Cut Tags Stats

cut_tags_stats.py

#!/usr/bin/env python
# -*- coding: utf-8 -*- 

import operator, sys

def split_tags(string):
    return [x for x in string.replace('{', '').split('}') if x]
        
class CutTagsStats:
    
    def __init__(self, filename):
        self.filename = filename
        self.data = {}
    
    def handle_line(self, tags, count):
        for t in tags:
            if t in self.data:
                self.data[t] += count
            else:
                self.data[t] = count
    
    def read_csv_file(self):
        fp = open(self.filename, "r")
        for line in fp:
            if len(line) == 1:
                break
                    
            line = line.split(',')
            
            if len(line) > 2: # old format like "{business},{3g},{arcep},{femtocell},{free},{free-mobile},{mobile}",4
                continue
            
            tags = split_tags(line[0])
            count = int(line[1])
            
            self.handle_line(tags, count)
        
    def sort_print(self):
        data = sorted(self.data.items(), key=operator.itemgetter(1), reverse=True)
        for line in data:
            print "%s,%d" % line

if __name__ == '__main__':
    if len(sys.argv) < 2:
        print "python cut_tags_stats.py file.csv"
    else:
        cut =  CutTagsStats(sys.argv[1])
        cut.read_csv_file()
        cut.sort_print()