Update Mathpix Notes archive online images to local (including updating links in the file).

update-mathpix-notes-archive-online-images-to-local.py

import os
import re
import requests
from urllib.parse import urlparse


def download_file(url, folder):
    parsed_url = urlparse(url)
    filename = os.path.basename(parsed_url.path)
    base, ext = os.path.splitext(filename)
    local_path = os.path.join(folder, filename)
    i = 1
    while os.path.exists(local_path):
        filename = f"{base}_{i}{ext}"
        local_path = os.path.join(folder, filename)
        i += 1

    try:
        resp = requests.get(url, stream=True, timeout=15)
        resp.raise_for_status()
        with open(local_path, "wb") as f:
            for chunk in resp.iter_content(1024):
                f.write(chunk)
        print(f"Downloaded: {url} -> {local_path}")
        return local_path
    except Exception as e:
        print(f"Failed to download {url}: {e}")
        return None


def update_markdown(md_path, assets_dir):
    with open(md_path, "r", encoding="utf-8") as file:
        content = file.read()

    pattern_md = re.compile(r"!\[.*?\]\((https?://[^\s)]+)\)")
    pattern_html = re.compile(r'<img\s+[^>]*src="(https?://[^"]+)"')

    urls = set(pattern_md.findall(content) + pattern_html.findall(content))
    if not urls:
        print(f"  No remote images found in {md_path}")
        return

    url_to_local = {}

    for url in urls:
        local_file = download_file(url, assets_dir)
        if local_file:
            rel_path = "assets/" + os.path.basename(local_file)
            url_to_local[url] = rel_path

    for url, rel_path in url_to_local.items():
        content = content.replace(url, rel_path)
        print(f"  Replaced {url} -> {rel_path}")

    with open(md_path, "w", encoding="utf-8") as file:
        file.write(content)


def process_all_markdowns(notes_dir="notes"):
    assets_dir = os.path.join(notes_dir, "assets")
    if not os.path.exists(assets_dir):
        os.makedirs(assets_dir, exist_ok=True)

    any_found = False
    for fname in os.listdir(notes_dir):
        md_path = os.path.join(notes_dir, fname)
        if os.path.isfile(md_path) and fname.endswith(".mmd"):
            print(f"Processing {md_path} ...")
            update_markdown(md_path, assets_dir)
            any_found = True
    if not any_found:
        print(f"No .mmd files found in {notes_dir}!")
    print("全部处理完成！")


if __name__ == "__main__":
    import sys

    # 支持命令行参数指定目录
    if len(sys.argv) > 1:
        process_all_markdowns(sys.argv[1])
    else:
        process_all_markdowns()