Script to add a `created` frontmatter property to any markdown file missing one

add_created_property.py

"""This script processes markdown files in the current directory and all subdirectories,
adding a `created` property to the frontmatter if it does not already exist.

Process:
1. Recursively identify markdown files (*.md) in the current folder and all subfolders.
2. Skip any directories specified in the `DIRS_TO_SKIP` tuple.
3. For each markdown file:
    a. Check if the frontmatter has a 'created' property.
    b. If 'created' exists, move on to the next file.
    c. If 'created' does not exist:
        i. If the filename contains a date in the format '%Y%m%d' or '%Y-%m-%d', set the 'created' property to that value formatted as '%Y-%m-%d'.
        ii. If the file's 'Created' metadata is within the past 24 hours, assume it's an error and use the file's 'Modified' metadata attribute to populate the 'created' property, formatted as 'YYYY-MM-DD HH:MM:SSZ'.
        iii. Otherwise, use the file's 'Created' metadata attribute to populate the 'created' property, formatted as 'YYYY-MM-DD HH:MM:SSZ'.

Usage:
1. Install frontmatter if not already present: `pip install python-frontmatter`.
2. Place the script in the root directory where your markdown files are located.
3. Run the script using `python add_created_property.py`.
"""

import os
import frontmatter
from datetime import datetime, timedelta
import re

# Directories to skip
DIRS_TO_SKIP = ("_templates",)


def get_file_created_time(file_path):
    """Get the creation time of a file and format it to the desired format.

    Args:
        file_path (str): The path to the file.

    Returns:
        str: The formatted creation time.
    """
    created_time = os.path.getctime(file_path)
    return datetime.utcfromtimestamp(created_time).strftime("%Y-%m-%d %H:%M:%SZ")


def get_file_modified_time(file_path):
    """Get the modified time of a file and format it to the desired format.

    Args:
        file_path (str): The path to the file.

    Returns:
        str: The formatted modified time.
    """
    modified_time = os.path.getmtime(file_path)
    return datetime.utcfromtimestamp(modified_time).strftime("%Y-%m-%d %H:%M:%SZ")


def extract_date_from_filename(filename):
    """Extract a date from the filename if it matches the patterns '%Y%m%d' or '%Y-%m-%d'.

    Args:
        filename (str): The name of the file.

    Returns:
        str: The extracted and formatted date, or None if no date is found.
    """
    date_patterns = [r"\d{8}", r"\d{4}-\d{2}-\d{2}"]
    for pattern in date_patterns:
        match = re.search(pattern, filename)
        if match:
            date_str = match.group()
            try:
                if len(date_str) == 8:
                    return datetime.strptime(date_str, "%Y%m%d").strftime("%Y-%m-%d")
                elif len(date_str) == 10:
                    return datetime.strptime(date_str, "%Y-%m-%d").strftime("%Y-%m-%d")
            except ValueError:
                continue
    return None


def process_markdown_files(directory):
    """Recursively process markdown files in the given directory to check and add 'created' property in frontmatter.

    Args:
        directory (str): The root directory to start the search.
    """
    for root, dirs, files in os.walk(directory):
        # Skip directories specified in DIRS_TO_SKIP
        dirs[:] = [d for d in dirs if d not in DIRS_TO_SKIP]

        for file in files:
            if file.endswith(".md"):
                file_path = os.path.join(root, file)
                try:
                    with open(file_path, "r", encoding="utf-8") as f:
                        post = frontmatter.load(f)
                except Exception as e:
                    print(f"Error loading frontmatter from {file_path}: {e}")
                    continue

                if "created" not in post.metadata:
                    filename_date = extract_date_from_filename(file)
                    if filename_date:
                        post.metadata["created"] = filename_date
                    else:
                        created_time = datetime.utcfromtimestamp(os.path.getctime(file_path))
                        if created_time >= datetime.utcnow() - timedelta(days=1):
                            modified_time = get_file_modified_time(file_path)
                            post.metadata["created"] = modified_time
                        else:
                            created_time_str = created_time.strftime("%Y-%m-%d %H:%M:%SZ")
                            post.metadata["created"] = created_time_str

                    post_content = frontmatter.dumps(post)
                    try:
                        with open(file_path, "w", encoding="utf-8") as f:
                            f.write(post_content)
                        print(f"Added 'created' to {file_path}")
                    except Exception as e:
                        print(f"Error writing to {file_path}: {e}")
                else:
                    print(f"'created' already exists in {file_path}")


if __name__ == "__main__":
    current_directory = os.getcwd()
    process_markdown_files(current_directory)