Mitchell and Webb Script Markov Generator

main.py

import os
import random

import markovify
import requests
from bs4 import BeautifulSoup

MAX_SEASON = 4
EPISODES_PER_SEASON = 6

SCRIPT_LENGTH = 100

BASE_URL = "https://www.springfieldspringfield.co.uk/view_episode_scripts.php?tv-show=that-mitchell-and-webb-look-2006&episode=s0{0}e0{1}"

CACHE_FILE = ".data"

full_data = ""

if os.path.isfile(CACHE_FILE):
    with open(CACHE_FILE, "r") as cached_data:
        full_data = cached_data.read()
else:
    data = []
    for season in range (1, MAX_SEASON + 1):
        for episode in range(1, EPISODES_PER_SEASON + 1):
            url = BASE_URL.format(season, episode)
            r = requests.get(url)
            html = r.text
            soup = BeautifulSoup(html, "html.parser")
            data_div = soup.find_all("div", class_="scrolling-script-container")[0]
            data.append(data_div.getText().replace(" -",".").replace(" #", "."))

    full_data = ".".join(data)
    with open(CACHE_FILE, "w") as cached_data:
        cached_data.write(full_data)


PEOPLE = [[0.25, "David Mitchell"], [0.25, "Robert Webb"], [0.15, "James Bachman"], [0.2, "Olivia Colman"], [0.15, "Paterson Joseph"]]

def select_person():
    r = random.random()
    s = 0
    for person in PEOPLE:
        s += person[0]
        if s > r:
            return person[1]

    return person[0]

text_model = markovify.Text(full_data, state_size=2)
for i in range(SCRIPT_LENGTH):
    person = select_person()
    sentence = text_model.make_sentence(tries=100)
    print("{0}: {1}".format(person, sentence))