Invert Pinboard HTML Backup for import into Linkwarden

invert_pinboard_export.py

from bs4 import BeautifulSoup

# Run: 
#    $ pip install beautifulsoup4
# For more info: https://github.com/linkwarden/linkwarden/issues/555

INPUT_FILE = "pinboard_export.html"


def reverse_bookmarks():
    # Read the input file
    with open(INPUT_FILE, "r", encoding="utf-8") as f:
        content = f.read()

    # Parse the HTML using BeautifulSoup
    soup = BeautifulSoup(content, "html.parser")

    # Find the main DL element
    main_dl = soup.find("dl")

    # Get all direct DT elements
    bookmarks = []

    # First preserve the p tag if it exists
    p_tag = None
    for child in main_dl.children:
        if child.name == 'p':
            p_tag = child
            break

    # Collect all DT and their following content until the next DT
    current_dt = None
    current_content = []
    for child in main_dl.children:
        if isinstance(child, str) and child.strip() == '':
            continue  # Skip empty text nodes

        if child.name == 'dt':
            # Store previous group if exists
            if current_dt is not None:
                bookmarks.append((current_dt, current_content))
            # Start new group
            current_dt = child
            current_content = []
        elif current_dt is not None:
            # Collect all content (tags or text) that follows the current DT
            current_content.append(child)

    # Don't forget the last group
    if current_dt is not None:
        bookmarks.append((current_dt, current_content))

    # Clear the DL element
    main_dl.clear()

    # Add back the p tag if it existed
    if p_tag:
        main_dl.append(p_tag)

    # Add the bookmarks in reverse order
    for dt, content_list in reversed(bookmarks):
        main_dl.append(dt)
        for content in content_list:
            main_dl.append(content)
    # Write the modified HTML to a new file
    output_filename = "reversed_" + INPUT_FILE
    with open(output_filename, "w", encoding="utf-8") as f:
        f.write(str(soup))


if __name__ == "__main__":
    reverse_bookmarks()