PushingScores/generate_links.py

import sys, os
import json
import re

with open('wordlist.json', 'r') as f:
    wordlist_dict = json.load(f)
#goes through every single file ending in html
#
path = "static/files/"
for path, subdirs, files in os.walk(path):
    for name in files:
        if name.endswith('html'):
            file = os.path.join(path, name)
            with open(file, 'r+', encoding="utf-8") as f:
                textfile = f.read()
                for word in wordlist_dict:
                    word = re.escape(word)
                    textfile = re.sub(r"(?<!<)(?<!</)(?<!ge\?)\b(%s)\b" %word, r"<a href='/diverge?search=\1'>\1</a>", textfile)
                f.truncate(0)
                f.write(textfile)
                f.truncate()

# print(textfile)