#!/usr/bin/env python
# -*- coding: utf-8 -*-
from vector import scalar, sim
from collections import defaultdict
from glob import glob
import cPickle
import sys, codecs
#sys.stdout = codecs.getwriter('utf-8')(sys.stdout)

#LOGFILE = open('languagelog.txt', 'a')
#LOGFILE = codecs.getwriter('utf-8')(LOGFILE) 

def load(dbname="/home/pat/.pylib/udhrbigrams.db"):
  return cPickle.load(open(dbname)) 

def freq(seq):
  fq = defaultdict(int)
  for elem in seq: fq[elem] += 1
  return fq

def ngrams(s,n):
  """
  >>> ngrams('abcddlagk', 3)
  ['abc', 'bcd', 'cdd', 'ddl', 'dla', 'lag', 'agk']
  """
  return [(s[i:i+n]) for i in range(len(s)-n+1)]

def bigrams(s): return ngrams(s, 2)

def top(d, n):
  """
  return all keys of dictionary d which have one of the top
  n values of all values in d
  >>> top(list('aaabbc',1))
  'a'
  """
  #if not hasattr(d, 'values'): d = freq(d)
  topvals = sorted(set(d.values()))[-n:]
  return [e for e in d if d[e] in topvals]

def uread(fname):
  """
  read file, return unicode text, we hope
  """
  try: 
    from BeautifulSoup import UnicodeDammit
    return UnicodeDammit(open(fname, 'U').read()).unicode
  except ImportError:
    return unicode(open(fname, 'U').read())

def ethnologue():
  from LanguageCodes import languages
  codes = languages.splitlines()
  codes = [unicode(line.strip()) for line in codes]
  code2lang = {} 
  for entry in codes:
    code, country, huh, language = entry.split('\t')
    code = code.strip()
    code2lang[code] = language.strip()
  return code2lang

code2lang = ethnologue()

def modeltext(text):
  fq = freq(bigrams(text))
  for k, v in fq.items():
  	if v == 1 or ' ' in k: 
	  baleeted = fq.pop(k)
  return fq

def id(unknown, threshhold=1):
  UNKNOWN = True
  allmodels = load()
  unknownmodel = modeltext(unknown)
  sims = {}
  for code, known in allmodels.items():
    sims[code] = sim(known, unknownmodel)
  res = []
  #LOGFILE.write("==========" + '\n\n')
  #LOGFILE.write(unknown.replace('\n', '') + '\n\n')
  for score, code in sorted([(v,k) for k,v in sims.items()]): 
    if code in code2lang: lgname = code2lang[code]
    else: lgname = code
    res.append(( score, lgname ))
    if score > 0.4: 
	  #LOGFILE.write( "%.2f\t%s\n" % (score, lgname) ) 
		UNKNOWN = False
    else: 
		UNKNOWN = True
  #if UNKNOWN: LOGFILE.write( "Unknown.\n" ) 
	
  return res


def _test():
  pass


if __name__ == "__main__":
  import sys
  if len(sys.argv) < 2: 
    _test() # TODO fix this
  res = id(uread(sys.argv[1]))
  res = dict(res)
  for score, code in sorted([(k,v) for k,v in res.items()]): 
    if code in code2lang: lgname = code2lang[code]
    else: lgname = code
    #print "%.2f" % score + '\t' + lgname
    print score, 
    print lgname
  #else: 
    #for answer in id(uread(sys.argv[1]), threshhold=3):
      #print code2lang[answer], answer