About
marc (markov chain) is a small, but flexible Markov chain generator.
Usage
marc is easy to use. To build a MarkovChain pass the object a sequence of items:
from marc import MarkovChain
sequence = [
'Rock', 'Rock', 'Rock', 'Paper', 'Rock', 'Scissors',
'Paper', 'Paper', 'Scissors', 'Rock', 'Scissors',
'Scissors', 'Paper', 'Scissors', 'Rock', 'Rock', 'Rock',
'Paper', 'Scissors', 'Scissors', 'Scissors', 'Rock'
]
chain = MarkovChain(sequence)
The learned transition matrix can be accessed through the matrix attribute:
print(chain.matrix)
# [[0.5, 0.25, 0.25], [0.2, 0.2, 0.6], [0.375, 0.25, 0.375]]
Though, the output is perhaps better viewed as a pandas DataFrame:
import pandas as pd
df = pd.DataFrame(
chain.matrix,
index=chain.encoder.index_,
columns=chain.encoder.index_
)
print(df)
# Rock Paper Scissors
# Rock 0.500 0.25 0.250
# Paper 0.200 0.20 0.600
# Scissors 0.375 0.25 0.375
Use the next method to generate the next state (seeded or unseeded):
chain.next('Rock')
# 'Rock'
chain.next()
# Paper
The next method can also generate multiple states with the n argument:
chain.next('Paper', n=5)
# ['Scissors', 'Paper', 'Rock', 'Paper', 'Scissors']
MarkovChain objects are iterable. This means that they can be passed directly to the next function:
next(chain)
# 'Scissors'
next(chain)
# Rock
Example
A fully worked example of marc in action (block text provided by quote):
import random
import re
from quote import quote
from marc import MarkovChain
quotes = quote('shakespeare', 250)
print(quotes[0])
# {'author': 'William Shakespeare',
# 'book': 'As You Like It',
# 'quote': 'The fool doth think he is wise, but the wise man knows himself to be a fool.'}
text = '\n'.join([q['quote'] for q in quotes])
text = text.lower()
tokens = re.findall(r"[\w']+|[.,!?;]", text)
tokens[:5]
# ['the', 'fool', 'doth', 'think', 'he']
chain = MarkovChain(tokens)
def generate_sentences(chain, n=2, length=(10, 20)):
for _ in range(n):
l = random.randint(length[0], length[1])
nonsense = ' '.join(chain.next(n=l))
print(nonsense)
generate_sentences(chain)
# and unless by some are fascinated by the hour upon the wind faithful
# those that hath had a very much as flaws go
Install
pip install -U marc
527 Dec 4, 2022
43 Dec 24, 2022
12 Dec 27, 2022
3 Jan 19, 2022
85 Dec 9, 2022
1 Nov 11, 2021
4 Jul 18, 2021
1k Dec 16, 2022
60 Nov 16, 2022
0 Sep 5, 2021
6k Jan 6, 2023
106 Nov 6, 2022
![[CVPR 2021] 'Searching by Generating: Flexible and Efficient One-Shot NAS with Architecture Generator'](https://github.com/eric8607242/SGNAS/raw/master/resource/SGNAS_framework.png)
35 Dec 1, 2022
40 Dec 1, 2022
2 Jul 23, 2022
1 Jan 3, 2022
5 Mar 09, 2022
5 Aug 01, 2022
48.4k Jan 09, 2023
1 Nov 11, 2021
6 Sep 29, 2022
174 Dec 31, 2022
8 Jul 29, 2022
431 Jan 04, 2023
2 Nov 16, 2022
595 Dec 27, 2022
1.6k Jan 04, 2023
8 Sep 10, 2022
5 Nov 19, 2021
13.2k Jan 02, 2023
13 Mar 31, 2022
62 Oct 10, 2022
111 Nov 27, 2022
65.7k Jan 03, 2023
76 Dec 06, 2022
235 Dec 18, 2022