lukas.pupkalipinski/Aniworld
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(providers): rotate, probe and fall back on 404

Browse Source 
Iterate providers actually advertised on the episode page (ordered by
SUPPORTED_PROVIDERS preference) instead of always re-resolving VOE.
Each candidate is HEAD-probed before yt-dlp runs, so dead links are
skipped immediately; direct video URLs use a streaming fast path that
bypasses yt-dlp; total failure now logs the exhausted provider list.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
This commit is contained in:
Lukas
2026-05-25 14:32:10 +02:00
parent c579235af0
commit a115215416
3 changed files with 629 additions and 24 deletions
Download Patch File Download Diff File Expand all files Collapse all files

61
docs/DEVELOPMENT.md
View File

@@ -308,3 +308,64 @@ If `/health` returns `unhealthy` status:
- Check network connectivity
- DNS failures are transient — warnings don't block startup
- Retry later to verify: `GET /health`
### Provider Failure Handling
Download providers (VOE, Doodstream, Vidmoly, Vidoza, SpeedFiles, Streamtape,
Luluvdo) regularly break: URLs expire, sites change their player markup, geo
blocks appear, and `yt-dlp` extractors lag behind upstream changes. The
`AniworldLoader.download()` flow is designed to fail fast and rotate.
**Rotation order**
1. The episode page is scraped for the providers AniWorld actually advertises.
2. Results are ordered by the preference in `DEFAULT_PROVIDERS`
(`provider_config.py`); providers not listed run last.
3. For each candidate the loader:
1. Calls `_check_url_alive()` — HEAD probe with GET fallback. Any 4xx
response or connection error skips the provider immediately.
2. Resolves the redirect via `_resolve_direct_link()` to obtain a direct
stream URL plus headers. Provider-specific extractors (e.g. `VOE`) are
preferred; unknown providers fall back to the embed URL so `yt-dlp` can
attempt extraction.
3. Tries `_try_direct_stream()` — straight `requests.get(stream=True)` when
`Content-Type` is `video/*` or `application/octet-stream`. This avoids
`yt-dlp` entirely for direct MP4 links.
4. Falls back to `yt-dlp` with the ffmpeg downloader for HLS streams.
4. On any failure, temp files are cleaned and the loop moves to the next
provider. When the chain is exhausted, the loader logs
`All download providers failed for S{season}E{episode} ...; tried=[...]`
to both the application log and `logs/download_errors.log`.
**Do not hardcode provider URLs.** Provider domains shift constantly (e.g.
Doodstream alternates between `dood.li`, `dood.so`, `dood.la`). Only the
referer hints in `PROVIDER_HEADERS` are persisted — discovery still happens
at runtime through AniWorld's redirect endpoint.
### Updating yt-dlp
When extractors break (typical symptoms: every provider HEAD probe succeeds
but `yt-dlp` raises `Unable to extract` or `HTTP Error 404`):
1. Check the upstream tracker first: https://github.com/yt-dlp/yt-dlp/issues
2. Upgrade in the conda environment:
```bash
conda run -n AniWorld pip install --upgrade yt-dlp
```
3. Smoke-test against a known-good episode before pinning a new floor in
`requirements.txt` (`yt-dlp>=YYYY.MM.DD`).
4. Re-run the provider test suite:
```bash
conda run -n AniWorld python -m pytest tests/unit/test_aniworld_provider.py -v
```
5. If a specific extractor is removed upstream, drop the provider from
`DEFAULT_PROVIDERS` rather than patching `yt-dlp` in tree.
### User Notification on Total Failure
`SeriesApp.download_episode()` already emits a `download_status="failed"`
WebSocket event when `loader.download()` returns `False`. Operators should
forward this to `notification_service.notify_download_failed()` so users see
a HIGH-priority alert. The loader keeps the failure detail in
`logs/download_errors.log` for post-mortem.

