round 2

gistfile1.txt

from string import ascii_lowercase
from collections import defaultdict
from pprint import pprint
 
def add_word(matrix, word):
    '''
    this modifies a letter count matrix to add in the pairs
    the matrix stores things as follows:
    matrix[first][second] = count of times first+second appear in word

    Note: consider if you want to have somethign that recognizes some letters
    are more common at the end of things
 
    Also note this does not work with one letter words, so consider how you want
    to handle
    '''
    word = word.lower()
    for a,b in zip(word,word[1:]):
        matrix[a][b] = matrix[a].get(b,0) + 1

def letter_probs(m):
    '''
    takes a part of the letter count matrix to build the stochastic matrix for
    a particular first letter.  I.e.:
    letter_probs(matrix['a']) will return a dictionary of k:v where:
        k = second letter
        v = percentage of times second letter appears after first
    
    Note: sum of all vs = 1
    '''
    total = sum([c for c in m.values()])
    result = {l:p/total for l,p in m.items()}
    return result

def build_stochastic(matrix):
    '''
    takes a letter count matrix and returns a new dictionary of matrixs
    each key in new matrix is a letter
    the value = a new stochastic matrix of following letters: probability
    '''

    result = {l:letter_probs(matrix[l]) for l in matrix}
    return result

matrix = defaultdict(dict)
add_word(matrix, 'hello')
add_word(matrix, 'jello')
add_word(matrix, 'ham')
add_word(matrix, 'dam')
add_word(matrix, 'dan')
add_word(matrix, 'dap')
#print(letter_probs(matrix['a']))
print(build_stochastic(matrix))