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

remove not useful text in source dlsite

Browse Source
This commit is contained in:
yoshiko2
2022-05-20 20:06:13 +08:00
parent 964518572c
commit da76d33b8b
1 changed files with 1 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
WebCrawler/dlsite.py
View File

@@ -9,6 +9,7 @@ def getTitle(html):
result = str(html.xpath('/html/head/title/text()')[0]) result = str(html.xpath('/html/head/title/text()')[0])
result = result[:result.rfind(' | DLsite')] result = result[:result.rfind(' | DLsite')]
result = result[:result.rfind(' [')] result = result[:result.rfind(' [')]
result = result.replace('【HD版】', '')
return result return result
def getActor(html): # //*[@id="center_column"]/div[2]/div[1]/div/table/tbody/tr[1]/td/text() def getActor(html): # //*[@id="center_column"]/div[2]/div[1]/div/table/tbody/tr[1]/td/text()
try: try: