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

Update 2.2

Browse Source
This commit is contained in:
Yoshiko 2020-01-22 02:19:40 +08:00 committed by GitHub
parent 0b5d68732f
commit 70f1b16b3c
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
4 changed files with 13 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
AV_Data_Capture.py
View File

@ -14,7 +14,7 @@ os.chdir(os.getcwd())
# ============global var=========== # ============global var===========
version='2.1' version='2.2'
config = ConfigParser() config = ConfigParser()
config.read(config_file, encoding='UTF-8') config.read(config_file, encoding='UTF-8')

2
config.ini
View File

@ -5,7 +5,7 @@ success_output_folder=JAV_output
soft_link=0 soft_link=0
[proxy] [proxy]
proxy= proxy=127.0.0.1:1081
timeout=10 timeout=10
retry=3 retry=3

17
javdb.py
View File

@ -3,6 +3,9 @@ from lxml import etree
import json import json
from bs4 import BeautifulSoup from bs4 import BeautifulSoup
from ADC_function import * from ADC_function import *
# import sys
# import io
# sys.stdout = io.TextIOWrapper(sys.stdout.buffer, errors = 'replace', line_buffering = True)
def getTitle(a): def getTitle(a):
try: try:
@ -61,15 +64,13 @@ def getTag(a):
return str(result1 + result2).strip('+').replace(",\\xa0", "").replace("'", "").replace(' ', '').replace(',,', '').lstrip(',') return str(result1 + result2).strip('+').replace(",\\xa0", "").replace("'", "").replace(' ', '').replace(',,', '').lstrip(',')
def getCover_small(a): def getCover_small(a):
html = etree.fromstring(a, etree.HTMLParser()) # //table/tr[1]/td[1]/text() html = etree.fromstring(a, etree.HTMLParser()) # //table/tr[1]/td[1]/text()
result = 'http:' + html.xpath( result = html.xpath("//div[@class='item-image fix-scale-cover']/img/@src")[0]
'//div[@id=\'videos\']/div[@class=\'grid columns\']/div[@class=\'grid-item column\'][1]/a[' if not 'https' in result:
'@class=\'box\']/div[@class=\'item-image fix-scale-cover\']/img/@src')[0] result = 'https:' + result
return result return result
def getCover(htmlcode): def getCover(htmlcode):
html = etree.fromstring(htmlcode, etree.HTMLParser()) html = etree.fromstring(htmlcode, etree.HTMLParser())
result = str(html.xpath('/html/body/section/div/div[2]/div[1]/a/img/@src')).strip(" ['']") result = str(html.xpath("//div[@class='column column-video-cover']/a/img/@src")).strip(" ['']")
if result == '':
result = str(html.xpath('/html/body/section/div/div[4]/div[1]/a/img/@src')).strip(" ['']")
return result return result
def getDirector(a): def getDirector(a):
html = etree.fromstring(a, etree.HTMLParser()) # //table/tr[1]/td[1]/text() html = etree.fromstring(a, etree.HTMLParser()) # //table/tr[1]/td[1]/text()
@ -138,7 +139,7 @@ def main(number):
'label': getLabel(b), 'label': getLabel(b),
'year': getYear(getRelease(b)), # str(re.search('\d{4}',getRelease(a)).group()), 'year': getYear(getRelease(b)), # str(re.search('\d{4}',getRelease(a)).group()),
'actor_photo': getActorPhoto(getActor(b)), 'actor_photo': getActorPhoto(getActor(b)),
'website': 'https://javdb.com' + result1, 'website': 'https://javdb3.com' + result1,
'source': 'javdb.py', 'source': 'javdb.py',
} }
if getNum(b) != number: # 与搜索到的番号不匹配 if getNum(b) != number: # 与搜索到的番号不匹配
@ -149,4 +150,4 @@ def main(number):
# main('DV-1562') # main('DV-1562')
# input("[+][+]Press enter key exit, you can check the error messge before you exit.\n[+][+]按回车键结束，你可以在结束之前查看和错误信息。") # input("[+][+]Press enter key exit, you can check the error messge before you exit.\n[+][+]按回车键结束，你可以在结束之前查看和错误信息。")
# print(get_html('https://javdb1.com/v/WwZ0Q')) # print(main('YMDD-178'))

4
update_check.json
View File

@ -1,5 +1,5 @@
{ {
"version": "2.1", "version": "2.2",
"version_show":"2.1", "version_show":"2.2",
"download": "https://github.com/yoshiko2/AV_Data_Capture/releases" "download": "https://github.com/yoshiko2/AV_Data_Capture/releases"
} }