337
src/core/providers/aniworld_provider.py
View File

@@ -249,6 +249,118 @@ class AniworldLoader(Loader):
logger.debug("Available languages for S%02dE%03d: %s, requested: %s, available: %s", season, episode, languages, language_code, is_available)
return is_available
def _check_url_alive(
self,
url: str,
headers: dict | None = None,
timeout: int = 10,
) -> bool:
"""Probe a provider URL with HEAD before committing to yt-dlp.
Skips dead providers quickly so the failover loop never blocks
waiting for yt-dlp to fail on a 404. Falls back to a streaming
GET when HEAD is not allowed by the upstream server.
Args:
url: URL to probe.
headers: Optional headers to forward with the probe.
timeout: Per-request timeout (seconds).
Returns:
True when the URL responds with a non-4xx status, else False.
"""
try:
response = self.session.head(
url,
headers=headers,
timeout=timeout,
allow_redirects=True,
)
if response.status_code == 405:
response = self.session.get(
url,
headers=headers,
timeout=timeout,
stream=True,
allow_redirects=True,
)
response.close()
if 400 <= response.status_code < 500:
logger.warning(
"Provider URL returned HTTP %s: %s",
response.status_code, url
)
return False
return True
except requests.RequestException as exc:
logger.warning("Provider URL unreachable %s: %s", url, exc)
return False
def _try_direct_stream(
self,
link: str,
output_path: str,
headers: dict | None,
timeout: int,
) -> bool:
"""Stream a direct video URL to disk without yt-dlp.
Used as a fast-path when the resolved provider link already points
at a downloadable video file (``Content-Type: video/*`` or
``application/octet-stream``). HLS and other non-video payloads
are rejected so the caller can fall back to yt-dlp.
Args:
link: Direct download URL.
output_path: Destination file path.
headers: Optional HTTP headers.
timeout: Per-request timeout (seconds).
Returns:
True on a successful save, False when the link is not a
direct video or the download fails.
"""
try:
with self.session.get(
link,
headers=headers,
timeout=timeout,
stream=True,
) as response:
if not response.ok:
logger.debug(
"Direct stream HEAD returned %s for %s",
response.status_code, link[:80]
)
return False
content_type = response.headers.get("Content-Type", "")
if not (
content_type.startswith("video/")
or content_type == "application/octet-stream"
):
logger.debug(
"Direct stream skipped, Content-Type=%s",
content_type
)
return False
logger.info(
"Direct stream download starting (type=%s)",
content_type
)
with open(output_path, "wb") as fh:
for chunk in response.iter_content(chunk_size=1024 * 1024):
if self._cancel_flag.is_set():
logger.info(
"Cancellation detected during direct stream"
)
return False
if chunk:
fh.write(chunk)
return True
except requests.RequestException as exc:
logger.warning("Direct stream download failed: %s", exc)
return False
def download(
self,
base_directory: str,
@@ -260,6 +372,11 @@ class AniworldLoader(Loader):
) -> bool:
"""Download episode to specified directory.
Iterates the providers actually advertised on the episode page
(ordered by SUPPORTED_PROVIDERS preference), probing each URL
before attempting an extraction so dead providers are skipped
immediately instead of stalling yt-dlp on a 404.
