NussNougate/nhentai
4
0
Fork 0
mirror of https://github.com/RicterZ/nhentai.git synced 2025-11-04 02:50:55 +01:00
Code Issues Packages Projects Releases Wiki Activity

tag page download #40

Browse Source
This commit is contained in:
RicterZ
2019-01-15 21:12:20 +08:00
parent b2d622f11a
commit 5c3dace937
1 changed files with 13 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

25
nhentai/parser.py
View File

@@ -281,20 +281,21 @@ def tag_parser(tag_name, max_page=1):
tag_name = tag_name.lower() tag_name = tag_name.lower()
tag_name = tag_name.replace(' ', '-') tag_name = tag_name.replace(' ', '-')
logger.info('Searching for doujinshi with tag \'{0}\''.format(tag_name)) for p in range(1, max_page + 1):
response = request('get', url='%s/%s' % (constant.TAG_URL, tag_name)).content logger.debug('Fetching page {0} for doujinshi with tag \'{1}\''.format(p, tag_name))
response = request('get', url='%s/%s?page=%d' % (constant.TAG_URL, tag_name, p)).content
html = BeautifulSoup(response, 'html.parser') html = BeautifulSoup(response, 'html.parser')
doujinshi_items = html.find_all('div', attrs={'class': 'gallery'}) doujinshi_items = html.find_all('div', attrs={'class': 'gallery'})
if not doujinshi_items: if not doujinshi_items:
logger.error('Cannot find doujinshi id of tag \'{0}\''.format(tag_name)) logger.error('Cannot find doujinshi id of tag \'{0}\''.format(tag_name))
return return
for i in doujinshi_items[:2]: for i in doujinshi_items:
doujinshi_id = i.a.attrs['href'].strip('/g') doujinshi_id = i.a.attrs['href'].strip('/g')
doujinshi_title = i.a.text.strip() doujinshi_title = i.a.text.strip()
doujinshi_title = doujinshi_title if len(doujinshi_title) < 85 else doujinshi_title[:82] + '...' doujinshi_title = doujinshi_title if len(doujinshi_title) < 85 else doujinshi_title[:82] + '...'
result.append({'title': doujinshi_title, 'id': doujinshi_id}) result.append({'title': doujinshi_title, 'id': doujinshi_id})
if not result: if not result:
logger.warn('No results for tag \'{}\''.format(tag_name)) logger.warn('No results for tag \'{}\''.format(tag_name))