Subf2m Provider: improve episode matching (#2081)

This commit is contained in:
Vitiko 2023-03-03 15:51:38 -04:00
parent 248e49de76
commit ad13f79d19
2 changed files with 61 additions and 3 deletions

View File

@ -5,8 +5,15 @@ import functools
import logging
import re
import time
import ssl
import urllib.parse
from urllib3 import poolmanager
from guessit import guessit
from requests import Session
from requests.adapters import HTTPAdapter
from bs4 import BeautifulSoup as bso
from guessit import guessit
from requests import Session
@ -113,6 +120,20 @@ _LANGUAGE_MAP = {
}
class _Adapter(HTTPAdapter):
def init_poolmanager(self, connections, maxsize, block=False):
ctx = ssl.create_default_context()
ctx.set_ciphers("DEFAULT@SECLEVEL=0")
ctx.check_hostname = False
self.poolmanager = poolmanager.PoolManager(
num_pools=connections,
maxsize=maxsize,
block=block,
ssl_version=ssl.PROTOCOL_TLS,
ssl_context=ctx,
)
class Subf2mProvider(Provider):
provider_name = "subf2m"
@ -141,6 +162,8 @@ class Subf2mProvider(Provider):
def initialize(self):
self._session = Session()
self._session.mount("https://", _Adapter())
self._session.verify = self._verify_ssl
self._session.headers.update({"user-agent": "Bazarr"})
@ -283,8 +306,11 @@ class Subf2mProvider(Provider):
if not clean_text:
continue
# It will return list values
guess = _memoized_episode_guess(clean_text)
# First try with the special episode matches for subf2m
guess = _get_episode_from_release(clean_text)
if guess is None:
guess = _memoized_episode_guess(clean_text)
if "season" not in guess:
if "complete series" in clean_text.lower():
@ -390,6 +416,24 @@ def _memoized_episode_guess(content):
)
_EPISODE_SPECIAL_RE = re.compile(
r"(season|s)\s*?(?P<x>\d{,2})\s?[-]\s?(?P<y>\d{,2})", flags=re.IGNORECASE
)
def _get_episode_from_release(release: str):
match = _EPISODE_SPECIAL_RE.search(release)
if match is None:
return None
try:
season, episode = [int(item) for item in match.group("x", "y")]
except (IndexError, ValueError):
return None
return {"season": [season], "episode": [episode]}
def _get_subtitle_from_item(item, language, episode_number=None):
release_info = [
rel.text.strip() for rel in item.find("ul", {"class": "scrolllist"})

View File

@ -1,5 +1,5 @@
import pytest
from subliminal_patch.providers import subf2m
from subliminal_patch.providers.subf2m import Subf2mProvider
from subliminal_patch.providers.subf2m import Subf2mSubtitle
from subzero.language import Language
@ -188,3 +188,17 @@ def test_download_subtitle_episode_with_title():
with Subf2mProvider() as provider:
provider.download_subtitle(sub)
assert sub.is_valid()
def test_get_episode_from_release():
assert subf2m._get_episode_from_release(
"Vinland Saga Season 2 - 05 [Crunchyroll][Crunchyroll] Vinland Saga Season 2 - 05"
) == {"season": [2], "episode": [5]}
def test_get_episode_from_release_return_none():
assert subf2m._get_episode_from_release("Vinland Saga Season 2 - Foo") is None
def test_get_episode_from_release_w_empty_match_return_none():
assert subf2m._get_episode_from_release("Vinland Saga - 02") is None