Args:
base_directory: Base download directory path
serie_folder: Filesystem folder name (metadata only, used for
@@ -308,12 +425,78 @@ class AniworldLoader(Loader):
temp_path = os.path.join(temp_dir, output_file)
logger.debug("Temporary path: %s", temp_path)
for provider in self.SUPPORTED_PROVIDERS:
logger.debug("Attempting download with provider: %s", provider)
link, header = self._get_direct_link_from_provider(
candidate_providers = self._select_providers_for_episode(
season, episode, key, language
)
logger.debug("Direct link obtained from provider")
if not candidate_providers:
logger.error(
"No providers advertised for S%02dE%03d (%s) in %s",
season, episode, key, language
)
self.clear_cache()
return False
tried: list[str] = []
for provider_name, redirect_url in candidate_providers:
tried.append(provider_name)
logger.debug("Attempting download with provider: %s", provider_name)
probe_headers = {"User-Agent": self.RANDOM_USER_AGENT}
if not self._check_url_alive(
redirect_url,
headers=probe_headers,
timeout=self.DEFAULT_REQUEST_TIMEOUT,
):
logger.info(
"Skipping provider %s, redirect URL not reachable",
provider_name
)
continue
try:
resolved = self._resolve_direct_link(
redirect_url, provider_name
)
except Exception as exc:
logger.warning(
"Provider %s link resolution failed: %s: %s",
provider_name, type(exc).__name__, exc
)
continue
if resolved is None:
logger.info(
"Provider %s returned no direct link", provider_name
)
continue
link, header = resolved
if self._cancel_flag.is_set():
logger.info("Cancellation requested before download start")
_cleanup_temp_file(temp_path)
self.clear_cache()
return False
if self._try_direct_stream(
link,
temp_path,
header,
self.DEFAULT_REQUEST_TIMEOUT,
) and os.path.exists(temp_path):
logger.debug(
"Direct stream succeeded with provider %s", provider_name
)
shutil.copyfile(temp_path, output_path)
os.remove(temp_path)
logger.info(
"Download completed successfully (direct): %s",
output_file
)
self.clear_cache()
return True
_cleanup_temp_file(temp_path)
cancel_flag = self._cancel_flag
@@ -321,7 +504,6 @@ class AniworldLoader(Loader):
if cancel_flag.is_set():
logger.info("Cancellation detected in progress hook")
raise DownloadCancelled("Download cancelled by user")
# Fire the event for progress
self.events.download_progress(d)
ydl_opts = {
@@ -333,7 +515,6 @@ class AniworldLoader(Loader):
'nocheckcertificate': True,
'logger': logger,
'progress_hooks': [events_progress_hook],
# Use ffmpeg for HLS streams and transport stream format
'downloader': 'ffmpeg',
'hls_use_mpegts': True,
}
@@ -343,9 +524,11 @@ class AniworldLoader(Loader):
logger.debug("Using custom headers for download")
try:
logger.info("Starting download: %s", output_file)
logger.info(
"Starting yt-dlp download with %s: %s",
provider_name, output_file
)
logger.debug("Download link: %s...", link[:100])
logger.debug("YDL options: %s", ydl_opts)
with YoutubeDL(ydl_opts) as ydl:
info = ydl.extract_info(link, download=True)
@@ -356,39 +539,151 @@ class AniworldLoader(Loader):
if os.path.exists(temp_path):
logger.debug("Moving file from temp to final destination")
# Use copyfile instead of copy to avoid metadata permission issues
shutil.copyfile(temp_path, output_path)
os.remove(temp_path)
logger.info("Download completed successfully: %s", output_file)
logger.info(
"Download completed successfully: %s", output_file
)
self.clear_cache()
return True
else:
logger.error("Download failed: temp file not found at %s", temp_path)
logger.error(
"Download failed: temp file not found at %s", temp_path
)
except DownloadCancelled:
logger.info("Download cancelled by user")
_cleanup_temp_file(temp_path)
self.clear_cache()
return False
except BrokenPipeError as e:
except BrokenPipeError as exc:
logger.error(
"Broken pipe error with provider %s: %s. "
"This usually means the stream connection was closed.",
provider, e
"Broken pipe error with provider %s: %s",
provider_name, exc
)
_cleanup_temp_file(temp_path)
continue
except Exception as e:
except Exception as exc:
logger.error(
"YoutubeDL download failed with provider %s: %s: %s",
provider, type(e).__name__, e
provider_name, type(exc).__name__, exc
)
_cleanup_temp_file(temp_path)
continue
break
# If we get here, all providers failed
logger.error("All download providers failed")
logger.error(
"All download providers failed for S%02dE%03d (%s) in %s. "
"Tried: %s. Episode may be unavailable on the source site.",
season, episode, key, language, ", ".join(tried) or "none"
)
download_error_logger.error(
"All providers failed for %s S%02dE%03d (%s); tried=%s",
key, season, episode, language, tried
)
_cleanup_temp_file(temp_path)
self.clear_cache()
return False
def _select_providers_for_episode(
self,
season: int,
episode: int,
key: str,
language: str,
) -> list[tuple[str, str]]:
"""Return ``[(provider_name, redirect_url), ...]`` for an episode.
