Allow to batch-process entire directories (7fd06006) · Commits · Fachschaft / coliverter

pyproject.toml

+1 −1

Original line number	Diff line number	Diff line
		[project]
		name = "coliverter"
		version = "0.3.4"
		version = "0.4.0"
		authors = [
		{ name="Jakob Moser", email="moser@cl.uni-heidelberg.de" },
		]

src/coliverter/convert.py

+76 −20

Original line number	Diff line number	Diff line
		from itertools import repeat
		from pathlib import Path
		from typing import Annotated

		from multiprocessing import Pool
		import typer

		from coliverter.FileFormat import FileFormat
		@@ -11,18 +13,71 @@ from coliverter.steps.html_to_pdf import html_to_pdf
		from coliverter.steps.md_to_html import md_to_html


		def _get_output_format(
		output_path: Path \| None, output_format: FileFormat \| None
		) -> FileFormat:
		"""
		Given a file format that might or might not be set, and an output path that might or might not be set,
		get the output format we want to use (in the worst case, by using some default).

		You can be sure that whatever you put in here, you will get a file format out.
		"""
		if output_format:
		# Well, that was easy
		return output_format

		try:
		return FileFormat(output_path.suffix.removeprefix(".") if output_path else None)
		except ValueError:
		return FileFormat.HTML


		def _convert_single(
		markdown_path: Path \| None, output_path: Path \| None, output_format: FileFormat
		) -> None:
		markdown = read(markdown_path)
		html = md_to_html(markdown)

		if output_format == FileFormat.HTML:
		write(html, output_path)
		return

		pdf = html_to_pdf(html)
		if output_format == FileFormat.PDF:
		write(pdf, output_path)
		return


		def _convert_single_guessing_output_path(
		markdown_file_path: Path,
		markdown_dir_path: Path,
		output_dir_path: Path,
		output_format: FileFormat,
		) -> None:
		_convert_single(
		markdown_file_path,
		output_dir_path
		/ (
		markdown_file_path.relative_to(markdown_dir_path).with_suffix(
		f".{output_format}"
		)
		),
		output_format,
		)


		@app.command()
		def convert(
		markdown_path: Annotated[
		Path \| None,
		typer.Argument(
		help="Path to the input Markdown file. If None, read Markdown from STDIN."
		help="Path to the input Markdown file. If None, read Markdown from STDIN. If a directory, recursively find all *.md files and process those."
		),
		] = None,
		output_path: Annotated[
		Path \| None,
		typer.Argument(
		help="Path to the output file. If None, write output to STDOUT."
		help="Path to the output file. If None, write output to STDOUT. If markdown_path is a directory, this must be a directory as well."
		),
		] = None,
		output_format: Annotated[
		@@ -35,24 +90,25 @@ def convert(
		"""
		Convert Markdown files into other formats, using the organization identity of Fachschaft Computerlinguistik.
		"""
		if output_format is None:
		try:
		output_format = FileFormat(
		output_path.suffix.removeprefix(".") if output_path else None
		)
		except ValueError:
		output_format = FileFormat.HTML

		markdown = read(markdown_path)

		html = md_to_html(markdown)

		if output_format == FileFormat.HTML:
		write(html, output_path)
		return
		# output_format might be none, but definitely_output_format definitely isn't.
		definitely_output_format = _get_output_format(output_path, output_format)

		pdf = html_to_pdf(html)
		if markdown_path and markdown_path.is_dir():
		if not output_path or not output_path.is_dir():
		raise ValueError(
		"output_path must be a directory, given that markdown_path is one."
		)

		if output_format == FileFormat.PDF:
		write(pdf, output_path)
		return
		with Pool() as p:
		p.starmap(
		_convert_single_guessing_output_path,
		zip(
		markdown_path.rglob("*.md"),
		repeat(markdown_path),
		repeat(output_path),
		repeat(definitely_output_format),
		),
		)
		else:
		_convert_single(markdown_path, output_path, definitely_output_format)