mirror of
https://github.com/unshackle-dl/unshackle.git
synced 2026-11-12 12:05:07 +00:00
imSp4rky
ffeb58f808
Fix DASH SegmentTimeline negative @r and Pattern, TTML sidecar sets with no contentType, HLS CHANNELS identifiers and Dolby channel masks; test against 408 public DASH/HLS/ISM manifests.
107 lines
3.9 KiB
Python
107 lines
3.9 KiB
Python
"""Parse every public test manifest in tests/fixtures/manifests (refresh: scripts/fetch_manifest_corpus.py)."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import json
|
|
from collections import Counter
|
|
from pathlib import Path
|
|
from typing import Any, Union
|
|
from unittest.mock import MagicMock
|
|
|
|
import pytest
|
|
|
|
from unshackle.core.manifests import DASH, HLS, ISM
|
|
from unshackle.core.tracks import Audio, Tracks, Video
|
|
from unshackle.core.utils.xml import load_xml
|
|
|
|
CORPUS = Path(__file__).resolve().parent.parent / "fixtures" / "manifests"
|
|
INDEX: list[dict[str, Any]] = json.loads((CORPUS / "index.json").read_text(encoding="utf-8"))
|
|
|
|
# file -> reason; strict, so a fix that makes one parse fails here until its entry is removed
|
|
KNOWN_FAILURES: dict[str, str] = {}
|
|
|
|
|
|
def cases(entries: list[dict[str, Any]]) -> list[Any]:
|
|
return [
|
|
pytest.param(e, id=e["file"], marks=pytest.mark.xfail(reason=KNOWN_FAILURES[e["file"]], strict=True))
|
|
if e["file"] in KNOWN_FAILURES
|
|
else pytest.param(e, id=e["file"])
|
|
for e in entries
|
|
]
|
|
|
|
|
|
def parse(entry: dict[str, Any]) -> Tracks:
|
|
"""Parse a fixture the way from_url does: XML from bytes, M3U8 decoded as UTF-8."""
|
|
data = (CORPUS / entry["file"]).read_bytes()
|
|
parser: Union[DASH, HLS, ISM]
|
|
if entry["kind"] == "hls":
|
|
parser = HLS.from_text(data.decode("utf-8"), entry["url"])
|
|
else:
|
|
parser = (DASH if entry["kind"] == "dash" else ISM)(load_xml(data), entry["url"])
|
|
return parser.to_tracks("en")
|
|
|
|
|
|
def most_per_set(tracks: list[Any]) -> int:
|
|
return max(Counter(id(t.data["dash"]["adaptation_set"]) for t in tracks).values(), default=0)
|
|
|
|
|
|
class NeedsNetwork(Exception):
|
|
pass
|
|
|
|
|
|
def test_index_lists_every_known_failure() -> None:
|
|
assert INDEX
|
|
assert set(KNOWN_FAILURES) <= {e["file"] for e in INDEX}
|
|
|
|
|
|
@pytest.mark.parametrize("entry", cases(INDEX))
|
|
def test_manifest_parses_to_tracks(entry: dict[str, Any]) -> None:
|
|
tracks = parse(entry)
|
|
|
|
assert tracks.videos or tracks.audio or tracks.subtitles
|
|
for track in tracks:
|
|
# HLS CODECS is only a SHOULD (RFC 8216 4.3.4.2), so an HLS track can honestly have none
|
|
assert track.codec is not None or entry["kind"] == "hls", track
|
|
assert track.language, track
|
|
|
|
expect = entry.get("expect")
|
|
if expect:
|
|
assert most_per_set(tracks.videos) == expect["max_num_video"]
|
|
assert most_per_set(tracks.audio) == expect["max_num_audio"]
|
|
for codec_type, codecs, found in (
|
|
(Video.Codec, expect["video_codec"], tracks.videos),
|
|
(Audio.Codec, expect["audio_codec"], tracks.audio),
|
|
):
|
|
if codecs:
|
|
assert {codec_type.from_codecs(c) for c in codecs.split(",")} <= {t.codec for t in found}
|
|
|
|
|
|
@pytest.mark.parametrize(
|
|
"entry",
|
|
[
|
|
e
|
|
for e in INDEX
|
|
if e["kind"] == "dash"
|
|
and e["file"] not in KNOWN_FAILURES
|
|
and b'type="dynamic"' not in (CORPUS / e["file"]).read_bytes()
|
|
],
|
|
ids=lambda e: e["file"],
|
|
)
|
|
def test_dash_segments_and_drm_resolve_offline(entry: dict[str, Any]) -> None:
|
|
"""Build the DRM and the segment list of every track; a track whose segment index must be fetched is skipped."""
|
|
session = MagicMock()
|
|
session.get.side_effect = NeedsNetwork
|
|
for track in parse(entry):
|
|
dash = track.data["dash"]
|
|
manifest, adaptation_set, representation = dash["manifest"], dash["adaptation_set"], dash["representation"]
|
|
DASH.get_drm(representation.findall("ContentProtection") + adaptation_set.findall("ContentProtection"))
|
|
period = dash["period"] if dash.get("period") is not None else manifest.find("Period")
|
|
try:
|
|
_, segments, _, _, _ = DASH.get_period_segments(
|
|
period, adaptation_set, representation, manifest, track, track.url, session, probe_kid=False
|
|
)
|
|
except NeedsNetwork:
|
|
continue
|
|
assert segments, track
|
|
assert all(url.startswith(("http://", "https://")) for url, _ in segments), track
|