pdf_rename_with_copyright_year

#!/usr/bin/env python3

from datetime import datetime
import os
import re
import subprocess
import sys

def extract_years_from_pdf(file_path, max_valid_year):
    """
    Extracts all valid copyright years from the content of a PDF.
    """
    try:
        # Use hex escape for the (C) symbol
        pdfgrep_command = [
            "pdfgrep", "-iP",
            r"(?:\xA9|copyright|\([cC]\))\s+.*\b\d{4}\b",
            file_path
        ]
        result = subprocess.run(
            pdfgrep_command,
            stdout=subprocess.PIPE,
            stderr=subprocess.PIPE,
            text=True,
            env={"LC_ALL": "C.UTF-8"}  # Ensure UTF-8 handling
        )

        # Debugging output
        print(f"Command: {' '.join(pdfgrep_command)}")
        print(f"stdout:\n{result.stdout}")
        print(f"stderr:\n{result.stderr}")

        # Extract valid 4-digit years from the output
        years = []
        for line in result.stdout.splitlines():
            matches = re.findall(r"\b\d{4}\b", line)
            print(f"Matched years in line: '{line}': {matches}")  # Debug matched years
            years.extend(int(year) for year in matches if 1500 <= int(year) <= max_valid_year)
        return years
    except Exception as e:
        print(f"Error processing {file_path}: {e}")
        return []

def determine_largest_valid_year(years):
    """
    Determines the largest valid year.
    """
    return max(years, default=None)

def rename_file(file_path, largest_year):
    """
    Renames the file to include the largest valid year if necessary.
    Normalizes the file extension to lowercase `.pdf`.
    """
    directory, filename = os.path.split(file_path)
    base, ext = os.path.splitext(filename)

    # Normalize the extension to lowercase
    ext = ext.lower()

    # Ensure we are dealing with a .pdf file
    if ext != ".pdf":
        print(f"Skipping non-PDF file: {filename}")
        return

    # Normalize file extension if not already lowercase
    normalized_filename = f"{base}.pdf"
    normalized_file_path = os.path.join(directory, normalized_filename)
    if file_path != normalized_file_path:
        os.rename(file_path, normalized_file_path)
        file_path = normalized_file_path
        filename = normalized_filename
        print(f"Renamed for extension normalization: {filename}")

    # Check if a valid year already exists at the end of the filename
    match = re.search(r"(\d{4})\.(medtype\.)?pdf$", filename, re.IGNORECASE)
    existing_year = int(match.group(1)) if match else None

    # Skip renaming if the existing year matches the largest valid year
    if existing_year == largest_year:
        print(f"Correct year already in file name: {filename}")
        return

    # Construct the new filename
    if ".medtype" in filename.lower():
        new_filename = re.sub(r"(\d{4}\.)?medtype\.pdf$", f"{largest_year}.medtype.pdf", filename, flags=re.IGNORECASE)
    else:
        new_filename = re.sub(r"(\d{4}\.)?pdf$", f"{largest_year}.pdf", filename, flags=re.IGNORECASE)

    # Perform the renaming
    new_file_path = os.path.join(directory, new_filename)
    if file_path != new_file_path:
        os.rename(file_path, new_file_path)
        print(f"Renamed: {filename} -> {new_filename}")

def main(file_paths):
    # Get the current year and calculate the max valid year
    current_year = datetime.now().year
    max_valid_year = current_year + 1

    for file_path in file_paths:
        # Ensure the file exists
        if not os.path.isfile(file_path):
            print(f"File not found: {file_path}")
            continue

        # Extract copyright years from the PDF content
        years = extract_years_from_pdf(file_path, max_valid_year)
        if not years:
            print(f"No copyright year found for {file_path}")
            continue

        # Determine the largest valid year
        largest_year = determine_largest_valid_year(years)
        if not largest_year:
            print(f"No valid copyright year found for {file_path}")
            continue

        # Rename the file if necessary
        rename_file(file_path, largest_year)

if __name__ == "__main__":
    if len(sys.argv) < 2:
        print("Usage: python rename_with_copyright_year.py <file1.pdf> [<file2.pdf> ...]")
        sys.exit(1)

    main(sys.argv[1:])