small datasets about markovify HOT 4 CLOSED

I am not a developer of the library, but AFAIK you should do something like the following:

markovify.NewlineText(text, state_size=1)

By default state_size is 2.

from markovify.

What is the length of the chain you use?

from markovify.

Heres my code snippet

import markovify

text = "I like you\nyou like pie"
text_model = markovify.NewlineText(text)

for i in range(10):
    sentence = text_model.make_sentence(test_output=False)
    print(sentence)
    if sencence in text.split('\n'):
        print('not new sentence')

print(text_model.chain.to_json())

Here's the JSON printed at the end, which I assume are the production rules.

 [
    [["___BEGIN__", "___BEGIN__"], {"I": 1, "you": 1}], 
    [["___BEGIN__", "I"], {"like": 1}], 
    [["I", "like"], {"you": 1}], 
    [["like", "you"], {"___END__": 1}], 
    [["___BEGIN__", "you"], {"like": 1}], 
    [["you", "like"], {"pie": 1}], 
    [["like", "pie"], {"___END__": 1}]
] 

It appears that the rules use word pairs as keys, e.g. the 3rd rule maps "I like" -> "you".
How do I change my model so that the production rules depend only on the previous word, not the two previous words. e.g. "like" -> "you" ?

from markovify.

Exactly what I needed, thank you :)

from markovify.