KevinMidboe/seasonedShows
1
0
Fork 0
Code Issues 20 Pull Requests 2 Packages Projects Releases 8 Wiki Activity

More general re matching for date and also returns empty dict of no candidates are found.

Browse Source
This commit is contained in:
Kevin
2017-10-22 18:50:50 +02:00
parent 0e3189ff11
commit f15df29801
1 changed files with 8 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
app/pirateSearch.py
View File

@@ -3,7 +3,7 @@
# @Author: KevinMidboe # @Author: KevinMidboe
# @Date: 2017-10-12 11:55:03 # @Date: 2017-10-12 11:55:03
# @Last Modified by: KevinMidboe # @Last Modified by: KevinMidboe
# @Last Modified time: 2017-10-21 12:25:09 # @Last Modified time: 2017-10-22 18:49:06
import sys, logging, re, json import sys, logging, re, json
from urllib import parse, request from urllib import parse, request
@@ -58,6 +58,9 @@ def return_re_match(string, re_statement):
m = re.search(re_statement, string) m = re.search(re_statement, string)
if 'Y-day' in m.group(): if 'Y-day' in m.group():
return datetime.timedelta(days=1).strftime('%m-%d %Y')
if 'Today' in m.group():
return datetime.datetime.now().strftime('%m-%d %Y') return datetime.datetime.now().strftime('%m-%d %Y')
return sanitize(m.group(), '\xa0', ' ') return sanitize(m.group(), '\xa0', ' ')
@@ -210,7 +213,7 @@ class piratebay(object):
info_text = torrentElement.find('font', class_='detDesc').get_text() info_text = torrentElement.find('font', class_='detDesc').get_text()
date = return_re_match(info_text, r"(\d+\-\d+\s\d+)|(Y\-day\s\d{2}\:\d{2})") date = return_re_match(info_text, r"(\d+\-\d+)|(Y\-day|Today)")
size = return_re_match(info_text, r"(\d+(\.\d+)?\s[a-zA-Z]+)") size = return_re_match(info_text, r"(\d+(\.\d+)?\s[a-zA-Z]+)")
# COULD NOT FIND HREF! # COULD NOT FIND HREF!
@@ -294,6 +297,8 @@ def chooseCandidate(torrent_list):
def searchTorrentSite(query, site='piratebay'): def searchTorrentSite(query, site='piratebay'):
pirate = piratebay() pirate = piratebay()
torrents_found = pirate.search(query, page=0, multiple_pages=3, sort='size') torrents_found = pirate.search(query, page=0, multiple_pages=3, sort='size')
candidates = {}
if (torrents_found):
candidates = chooseCandidate(torrents_found) candidates = chooseCandidate(torrents_found)
print(json.dumps(candidates)) print(json.dumps(candidates))