Script that performs mail merge using pandoc, jinja2, and pypdf2.

pandoc-merge.py

#!/usr/bin/env python3
# /// script
# requires-python = ">=3.12"
# dependencies = [
#     "jinja2",
#     "pypdf2",
#     "tqdm",
# ]
# ///
import argparse
import csv
import logging
import subprocess
import sys
import tempfile
from pathlib import Path

from jinja2 import Template
from PyPDF2 import PdfMerger
from tqdm import tqdm


def merge(input_file, csv_file, output_file, extra_args):
    template = Template(input_file.read_text())
    tempdir = tempfile.TemporaryDirectory(prefix="pandoc_merge_", dir=input_file.parent)
    tempdir_path = Path(tempdir.name)
    logging.debug(f"Using temporary directory {tempdir_path}.")
    output_files = []
    with csv_file.open() as csvfile:
        # Make list to get proper progress bar
        rows = list(enumerate(csv.DictReader(csvfile)))
        for i, data in tqdm(rows):
            content = template.render(**data)
            tmpname = f"temp{i:05d}"
            tmppath = (tempdir_path / tmpname).with_suffix(output_file.suffix)
            output_files.append(tmppath)
            cmd = ["pandoc", "-o", tmppath] + extra_args
            logging.debug(f"Process file {tmppath.name}")
            subprocess.run(cmd, input=content, text=True, check=True)
    merger = PdfMerger()
    for pdf in output_files:
        merger.append(pdf)
    with output_file.open("wb") as outfile:
        merger.write(outfile)
    tempdir.cleanup()


def main():
    # Parse commandline arguments
    arg_parser = argparse.ArgumentParser()
    arg_parser.add_argument("-v", "--verbose", action="store_true")
    arg_parser.add_argument("-c", "--csv", type=Path)
    arg_parser.add_argument("-o", "--outfile", type=Path)
    arg_parser.add_argument("input", type=Path)
    args, extra_args = arg_parser.parse_known_args()
    # Set up logging
    if args.verbose:
        level = logging.DEBUG
    else:
        level = logging.ERROR
    logging.basicConfig(level=level)
    # Return exit value
    merge(args.input, args.csv, args.outfile, extra_args)
    return 0


if __name__ == "__main__":
    sys.exit(main())