anki-decks/fromtext.py

72 lines
2 KiB
Python
Executable file

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
import os
import string
import sys
import pathlib
def _sanitize_line(l: str) -> str:
if '\n' in l and l.startswith('"') and l.endswith('"'):
return l[1:-1].replace('""', '"')
return l
def _get_path(basepath: pathlib.Path, deckname: str) -> pathlib.Path:
for p in basepath.iterdir():
if p.name == deckname:
return p
if p.name.startswith(f'{deckname}.'):
return p
p = basepath / deckname
p.mkdir()
return p
def process_line(
*,
columns: list[str],
columns_indices: list[int],
deck_index: int,
target_dir: pathlib.Path,
) -> None:
columns = [_sanitize_line(l) for l in columns]
deckpath = _get_path(target_dir, columns[deck_index])
columns = [columns[i] for i in columns_indices]
slug = ''.join(c for c in columns[0].lower() if c in string.ascii_lowercase + string.digits + ' ')
slug = slug.replace(' ', '-')
base_filename = deckpath / slug
for i, c in enumerate(columns):
with (base_filename.parent / (base_filename.name + f'.{i}.html')).open('w') as fp:
fp.write(c)
if not c.endswith('\n'):
fp.write('\n')
def main(
*,
target_column_count: int = 6,
columns_indices: list[int] = [3, 4],
deck_index: int = 2,
sep: str = '\t',
):
target_dir = pathlib.Path(next(iter(sys.argv[1:] or ['.'])))
buff = ''
process_header = True
for l in sys.stdin.readlines():
if process_header and l.startswith('#'):
continue
else:
process_header = False
buff += l
split_ = buff.split(sep)
if len(split_) == target_column_count:
process_line(
columns=split_,
columns_indices=columns_indices,
deck_index=deck_index,
target_dir=target_dir,
)
buff = ''
if __name__ == '__main__':
main()