catdog2/urlbot-native-trex
1
0
Fork 0
mirror of http://aero2k.de/t/repos/urlbot-native.git synced 2017-09-06 15:25:38 +02:00
Code Issues Releases Activity

remove unused str_sim part

Browse Source
This commit is contained in:
urlbot
2014-12-07 14:06:58 +01:00
parent 8d31036e0d
commit 894f19f16e
1 changed files with 0 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
urlbot.py
View File

@@ -4,7 +4,6 @@
import sys, os, stat, re, time, pickle, random import sys, os, stat, re, time, pickle, random
import urllib.request, urllib.parse, urllib.error, html.parser import urllib.request, urllib.parse, urllib.error, html.parser
from common import * from common import *
from strsim import str_sim
try: try:
from local_config import conf, set_conf from local_config import conf, set_conf
@@ -179,18 +178,6 @@ def extract_url(data):
if 0 == status: if 0 == status:
title = title.strip() title = title.strip()
lev_url = re.sub(r'https?://[^/]*/', '', url)
lev_res = levenshtein(lev_url, title)
sim = str_sim(title, lev_url)
sim_len_title = len(sim)
sim_len_url = len(sim[0])
sim_sum = sum([sum(a) for a in sim])
obj = conf_load()
obj['lev'].append((lev_res, title, url))
obj['sim'].append((sim_sum, sim_len_title, sim_len_url, title, url))
conf_save(obj)
message = 'Title: %s: %s' %(title, url) message = 'Title: %s: %s' %(title, url)
elif 1 == status: elif 1 == status: