import codecs
from collections import defaultdict

class Corpus:

    def __init__(self, filename=None):
        self.freqs = defaultdict(int)
        if filename == None:
            self.name = 'None'
            self.read = True
        else:
            self.name = filename
            self.read = False
                        
    def __read_data(self):
        with codecs.open(self.name,encoding='utf-8') as f:
            for l in f:
                (w,i) = l[:-1].split('\t')
                self.freqs[w] = int(i)

    def add_frequency_to_table(self,words):
        if not self.read:
            self.__read_data()
            self.read = True
        # Words that start with '*' will not be counted
        # Variants of a word may be separated by ','. Each variant will
        # be counted.
        words = sum([w.split(',') for w in words if not w.endswith('*')],[])
        return [(w,self.freqs[w]) for w in words]