The corpus I am using is just one I found online. The corpus you choose is central to generating realistic text.
# Get raw text as string with open("brown.txt") as f: text = f.read()
Build Markov Chain
# Build the model. text_model = markovify.Text(text)
Generate One Tweet
# Print three randomly-generated sentences of no more than 140 characters for i in range(3): print(text_model.make_short_sentence(140))
Within a month, calls were still productive and most devotees of baseball attended the dozens of them. Even death, therefore, has a leather bolo drawn through a local rajah in 1949. He had a rather sharp and confident.