEPARSER", False), \
patch("app.services.diagnose._search_dates", None):
since, until, keywords = parse_time_window("plex stopped playing audio")
assert keywords == "plex stopped playing audio"
diff = (datetime.fromisoformat(until) - datetime.fromisoformat(since)).total_seconds()
@ -50,3 +51,13 @@ def test_keywords_cleaned_of_extra_spaces():
patch("app.services.diagnose._search_dates", return_value=[("around 2pm", dt)]):
_, _, keywords = parse_time_window("plex stopped audio around 2pm extra")
assert " " not in keywords
def test_diagnose_with_explicit_window_sets_time_detected(tmp_path):
from app.ingest.pipeline import ensure_schema
db = tmp_path / "test.db"
ensure_schema(db)
result = diagnose(db, query="plex", since="2026-05-11T14:00:00+00:00", until="2026-05-11T15:00:00+00:00")
assert result["summary"]["time_detected"] is True
assert result["summary"]["total"] == 0 # empty DB
assert result["summary"]["window_start"] == "2026-05-11T14:00:00+00:00"