# WikiCFP crawl script

Usage: 

$ ./all.sh [total count of wikicfp]

Must know:
1. You can see the event id of the latest CFP here [http://www.wikicfp.com/cfp/allcfp] and use it as total count.
2. The script honors the requirement of WikiCFP that crawler should crawl no more than one page every 5 seconds [http://www.wikicfp.com/cfp/data.jsp]. And I do not recommend you to modify it.
3. The script takes approximately 5*[total count] seconds to crawl, so it's going to be a long time. 
4. You will need python 2.x (2.7.5 tested), pycurl, beautifulsoup4, bash and wget to run the script.

The script will download all cfps and conference series, do information extraction and generate following files:

- item.txt

  It is a tsv (tab separated vetor) file in which every line is:

  [conferenceID] [seriesID] [conferenceDeadline] [conferenceName]

- meta.txt

  It is a json file in which every line is a json string that described several meta data of a conference. The itemID part is identical to conferenceID in item.txt.

- posted.txt

  It is a tsv file in which every line is a post behaviour (who posted what):

  [userID] [conferenceID] 2

  The third column is always 2.

- series.tsv

  It is a tsv file in which every line is:

  [conferenceID] [seriesID] [conferenceName] [seriesName]

- series.txt

  It is a tsv file in which every line is:

  [seriesID] [seriesName]

- tracked.txt

  It is a tsv file in which every line is a track behaviour (who tracked what):

  [userID] [conferenceID] 1

  The third column is always 1.