Filters by requested language and orders results by
``SUPPORTED_PROVIDERS`` preference so the failover chain matches
operator expectations. Returns an empty list when nothing is
advertised on the page.
"""
if not self.is_language(season, episode, key, language):
logger.warning(
"Language %s not advertised for S%02dE%03d (%s)",
language, season, episode, key
)
return []
language_code = self._get_language_key(language)
providers = self._get_provider_from_html(season, episode, key)
ordered: list[tuple[str, str]] = []
preferred = list(self.SUPPORTED_PROVIDERS)
for name in preferred:
lang_map = providers.get(name)
if lang_map and language_code in lang_map:
ordered.append((name, lang_map[language_code]))
for name, lang_map in providers.items():
if name in preferred:
continue
if language_code in lang_map:
ordered.append((name, lang_map[language_code]))
return ordered
def _resolve_direct_link(
self,
redirect_url: str,
provider_name: str,
) -> tuple[str, dict] | None:
"""Resolve a provider redirect URL into a direct stream link.
Follows the redirect to the embedded player, then delegates to a
provider-specific extractor (when registered) or returns the
embed URL itself so yt-dlp can attempt extraction.
Args:
redirect_url: AniWorld redirect URL.
provider_name: Provider key (e.g. ``"VOE"``).
Returns:
``(direct_link, headers)`` tuple or None when extraction fails.
"""
try:
embedded = self.session.get(
redirect_url,
timeout=self.DEFAULT_REQUEST_TIMEOUT,
headers={"User-Agent": self.RANDOM_USER_AGENT},
allow_redirects=True,
).url
except requests.RequestException as exc:
logger.warning(
"Failed resolving redirect for %s: %s", provider_name, exc
)
return None
try:
extractor = self.Providers.GetProvider(provider_name)
except (KeyError, AttributeError):
extractor = None
if extractor is not None:
try:
return extractor.get_link(
embedded, self.DEFAULT_REQUEST_TIMEOUT
)
except Exception as exc:
logger.warning(
"Custom extractor %s failed: %s",
provider_name, exc
)
return None
header_list = self.PROVIDER_HEADERS.get(provider_name)
header_dict = self._parse_provider_headers(header_list)
return embedded, header_dict
@staticmethod
def _parse_provider_headers(
header_list: list | None,
) -> dict[str, str]:
"""Convert legacy ``"Name: value"`` header strings to a dict."""
if not header_list:
return {}
parsed: dict[str, str] = {}
for entry in header_list:
if not isinstance(entry, str) or ":" not in entry:
continue
name, _, value = entry.partition(":")
parsed[name.strip()] = value.strip().strip('"')
return parsed
def get_site_key(self) -> str:
"""Get the site key for this provider."""
return "aniworld.to"

249
tests/unit/test_aniworld_provider.py
View File

@@ -1,9 +1,11 @@
"""Unit tests for aniworld_provider.py - Anime catalog scraping, episode listing, streaming link extraction."""
import json
import os
from unittest.mock import MagicMock, Mock, patch
import pytest
import requests
from src.core.providers.aniworld_provider import AniworldLoader
@@ -472,3 +474,250 @@ class TestAniworldEvents:
# Fire event - handler should NOT be called
loader.events.download_progress({"status": "downloading"})
handler.assert_not_called()
class TestAniworldHealthCheck:
"""Tests for the _check_url_alive HEAD probe."""
