Created
April 17, 2021 14:34
-
-
Save blopa/a6e49c798cf0fd5c875025c6ae282064 to your computer and use it in GitHub Desktop.
jekyll-markdown-script.py
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 bs4 import BeautifulSoup | |
| import unidecode | |
| import pprint | |
| import json | |
| import yaml | |
| import os | |
| PHP_ARRAYS = "<?php $redirect_map = [" | |
| TAGS = [] | |
| CATEGORIES = [] | |
| def main(): | |
| languages = ["br", "en"] | |
| add_php_redirect_array = True | |
| create_tags_json = True | |
| create_categories_json = True | |
| # create_posts("2008-09-29-iphone.md", False, True, False, ["en", "br"]) | |
| files = os.listdir("_posts") | |
| for f in files: | |
| print("Processing file: " + f) | |
| create_posts(f, False, True, add_php_redirect_array, languages) | |
| if add_php_redirect_array: | |
| global PHP_ARRAYS | |
| PHP_ARRAYS += "];"; | |
| path = "arrays.php" | |
| php_file = open(path, mode="w+", encoding="utf-8") | |
| php_file.write(PHP_ARRAYS) | |
| print("Saving PHP file: " + path) | |
| php_file.close() | |
| if create_tags_json: | |
| global TAGS | |
| local_tags = list(dict.fromkeys(TAGS)) | |
| path = "tags.json" | |
| php_file = open(path, mode="w+", encoding="utf-8") | |
| php_file.write(json.dumps(local_tags)) | |
| print("Saving TAGS file: " + path) | |
| php_file.close() | |
| if create_categories_json: | |
| global CATEGORIES | |
| local_tags = list(dict.fromkeys(CATEGORIES)) | |
| path = "categories.json" | |
| php_file = open(path, mode="w+", encoding="utf-8") | |
| php_file.write(json.dumps(local_tags)) | |
| print("Saving CATEGORIES file: " + path) | |
| php_file.close() | |
| def get_proper_category_name(category): | |
| if category == "geral": | |
| return "general" | |
| if category == "celulares": | |
| return "cellphones" | |
| if category == "engraçado": | |
| return "funny" | |
| if category == "hardware": | |
| return "hardware" | |
| if category == "dicas": | |
| return "tips" | |
| if category == "games": | |
| return "games" | |
| if category == "sistemas": | |
| return "software" | |
| if category == "filmes": | |
| return "movies" | |
| if category == "música": | |
| return "music" | |
| if category == "notícias": | |
| return "news" | |
| if category == "nada": | |
| return "general" | |
| if category == "programação": | |
| return "coding" | |
| if category == "bandas": | |
| return "music" | |
| if category == "shows": | |
| return "music" | |
| if category == "review": | |
| return "review" | |
| if category == "unboxing": | |
| return "unboxing" | |
| if category == "eventos": | |
| return "events" | |
| if category == "software": | |
| return "software" | |
| if category == "viagens": | |
| return "trips" | |
| return category | |
| def create_posts(filename, use_categories, check_translation, add_php_redirect_array, languages): | |
| #pp = pprint.PrettyPrinter(indent=4) | |
| file = open("_posts/" + filename, mode='r', encoding='UTF8') | |
| read = file.read().split("---") | |
| content = read | |
| yml = content[1] | |
| try: | |
| yml = yaml.safe_load(yml) | |
| #print(yml["categories"]) | |
| except yaml.YAMLError as exc: | |
| print(exc) | |
| new_content = "---\n" | |
| titles_translations = [] | |
| posts_translations = [] | |
| if "id" in yml: | |
| new_content += "id: " + str(yml["id"]) + "\n" | |
| #if "layout" in yml: | |
| #new_content += "layout: " + yml["layout"] + "\n" | |
| global CATEGORIES | |
| if "category" in yml: | |
| category_name = get_proper_category_name(yml["category"].lower()) | |
| CATEGORIES.append(category_name) | |
| new_content += "category: [" + category_name + "]\n" | |
| if "author" in yml: | |
| new_content += "author: " + yml["author"].replace("blopa", "Pablo Montenegro") + "\n" | |
| if "title" in yml: | |
| if check_translation: | |
| title = yml["title"].strip().strip("'").strip('"') | |
| if "<!--:-->" not in title: | |
| new_title = "" | |
| for l in languages: | |
| new_title += "<!--:" + l + "-->" + title + "<!--:-->" | |
| title = new_title | |
| titles = list(filter(None, title.split("<!--:-->"))) | |
| #titles = title.split("<!--:-->") | |
| titles_translations = titles | |
| else: | |
| new_content += "title: " + yml["title"] + "\n" | |
| if "categories" in yml: | |
| if use_categories: | |
| new_content += "categories: [" | |
| for cat in yml["categories"]: | |
| category_name = get_proper_category_name(cat.lower()) | |
| CATEGORIES.append(category_name) | |
| new_content += category_name + ", " | |
| else: | |
| if "category" not in yml: | |
| category_name = get_proper_category_name(yml["categories"][0].lower()) | |
| CATEGORIES.append(category_name) | |
| new_content += "categories: [" + category_name | |
| if new_content.endswith(", "): | |
| new_content = new_content[:-2] | |
| new_content += "]\n" | |
| if "tags" in yml: | |
| global TAGS | |
| new_content += "tags: [" | |
| for tag in yml["tags"]: | |
