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

Update mgstage.py

Browse Source
This commit is contained in:
Feng4
2020-12-27 00:28:47 +08:00
committed by GitHub
parent 4c9cef5ec7
commit 4816affc38
1 changed files with 13 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
WebCrawler/mgstage.py
View File

@@ -95,6 +95,18 @@ def getSeries(a):
result2 = str(html.xpath('//th[contains(text(),"シリーズ")]/../td/text()')).strip(" ['']").strip('\\n ').strip( result2 = str(html.xpath('//th[contains(text(),"シリーズ")]/../td/text()')).strip(" ['']").strip('\\n ').strip(
'\\n') '\\n')
return str(result1 + result2).strip('+').replace("', '", '').replace('"', '') return str(result1 + result2).strip('+').replace("', '", '').replace('"', '')
def getExtrafanart(htmlcode): # 获取剧照
html_pather = re.compile(r'<dd>\s*?<ul>[\s\S]*?</ul>\s*?</dd>')
html = html_pather.search(htmlcode)
if html:
html = html.group()
extrafanart_pather = re.compile(r'<a class=\"sample_image\" href=\"(.*?)\"')
extrafanart_imgs = extrafanart_pather.findall(html)
if extrafanart_imgs:
return extrafanart_imgs
return ''
def main(number2): def main(number2):
number=number2.upper() number=number2.upper()
htmlcode=str(get_html('https://www.mgstage.com/product/product_detail/'+str(number)+'/',cookies={'adc':'1'})) htmlcode=str(get_html('https://www.mgstage.com/product/product_detail/'+str(number)+'/',cookies={'adc':'1'}))
@@ -115,6 +127,7 @@ def main(number2):
'imagecut': 0, 'imagecut': 0,
'tag': getTag(a), 'tag': getTag(a),
'label':getLabel(a), 'label':getLabel(a),
'extrafanart': getExtrafanart(htmlcode),
'year': getYear(getRelease(a)), # str(re.search('\d{4}',getRelease(a)).group()), 'year': getYear(getRelease(a)), # str(re.search('\d{4}',getRelease(a)).group()),
'actor_photo': '', 'actor_photo': '',
'website':'https://www.mgstage.com/product/product_detail/'+str(number)+'/', 'website':'https://www.mgstage.com/product/product_detail/'+str(number)+'/',