def test_returns_true_on_200(self, loader):
loader.session.head.return_value = MagicMock(status_code=200)
assert loader._check_url_alive("https://provider/x") is True
def test_returns_false_on_404(self, loader):
loader.session.head.return_value = MagicMock(status_code=404)
assert loader._check_url_alive("https://provider/x") is False
def test_returns_false_on_403(self, loader):
loader.session.head.return_value = MagicMock(status_code=403)
assert loader._check_url_alive("https://provider/x") is False
def test_falls_back_to_get_when_head_disallowed(self, loader):
loader.session.head.return_value = MagicMock(status_code=405)
get_resp = MagicMock(status_code=200)
get_resp.close = MagicMock()
loader.session.get.return_value = get_resp
assert loader._check_url_alive("https://provider/x") is True
loader.session.get.assert_called_once()
def test_returns_false_on_connection_error(self, loader):
loader.session.head.side_effect = requests.ConnectionError("boom")
assert loader._check_url_alive("https://provider/x") is False
class TestAniworldDirectStream:
"""Tests for the _try_direct_stream fast-path."""
def _build_response(self, status, content_type, body=b""):
resp = MagicMock()
resp.ok = status < 400
resp.status_code = status
resp.headers = {"Content-Type": content_type}
resp.iter_content = MagicMock(return_value=[body])
resp.__enter__ = MagicMock(return_value=resp)
resp.__exit__ = MagicMock(return_value=False)
return resp
def test_skips_non_video_content(self, loader, tmp_path):
target = tmp_path / "out.mp4"
loader.session.get.return_value = self._build_response(
200, "text/html"
)
assert loader._try_direct_stream(
"https://x", str(target), None, 10
) is False
assert not target.exists()
def test_writes_video_content(self, loader, tmp_path):
target = tmp_path / "out.mp4"
loader.session.get.return_value = self._build_response(
200, "video/mp4", body=b"abc123"
)
assert loader._try_direct_stream(
"https://x", str(target), None, 10
) is True
assert target.read_bytes() == b"abc123"
def test_returns_false_on_http_error(self, loader, tmp_path):
target = tmp_path / "out.mp4"
loader.session.get.return_value = self._build_response(
404, "video/mp4"
)
assert loader._try_direct_stream(
"https://x", str(target), None, 10
) is False
def test_returns_false_on_request_exception(self, loader, tmp_path):
loader.session.get.side_effect = requests.RequestException("nope")
assert loader._try_direct_stream(
"https://x", str(tmp_path / "out.mp4"), None, 10
) is False
class TestAniworldProviderSelection:
"""Tests for _select_providers_for_episode ordering and filtering."""
def test_orders_by_supported_preference(self, loader):
loader.is_language = MagicMock(return_value=True)
loader._get_provider_from_html = MagicMock(return_value={
"Vidoza": {1: "https://aniworld.to/redirect/2"},
"VOE": {1: "https://aniworld.to/redirect/1"},
})
result = loader._select_providers_for_episode(1, 1, "k", "German Dub")
assert [name for name, _ in result] == ["VOE", "Vidoza"]
def test_filters_by_language(self, loader):
loader.is_language = MagicMock(return_value=True)
loader._get_provider_from_html = MagicMock(return_value={
"VOE": {2: "https://aniworld.to/redirect/1"}, # English only
})
result = loader._select_providers_for_episode(1, 1, "k", "German Dub")
assert result == []
def test_returns_empty_when_language_unavailable(self, loader):
loader.is_language = MagicMock(return_value=False)
loader._get_provider_from_html = MagicMock()
result = loader._select_providers_for_episode(1, 1, "k", "German Dub")
assert result == []
loader._get_provider_from_html.assert_not_called()
class TestAniworldDownloadFailover:
"""Tests for the failover rotation in download()."""
@pytest.fixture
def patched_loader(self, loader, tmp_path):
"""Loader with side-effect heavy methods stubbed."""
