Convert tsv files in "Kaggle Display Advertising Challenge Dataset" into vwopal wabbit format files

tsv_to_vw.py

#!/usr/bin/env python
# conding:utf-8

import sys

def main():
    for index, line in enumerate(sys.stdin):
        elems = line.rstrip().split("\t")
        label, nums, categories = elems[0], elems[1:14], elems[14:40]
        vw_label = "1" if label == "1" else "-1"
        num_fstr = " ".join(["I{}:{}".format(i, v) for i, v in enumerate(nums) if len(v) > 0])
        cat_fstr = " ".join(["{}".format(v) for v in categories if len(v) > 0])
        print("{} '{} |i {} |c {}".format(vw_label, index, num_fstr, cat_fstr))

if __name__ == "__main__":
    main()