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



sourcetext = 'data/faq-fr.txt'
targettext = 'data/faq-en.txt'

sourcetext = 'data/gainsbourg-en.txt'
targettext = 'data/gainsbourg-fr.txt'

sourcetext = "/home/pat/l/pt/doingbusinessinbrazil/brazilbiz-en.txt"
targettext = "/home/pat/l/pt/doingbusinessinbrazil/brazilbiz-pt.txt"

sourcetext = 'data/mis/en.txt'
targettext = 'data/mis/fr.txt'

sourcetext = 'data/notredame/en.txt'
targettext = 'data/notredame/fr.txt'

sourcetext = '/home/pat/l/udhr/data/udhr_eng.txt'
targettext = '/home/pat/l/udhr/data/udhr_por.txt'

sourcetext = 'data/x.txt'
targettext = 'data/y.txt'


source = align.UnicodeDammit(open(sourcetext, 'U').read()).unicode
target = align.UnicodeDammit(open(targettext, 'U').read()).unicode

sourcetokens = source.split()
for w in sourcetokens: print w
targettokens = target.split()
sourcewords = set(sourcetokens)
targetwords = set(targettokens)

sourcedistributions = align.Collection()
targetdistributions = align.Collection()

for i,w in enumerate(sourcewords):
    sourcedistributions[w] = i
    print i, w, sourcedistributions[w]

for i,w in enumerate(targetwords):
    targetdistributions[w] = i

def normalize(seq): 
    print seq
    maximum = max(seq)
    try: return [ int(float(n)/maximum * 100) for n in seq]
    except ZeroDivisionError: 
        raise  "doh! maximum is: %d" % maximum

def hitfreq(word, distributions): 
 return align.freq(normalize(distributions[word]))

def compare(a,b):
 sourcewordvec = hitfreq(a, targetdistributions)
 targetwordvec = hitfreq(b, sourcedistributions)
 answer = align.sim(sourcewordvec, targetwordvec)
 if answer > 0: return answer

def findsims(sourceword):
 sims = []
 for targetword in list(targetwords):
  res = (compare(targetword, sourceword), sourceword, targetword)
  sims.append(res)
 return sorted(sims)

def top(n, dictionary):
 vals = set(dictionary.values())[:100]
 tops = sorted(list(vals))[-n:]
 return [k for k in dictionary if dictionary[k] in tops]
 

if __name__ == "__main__":
 import sys
 query = unicode(sys.argv[1])
 biz = findsims(query)
 for x, e, p  in biz[-10:]: 
  print "» %f\t%s\t%s" % (x, e, p)
 print u"out of " + unicode(len(sourcedistributions[query]))