loader.get_title = MagicMock(return_value="Anime")
loader._select_providers_for_episode = MagicMock(return_value=[
("VOE", "https://aniworld.to/redirect/1"),
("Doodstream", "https://aniworld.to/redirect/2"),
])
loader._check_url_alive = MagicMock(return_value=True)
loader._try_direct_stream = MagicMock(return_value=False)
loader.clear_cache = MagicMock()
loader._resolve_direct_link = MagicMock(
return_value=("https://cdn/video.m3u8", {"Referer": "https://x"})
)
return loader
def test_skips_provider_when_url_dead(self, patched_loader, tmp_path):
# First provider URL fails health check, second succeeds and downloads
patched_loader._check_url_alive.side_effect = [False, True]
def fake_ytdl(opts):
outpath = opts["outtmpl"]
os.makedirs(os.path.dirname(outpath), exist_ok=True)
with open(outpath, "wb") as fh:
fh.write(b"data")
ydl = MagicMock()
ydl.__enter__ = MagicMock(return_value=ydl)
ydl.__exit__ = MagicMock(return_value=False)
ydl.extract_info = MagicMock(return_value={"title": "t"})
return ydl
with patch(
"src.core.providers.aniworld_provider.YoutubeDL",
side_effect=fake_ytdl,
):
result = patched_loader.download(
str(tmp_path), "Anime", 1, 1, "k", "German Dub"
)
assert result is True
assert patched_loader._check_url_alive.call_count == 2
# Only second provider (Doodstream) attempted resolve
patched_loader._resolve_direct_link.assert_called_once_with(
"https://aniworld.to/redirect/2", "Doodstream"
)
def test_falls_back_to_next_provider_on_ytdl_error(
self, patched_loader, tmp_path
):
calls = {"n": 0}
def fake_ytdl(opts):
calls["n"] += 1
if calls["n"] == 1:
raise Exception("HTTP 404 from VOE")
outpath = opts["outtmpl"]
os.makedirs(os.path.dirname(outpath), exist_ok=True)
with open(outpath, "wb") as fh:
fh.write(b"ok")
ydl = MagicMock()
ydl.__enter__ = MagicMock(return_value=ydl)
ydl.__exit__ = MagicMock(return_value=False)
ydl.extract_info = MagicMock(return_value={"title": "t"})
return ydl
with patch(
"src.core.providers.aniworld_provider.YoutubeDL",
side_effect=fake_ytdl,
):
result = patched_loader.download(
str(tmp_path), "Anime", 1, 1, "k", "German Dub"
)
assert result is True
assert calls["n"] == 2
def test_uses_direct_stream_when_available(
self, patched_loader, tmp_path
):
def write_direct(link, output, headers, timeout):
os.makedirs(os.path.dirname(output), exist_ok=True)
with open(output, "wb") as fh:
fh.write(b"vid")
return True
patched_loader._try_direct_stream.side_effect = write_direct
with patch(
"src.core.providers.aniworld_provider.YoutubeDL"
) as mock_ydl:
result = patched_loader.download(
str(tmp_path), "Anime", 1, 1, "k", "German Dub"
)
assert result is True
mock_ydl.assert_not_called()
def test_returns_false_when_all_providers_fail(
self, patched_loader, tmp_path, caplog
):
with patch(
"src.core.providers.aniworld_provider.YoutubeDL",
side_effect=Exception("HTTP 404"),
):
result = patched_loader.download(
str(tmp_path), "Anime", 1, 1, "k", "German Dub"
)
assert result is False
assert "All download providers failed" in caplog.text
# Both providers attempted
assert patched_loader._resolve_direct_link.call_count == 2
def test_returns_false_when_no_providers_advertised(
self, patched_loader, tmp_path, caplog
):
patched_loader._select_providers_for_episode.return_value = []
result = patched_loader.download(
str(tmp_path), "Anime", 1, 1, "k", "German Dub"
)
assert result is False
assert "No providers advertised" in caplog.text
class TestAniworldHeaderParsing:
"""_parse_provider_headers normalizes legacy strings to dict."""
def test_parses_referer(self):
result = AniworldLoader._parse_provider_headers(
['Referer: "https://vidmoly.to"']
)
assert result == {"Referer": "https://vidmoly.to"}
def test_handles_none(self):
assert AniworldLoader._parse_provider_headers(None) == {}
def test_skips_malformed_entries(self):
result = AniworldLoader._parse_provider_headers(
["not-a-header", "Key: value"]
)
assert result == {"Key": "value"}