#!/usr/bin/python
# Fuzzy license searcher
# Author: Colin Walters <walters@redhat.com>
# This file is hereby placed into the public domain.

import os,sys,re
import difflib

TABLE_URL = 'http://fedoraproject.org/wiki/Licensing?action=raw'

licenses = {
'GPLv2+': {'fragments': ['GNU General Public License'],
           'seqs': ['under the terms of the GNU General Public License', 'version 2 of the License', 'or (at your option) any later version']},
'ASL 2.0': {'seqs': ['under the Apache License, Version 2.0']},
}
for license,matchdata in licenses.iteritems():
  matchdata['seqs'] = map(lambda seq: difflib.SequenceMatcher(None, None, seq), matchdata['seqs'])

def file_licenses(fname, threshold=0.7):
  for license,matchdata in licenses.iteritems():
    seqs = matchdata['seqs']
    seq_hits = map(lambda x: 0, seqs)
    for i,line in enumerate(open(fname)):
      for seqi,seq in enumerate(seqs):
        seq.set_seq1(line)
        ratio = seq.quick_ratio()
        if ratio > seq_hits[seqi]:
          ratio = seq.ratio()
          if ratio > seq_hits[seqi]:
            seq_hits[seqi] = ratio
    avg_ratio = reduce(lambda x,y: x+y, seq_hits)/len(seq_hits)
    if avg_ratio >= threshold:
      yield (license, avg_ratio)

def dirtree_licenses(dirname):
  for dirname,dirs,files in os.walk(dirname):
    for fname in files:
      fpath = os.path.join(dirname,fname)
      for license,ratio in file_licenses(fpath):
        yield (fpath, license, ratio)
      
if __name__ == '__main__':
  matches = {}
  for fpath,license,ratio in dirtree_licenses(sys.argv[1]):
    if license not in matches or (matches[license][0] < ratio):
      matches[license] = (ratio, fpath)
      sys.stdout.write('%s %s: %s\n' % (license, ratio, fpath))