Created
September 7, 2023 02:09
-
-
Save k5trismegistus/064bb6e973d6a618a1a543c265794499 to your computer and use it in GitHub Desktop.
A script for splitting blog data exported in MT format into individual text files for training with llama-index by article.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
from dateutil.parser import parse | |
from bs4 import BeautifulSoup | |
BLOG_TITLE = <ブログ名> | |
EXPORTED = <エクスポートファイル名> | |
class MtArticle(): | |
def __init__(self): | |
self.title = None | |
self.date = None | |
self.body = None | |
self.published = False | |
def __str__(self): | |
return self.title | |
class MtParser(): | |
def __init__(self): | |
self.mode = 'header' | |
self.contents = [] | |
def parse(self, file): | |
content = MtArticle() | |
line = file.readline() | |
while line: | |
if line.startswith('--------'): | |
self.contents.append(content) | |
content = MtArticle() | |
self.mode = 'header' | |
elif line.startswith('-----'): | |
self.mode = 'select' | |
if self.mode == 'header': | |
self.__parse_header(content, line) | |
if self.mode == 'body': | |
self.__parse_body(content, line) | |
if self.mode == 'comment': | |
pass | |
if self.mode == 'select': | |
self.__select_mode(line) | |
line = file.readline() | |
def save_text(self, only_published=True): | |
to_save_contents = [c for c in self.contents if c.published] if only_published else self.contents | |
for content in to_save_contents: | |
body_soup = BeautifulSoup(content.body, 'html.parser') | |
body_text = body_soup.get_text() | |
prefix =str(content.date.year) + '-' + str(content.date.month) + '-' + str(content.date.day) | |
with open('export/' + prefix + '_' + content.title.replace('/', '') + '.txt', 'w') as f: | |
f.write(f'{BLOG_TITLE} というブログの筆者は、 「{content.title.strip()}」 という記事を書きました。\n\n') | |
f.write('---------\n\n') | |
f.write(body_text) | |
def __select_mode(self, line): | |
if line.startswith('BODY:'): | |
self.mode = 'body' | |
if line.startswith('COMMENT:'): | |
self.mode = 'comment' | |
if line.startswith('EXCERPT:'): | |
self.mode = 'excerpt' | |
def __parse_header(self, content, line): | |
if line.startswith('TITLE: '): | |
content.title = line[7:] | |
if line.startswith('STATUS: Publish'): | |
content.published = True | |
if line.startswith('DATE: '): | |
content.date = parse(line[6:]) | |
def __parse_body(self, content, line): | |
if content.body is None: | |
content.body = line | |
else: | |
content.body += line | |
if __name__ == '__main__': | |
with open(EXPORTED) as f: | |
mp = MtParser() | |
mp.parse(f) | |
mp.save_text() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment