Kaggle input data fromatter for Home Depot Product Search Relevance

attribute_flattener.py

import sys

"""
usage: python attribute_flattener.py [include-keys=yes] [override filename]
generates <file_name>.out file
contents are all attributes are flattened to single line per key

NOte: Keep attributes.csv in current folder and run 
python attribute_flattener.py
"""
include_key = False
file_name = 'attributes.csv'

if len(sys.argv) > 1 and sys.argv[1] == 'include-key=yes':
		include_key = True
if len(sys.argv) > 2 :
		file_name = sys.argv[2]
		include_key = true

lines = [line.rstrip('\n') for line in open(file_name)]
print 'done reading file ' + file_name
out = {}
for l in lines:
		if l == ",,": continue
		a = l.replace("'", " ").replace("\"", " ")
		a = a.split(",")
		if a[0] in out:
			if include_key:
				out[a[0]] += " ".join(a[1:])
			else:
				out[a[0]] += " ".join(a[2:])
		else:
			if include_key:
				out[a[0]] = " ".join(a[1:])
			else:
				out[a[0]] = " ".join(a[2:])

print 'done flattening attributes'
outf = open(file_name+'.out', 'w')
for i in out.keys():
	outf.write(i + "," + out[i] + "\n")
outf.close()
print 'done writing to outfile ' + file_name+'.out'