bacow/AV_Data_Capture
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update javdb source

Browse Source
This commit is contained in:
yoshiko2
2021-05-20 20:41:14 +08:00
parent 210b2cf6b8
commit ab50ec6b30
1 changed files with 3 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
WebCrawler/javdb.py
View File

@@ -51,7 +51,7 @@ def getActorPhoto(html): #//*[@id="star_qdt"]/li/a/img
actor = actor_prether.findall(actoralls) actor = actor_prether.findall(actoralls)
actor_photo = {} actor_photo = {}
for i in actor: for i in actor:
actor_photo[i[1]] = getaphoto('https://javdb.com'+i[0]) actor_photo[i[1]] = getaphoto('https://' + javdb_site + '.com'+i[0])
return actor_photo return actor_photo
@@ -217,7 +217,7 @@ def main(number):
javdb_url = 'https://' + javdb_site + '.com/search?q=' + number + '&f=all' javdb_url = 'https://' + javdb_site + '.com/search?q=' + number + '&f=all'
query_result = get_html(javdb_url, cookies=javdb_cookies) query_result = get_html(javdb_url, cookies=javdb_cookies)
except: except:
query_result = get_html('https://javdb8.com/search?q=' + number + '&f=all') query_result = get_html('https://javdb9.com/search?q=' + number + '&f=all')
html = etree.fromstring(query_result, etree.HTMLParser()) # //table/tr[1]/td[1]/text() html = etree.fromstring(query_result, etree.HTMLParser()) # //table/tr[1]/td[1]/text()
# javdb sometime returns multiple results, # javdb sometime returns multiple results,
# and the first elememt maybe not the one we are looking for # and the first elememt maybe not the one we are looking for
@@ -233,7 +233,7 @@ def main(number):
javdb_detail_url = 'https://' + javdb_site + '.com' + correct_url javdb_detail_url = 'https://' + javdb_site + '.com' + correct_url
detail_page = get_html(javdb_detail_url, cookies=javdb_cookies) detail_page = get_html(javdb_detail_url, cookies=javdb_cookies)
except: except:
detail_page = get_html('https://javdb8.com' + correct_url) detail_page = get_html('https://' + javdb_site + '.com' + correct_url)
# no cut image by default # no cut image by default
imagecut = 3 imagecut = 3