Mirrors/gallery-dl
1
0
Fork 0
mirror of https://github.com/mikf/gallery-dl.git synced 2024-11-21 18:22:30 +01:00
Code Issues Releases Wiki Activity

[bilibili] support large articles

Browse Source
This commit is contained in:
hdk5 2024-11-10 15:18:03 +02:00
parent 6eef3e3495
commit fc59e0fb14
1 changed files with 19 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
gallery_dl/extractor/bilibili.py
View File

@ -9,7 +9,7 @@
"""Extractors for https://www.bilibili.com/""" """Extractors for https://www.bilibili.com/"""
from .common import Extractor, Message from .common import Extractor, Message
from .. import text, exception from .. import text, util, exception
class BilibiliExtractor(Extractor): class BilibiliExtractor(Extractor):
@ -50,16 +50,21 @@ class BilibiliArticleExtractor(BilibiliExtractor):
def items(self): def items(self):
article = self.api.article(self.item) article = self.api.article(self.item)
article["username"] = article["modules"]["module_author"]["name"]
article["id"] = article["id_str"]
dynamic_major = article["modules"]["module_dynamic"]["major"] # Flatten modules list
if dynamic_major["type"] == "MAJOR_TYPE_OPUS": modules = {}
urls = [pic["url"] for pic in dynamic_major["opus"]["pics"]] for module in article["detail"]["modules"]:
else: del module['module_type']
urls = [] modules.update(module)
self.log.warning("%s: Unsupported article type '%s'", article["detail"]["modules"] = modules
article["id"], dynamic_major["type"])
article["username"] = modules["module_author"]["name"]
urls = []
for paragraph in modules['module_content']['paragraphs']:
pics = paragraph.get('pic', {}).get('pics', [])
for pic in pics:
urls.append(pic['url'])
yield Message.Directory, article yield Message.Directory, article
for article["num"], url in enumerate(urls, 1): for article["num"], url in enumerate(urls, 1):
@ -94,9 +99,7 @@ class BilibiliAPI():
break break
def article(self, article_id): def article(self, article_id):
endpoint = "/detail" url = "https://www.bilibili.com/opus/{}".format(article_id)
params = { response = self.extractor.request(url)
"id": article_id, return util.json_loads(text.extr(
"features": "itemOpusStyle", response.text, "window.__INITIAL_STATE__=", ";"))
}
return self._call(endpoint, params)["data"]["item"]