| if tag is not "Add new tag" and tag is not "Adicionar nova tag": | |
| tag_name = tag.replace("'", "").lower() | |
| TAGS.append(tag_name) | |
| new_content += "'" + tag_name + "', " | |
| if not use_categories and "categories" in yml: | |
| # add categories as tags | |
| for cat in yml["categories"]: | |
| category_name = get_proper_category_name(cat.lower()) | |
| CATEGORIES.append(category_name) | |
| new_content += "'" + category_name + "', " | |
| if new_content.endswith(", "): | |
| new_content = new_content[:-2] | |
| new_content += "]\n" | |
| else: | |
| TAGS.append("geral") | |
| new_content += "tags: ['geral']\n" | |
| if "categories" not in yml: | |
| CATEGORIES.append(get_proper_category_name("geral")) | |
| new_content += "categories: ['" + get_proper_category_name("geral") + "']\n" | |
| if "category" not in yml and "categories" not in yml: | |
| CATEGORIES.append(get_proper_category_name("geral")) | |
| new_content += "categories: ['" + get_proper_category_name("geral") + "']\n" | |
| post = content[2] | |
| if post is not None: | |
| post = post.strip().strip("'").strip('"') | |
| if "<!--:-->" not in post: | |
| new_post = "" | |
| for l in languages: | |
| new_post += "<!--:" + l + "-->" + post + "<!--:-->" | |
| post = new_post | |
| posts = list(filter(None, post.split("<!--:-->"))) | |
| #titles = title.split("<!--:-->") | |
| posts_translations = posts | |
| files = {} | |
| filenames = {} | |
| # print(new_content) | |
| for lang in languages: | |
| lang_key = lang | |
| if lang == "br": | |
| lang_key = "pt-br" | |
| #if titles_translations.startswith("<!--:" + lang + "-->"): | |
| for tit in titles_translations: | |
| if tit.startswith("<!--:" + lang + "-->"): | |
| if lang_key not in files: | |
| files[lang_key] = new_content | |
| san_title = tit.replace("<!--:" + lang + "-->", "").replace("'", "`") | |
| filenames[lang_key] = filename[:11] + unidecode.unidecode(san_title.lower().replace("'", "").replace("’", "").replace("`", "").replace("…", "").replace('"', "").replace("-", "").replace(" ", "-").replace("?", "").replace(":", "").replace(",", "").replace(";", "").replace("!", "").replace(".", "-").replace("\\", "").replace("/", "").replace("$", "").replace("’", "").replace("”", "").replace("…", "").replace("–", "").replace("&", "").replace("‘", "").replace("“", "")).replace("--", "-").replace("]", "-").replace("[", "-").replace(")", "-").replace("(", "-") + "." + lang_key +".md" | |
| file_ending = "." + lang_key + ".md" | |
| filenames[lang_key] = filenames[lang_key].replace("--", "-").replace("--", "-").replace("--", "-").replace("--", "-").replace("-" + file_ending, file_ending) | |
| clean_filename = filenames[lang_key][11:].replace(file_ending, "") | |
| if "id" in yml and add_php_redirect_array: | |
| # TODO redo links without category or not | |
| date = filename[:11].replace("-", "/") | |
| if "category" in yml: | |
| cat = yml["category"] | |
| elif "categories" in yml: | |
| if use_categories: | |
| dunno = 1 # TODO | |
| else: | |
| cat = unidecode.unidecode(yml["categories"][0]).lower() | |
| global PHP_ARRAYS | |
| PHP_ARRAYS += str(yml["id"]) + " => 'https://pablomontenegro.com.br/" + lang_key + "/blog/" + cat + "/" + date + clean_filename + "', " | |
| # print("This is the raw_filename: " + filenames[lang_key]) | |
| # print("This is the clean_filename: " + clean_filename) | |
| files[lang_key] += "path: " + clean_filename + "\n" | |
| # files[lang_key] += "cover: " + date[:7] + "\n" | |
| files[lang_key] += "cover: ../../../src/images/default-cover.png\n" | |
| files[lang_key] += "date: " + date[:10].replace("/", "-") + "\n" #yyy-mm-dd | |
| clean_title = san_title.replace("’", "’").replace("”", '"').replace("…", "...").replace("–", "-").replace("&", "&").replace("‘", "`").replace("“", '"') | |
| files[lang_key] += "excerpt: '" + clean_title + "'\n" | |
| files[lang_key] += "title: '" + clean_title + "'\n---\n" | |
| # print(files[lang_key]) | |
| for p in posts_translations: | |
| con = p.strip() | |
| # print("Ve se comeca com <!--:" + lang + "--> : ") | |
| # print(con) | |
| if con.startswith("<!--:" + lang + "-->"): | |
| # print("Entrei na linguagem " + lang_key) | |
| if lang_key not in files: | |
| files[lang_key] = new_content | |
| san_con = con.replace("<!--:" + lang + "-->", "") | |
| san_con = BeautifulSoup(san_con).prettify() | |
| files[lang_key] += san_con + "\n" | |
| # print(files[lang_key]) | |
| #print(new_content) | |
| # print(titles_translations) | |
| # print(posts_translations) | |
| #print(files["br"]) | |
| # print(files["en"]) | |
| for lang_key in files.keys(): | |
| fullpath = "posts/" + filenames[lang_key] | |
| final_file = open(fullpath, mode="w+", encoding="utf-8") | |
| final_file.write(files[lang_key]) | |
| print("Saving file: " + fullpath) | |
| final_file.close() | |
| if __name__ == '__main__': | |
| main() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment