soulsync/tests/test_video_scanner.py
BoulderBadgeDad 8b01ada68a video: pause enrichment workers during library scans, resume after
Every scan (incremental / full / deep, both entry points) now steps the
enrichment workers aside to cut DB lock contention — same as music. Mirrors
music's contract exactly: pause ONLY workers that were running (a user's manual
pause is left alone), track which we paused, and resume just those in a finally
so success, cancel, and error all clean up. Auto-pause is transient
(persist=False) so it never leaks into the saved <service>_paused flag.

Hooks are injected by get_video_scanner; the scanner is inert without them
(tests build it directly, no engine spun up). Isolated to the video side.
2026-06-14 13:55:59 -07:00

249 lines
9.7 KiB
Python

"""Seam tests for the video library scanner (experimental branch).
The scanner is server-agnostic: it consumes a media source that yields
normalized dicts. We drive it with a fake source so the scan/prune/error logic
is fully tested without a live Plex/Jellyfin. Also guards that core/video/
imports nothing from the music database.
"""
from __future__ import annotations
from pathlib import Path
import pytest
from database.video_database import VideoDatabase
from core.video.scanner import VideoLibraryScanner
class FakeSource:
server_name = "plex"
def __init__(self, movies, shows):
self._movies, self._shows = movies, shows
self.incremental_calls = []
def iter_movies(self, incremental=False):
self.incremental_calls.append(("movies", incremental))
return iter(self._movies)
def iter_shows(self, incremental=False):
self.incremental_calls.append(("shows", incremental))
return iter(self._shows)
@pytest.fixture()
def db(tmp_path):
return VideoDatabase(database_path=str(tmp_path / "video_library.db"))
def test_scan_sync_populates_library(db):
movies = [{"server_id": "m1", "title": "A", "file": {"relative_path": "a.mkv", "size_bytes": 5}}]
shows = [{"server_id": "s1", "title": "Show", "seasons": [
{"season_number": 1, "episodes": [
{"episode_number": 1, "title": "E1", "file": {"relative_path": "e1.mkv"}}]}]}]
st = VideoLibraryScanner(db).scan_sync(lambda: FakeSource(movies, shows))
assert st["state"] == "done"
assert (st["movies"], st["shows"], st["episodes"]) == (1, 1, 1)
lib = db.dashboard_stats()["library"]
assert (lib["movies"], lib["shows"], lib["episodes"]) == (1, 1, 1)
def test_scan_pauses_and_resumes_enrichment_workers(db):
events = []
scanner = VideoLibraryScanner(
db, pause_workers=lambda: events.append("pause"),
resume_workers=lambda: events.append("resume"))
scanner.scan_sync(lambda: FakeSource([{"server_id": "m1", "title": "A"}], []))
assert events == ["pause", "resume"] # paused first, resumed after
def test_scan_resumes_enrichment_even_on_error(db):
events = []
scanner = VideoLibraryScanner(
db, pause_workers=lambda: events.append("pause"),
resume_workers=lambda: events.append("resume"))
def boom():
raise RuntimeError("server blew up")
scanner.scan_sync(boom) # source_factory raises
assert scanner.get_status()["state"] == "error"
assert events == ["pause", "resume"] # resumed despite the failure
def test_scan_resumes_enrichment_on_cancel(db):
events = []
scanner = VideoLibraryScanner(
db, pause_workers=lambda: events.append("pause"),
resume_workers=lambda: events.append("resume"))
class S:
server_name = "plex"
def counts(self, incremental=False):
return {"movies": 5, "shows": 0}
def iter_movies(self, incremental=False):
for i in range(5):
scanner._cancel = True # cancel mid-iteration
yield {"server_id": "m%d" % i, "title": "M%d" % i}
def iter_shows(self, incremental=False):
return iter([])
scanner.scan_sync(lambda: S())
assert scanner.get_status()["state"] == "cancelled"
assert events == ["pause", "resume"] # resumed after cancel too
def test_scanner_with_no_hooks_does_not_pause(db):
# Default construction (as in tests / when no engine wired) must be inert.
scanner = VideoLibraryScanner(db)
st = scanner.scan_sync(lambda: FakeSource([{"server_id": "m1", "title": "A"}], []))
assert st["state"] == "done" # scans fine without hooks
def test_deep_scan_prunes_removed_items(db):
scanner = VideoLibraryScanner(db)
scanner.scan_sync(lambda: FakeSource(
[{"server_id": "m1", "title": "A"}, {"server_id": "m2", "title": "B"}], []), mode="deep")
assert db.dashboard_stats()["library"]["movies"] == 2
scanner.scan_sync(lambda: FakeSource([{"server_id": "m1", "title": "A"}], []), mode="deep")
assert db.dashboard_stats()["library"]["movies"] == 1
def test_full_refresh_does_not_prune(db):
# 'full' refreshes/adds but never removes — only 'deep' prunes.
scanner = VideoLibraryScanner(db)
scanner.scan_sync(lambda: FakeSource(
[{"server_id": "m1", "title": "A"}, {"server_id": "m2", "title": "B"}], []), mode="deep")
