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

madou priority against javdb

Browse Source
This commit is contained in:
lededev
2022-03-06 01:31:43 +08:00
parent 9e332b0d02
commit 54d8f3af87
2 changed files with 7 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
WebCrawler/__init__.py
View File

@@ -95,6 +95,10 @@ def get_data_from_json(file_number, oCC):
"rj" in lo_file_number or "vj" in lo_file_number "rj" in lo_file_number or "vj" in lo_file_number
): ):
sources.insert(0, sources.pop(sources.index("dlsite"))) sources.insert(0, sources.pop(sources.index("dlsite")))
elif "madou" in sources and (
re.match(r"^md[0-9]{4}$", lo_file_number)
):
sources.insert(0, sources.pop(sources.index("madou")))
elif re.match(r"^[a-z0-9]{3,}$", lo_file_number): elif re.match(r"^[a-z0-9]{3,}$", lo_file_number):
if "javdb" in sources: if "javdb" in sources:
sources.insert(0, sources.pop(sources.index("javdb"))) sources.insert(0, sources.pop(sources.index("javdb")))

5
WebCrawler/madou.py
View File

@@ -1,3 +1,5 @@
import sys
sys.path.append('../')
from bs4 import BeautifulSoup # need install from bs4 import BeautifulSoup # need install
from lxml import etree # need install from lxml import etree # need install
from pyquery import PyQuery as pq # need install from pyquery import PyQuery as pq # need install
@@ -5,10 +7,8 @@ from ADC_function import *
import json import json
import re import re
from lib2to3.pgen2 import parse from lib2to3.pgen2 import parse
import sys
from urllib.parse import urlparse, unquote from urllib.parse import urlparse, unquote
sys.path.append('../')
def getActorPhoto(html): def getActorPhoto(html):
@@ -162,3 +162,4 @@ def main(number):
if __name__ == '__main__': if __name__ == '__main__':
print(main('MD0094')) print(main('MD0094'))
print(main('MD0222'))