← Back to branch summary

~grosmoteur/grosmoteur/trunk

« back to all changes in this revision

Viewing changes to lib/spiderConfigurations/libe.gro.cfg

  • Committer: kim
  • Date: 2014-02-22 17:13:52 UTC
  • Revision ID: kmgrds@gmail.com-20140222171352-iodi07d46m2boftw
2014

Show diffs side-by-side

added added

removed removed

Lines of Context:
lib/spiderConfigurations/libe.gro.cfg
 
1
[spider configuration]
 
2
startURL = liberation.fr
 
3
startUrlFile = ""
 
4
searchEngineQuery = ""
 
5
searchEngineFilter = ""
 
6
searchEngineAppId = /oQLMwXhI0Ygpz/1Ma/Oce+VrWJL3mbjsJL4aA3Jo78=
 
7
downloadURL = /liberation.fr
 
8
downloadAvoidURL = (\.blog|token)
 
9
pageRestriction = ""
 
10
linkDomain = liberation.fr
 
11
linkAvoidURL = ""
 
12
defaultEncoding = ""
 
13
httpProxyName = http://127.0.0.1:8087
 
14
httpsProxyName = http://127.0.0.1:8087
 
15
pageLinkRestrict = yes
 
16
pageIgnoreCase = yes
 
17
takePdf = no
 
18
redirectMustMatch = yes
 
19
obeyRobots = no
 
20
downloadRestriction = yes
 
21
linkRestriction = yes
 
22
spiderTraps = yes
 
23
everyServer = yes
 
24
maxOneServer = no
 
25
followRedirects = yes
 
26
timeout = yes
 
27
proxyGroupBox = no
 
28
dataAppend = no
 
29
dataOverwrite = no
 
30
dataErase = yes
 
31
fromdatabase = no
 
32
startWithURL = yes
 
33
startWithUrlFile = no
 
34
startWithSearchEngine = no
 
35
onlySearchEngineResults = yes
 
36
yContinueFromResults = no
 
37
breadthFirst = yes
 
38
depthFirst = no
 
39
expert = yes
 
40
levelFrom = -1
 
41
levelTo = -1
 
42
maxPages = 25
 
43
maxSentences = 0
 
44
maxSubdomains = 0
 
45
maxMb = 0
 
46
diskMin = 300
 
47
everyServerSeconds = 0
 
48
pageByServer = 0
 
49
timeoutsec = 2
 
50
trytimes = 0
 
51
nbThreads = 6
 
52
location = automatic location detection
 
53
language = any language
 
54
userAgent = Mozilla/5.0 (compatible; gromoteur/1.0; http://gromoteur.ilpga.fr/)