scanner.scan_sync(lambda: FakeSource([{"server_id": "m1", "title": "A"}], []), mode="full")
assert db.dashboard_stats()["library"]["movies"] == 2 # m2 NOT pruned
def test_empty_deep_scan_does_not_wipe_library(db):
# Safety: a deep scan that returns nothing (transient failure) must NOT prune.
scanner = VideoLibraryScanner(db)
scanner.scan_sync(lambda: FakeSource([{"server_id": "m1", "title": "A"}], []), mode="deep")
scanner.scan_sync(lambda: FakeSource([], []), mode="deep")
assert db.dashboard_stats()["library"]["movies"] == 1
def test_incremental_mode_requests_incremental_from_source(db):
# Populate past the small-library fallback so incremental stays incremental.
for i in range(60):
db.upsert_movie("plex", {"server_id": "p%d" % i, "title": str(i)})
src = FakeSource([{"server_id": "m1", "title": "A"}], [])
VideoLibraryScanner(db).scan_sync(lambda: src, mode="incremental")
assert ("movies", True) in src.incremental_calls
assert ("shows", True) in src.incremental_calls
def test_scan_sync_no_source_reports_error(db):
st = VideoLibraryScanner(db).scan_sync(lambda: None)
assert st["state"] == "error" and "error" in st
def test_scan_reports_percent_from_counts(db):
class S:
server_name = "plex"
def counts(self, incremental=False):
return {"movies": 4, "shows": 0}
def iter_movies(self, incremental=False):
return iter([{"server_id": "m%d" % i, "title": str(i)} for i in range(4)])
def iter_shows(self, incremental=False):
return iter([])
st = VideoLibraryScanner(db).scan_sync(lambda: S())
assert st["state"] == "done"
assert st["percent"] == 100
def test_scan_cancel_stops_midway(db):
scanner = VideoLibraryScanner(db)
def gen():
yield {"server_id": "m1", "title": "A"}
scanner.cancel() # request stop after the first item
yield {"server_id": "m2", "title": "B"}
class S:
server_name = "plex"
def counts(self, incremental=False):
return {"movies": 2, "shows": 0}
def iter_movies(self, incremental=False):
return gen()
def iter_shows(self, incremental=False):
return iter([])
st = scanner.scan_sync(lambda: S())
assert st["state"] == "cancelled"
assert db.dashboard_stats()["library"]["movies"] == 1 # only the first was saved
def test_incremental_skips_known_and_early_stops(db):
# Past the small-library fallback, with everything already known.
for i in range(60):
db.upsert_movie("plex", {"server_id": "k%d" % i, "title": "K%d" % i})
new_item = {"server_id": "new1", "title": "New"}
known = [{"server_id": "k%d" % i, "title": "K%d" % i} for i in range(60)]
class S:
server_name = "plex"
def counts(self, incremental=False):
return {"movies": 61, "shows": 0}
def iter_movies(self, incremental=False):
assert incremental is True # NOT fallen back (library big enough)
return iter([new_item] + known) # one new, then a long run of known
def iter_shows(self, incremental=False):
return iter([])
st = VideoLibraryScanner(db).scan_sync(lambda: S(), mode="incremental")
assert st["state"] == "done"
assert st["movies"] == 1 # only the new one; known skipped
assert db.table_count("movies") == 61
def test_incremental_falls_back_to_full_on_small_library(db):
captured = {}
class S:
server_name = "plex"
def counts(self, incremental=False):
return {"movies": 3, "shows": 0}
def iter_movies(self, incremental=False):
captured["incremental"] = incremental
return iter([{"server_id": "m%d" % i, "title": str(i)} for i in range(3)])
def iter_shows(self, incremental=False):
return iter([])
st = VideoLibraryScanner(db).scan_sync(lambda: S(), mode="incremental")
assert captured["incremental"] is False # empty DB (<50) -> full pass
assert st["movies"] == 3
def test_parse_plex_guids():
from core.video.sources import _parse_plex_guids
class _G:
def __init__(self, gid): self.id = gid
class _Obj:
def __init__(self, guids): self.guids = guids
got = _parse_plex_guids(_Obj([_G("imdb://tt1375666"), _G("tmdb://27205"), _G("tvdb://121361")]))
assert got == {"tmdb_id": 27205, "imdb_id": "tt1375666", "tvdb_id": 121361}
assert _parse_plex_guids(_Obj([])) == {"tmdb_id": None, "imdb_id": None, "tvdb_id": None}
def test_parse_jellyfin_providers():
from core.video.sources import _parse_jf_providers
got = _parse_jf_providers({"ProviderIds": {"Imdb": "tt123", "Tmdb": "27205", "Tvdb": "121361"}})
assert got == {"imdb_id": "tt123", "tmdb_id": 27205, "tvdb_id": 121361}
assert _parse_jf_providers({}) == {"imdb_id": None, "tmdb_id": None, "tvdb_id": None}
def test_core_video_imports_nothing_from_music():
base = Path(__file__).resolve().parent.parent / "core" / "video"
for py in base.glob("*.py"):
for line in py.read_text(encoding="utf-8").splitlines():
s = line.strip()
if s.startswith("import ") or s.startswith("from "):
assert "music" not in s.lower(), f"{py.name}: music import leaked: {s!r}"