Convert ODT (Open Document Text) to Markdown. Small subset of markers is supported

odt2md.py

# /// script
# requires-python = ">=3.12"
# dependencies = [
#     "isodate",
#     "odfdo",
# ]
# ///
import dataclasses
import datetime as dt
import sys
from collections.abc import Iterable
from pathlib import Path

import isodate
from odfdo import Body, Document, Element, ListItem, PageBreak, Paragraph, Span


@dataclasses.dataclass
class Entry:
    date: dt.date
    text: str


class Parser:
    def __init__(self):
        self._current_entry = ''
        self._current_date = dt.date(2009, 1, 1)  # first entry as no date
        self._current_nest = 0

    def finish_entry(self):
        return Entry(
            date=self._current_date,
            text=self._current_entry,
        )

    def process_children(self, children: Iterable[Element], tail=False):
        for child in children:
            yield from self.visit_element(child)
            if tail and child.tail:
                self._current_entry += child.tail

    def visit_body(self, body: Body):
        yield from self.process_children(body.children)
        yield self.finish_entry()

    def visit_soft_page_break(self, elem: Element):
        yield from self.visit_line_break(elem)

    def visit_p(self, elem: Paragraph):
        self._current_entry += elem.text
        yield from self.process_children(elem.children, True)
        if self._current_entry:
            if self._current_nest == 0:
                self._current_entry += '\n\n'
            else:
                self._current_entry += '\n'

    visit_h = visit_p

    def visit_list_item(self, elem: ListItem, ordered: bool = False):
        self._current_entry += ' ' * (self._current_nest-1) + '- '
        yield from self.process_children(elem.children, True)

    def visit_line_break(self, elem: PageBreak):
        self._current_entry += '\n'
        if elem.tail:
            self._current_entry += elem.tail

        yield from ()

    def visit_s(self, elem: Element):
        if elem.tail:
            self._current_entry += elem.tail

        yield from ()


    def visit_date(self, elem: Element):
        if self._current_entry:
            yield self.finish_entry()
        self._current_date = dt.datetime.fromisoformat(elem.attributes['text:date-value'])
        if adjust_text := elem.attributes.get('text:date-adjust'):
            self._current_date += isodate.parse_duration(adjust_text)
        self._current_date = self._current_date.date()
        self._current_entry = ''

    def visit_element(self, elem: Element):
        match elem:
            case Element(tag='text:list'):
                self._current_nest += 1
                yield from self.process_children(elem.children)
                self._current_nest -= 1
            case _:
                tag = elem.tag.split(':')[1].replace('-', '_')
                visitor_name = f'visit_{tag}'
                if hasattr(self, visitor_name):
                    yield from getattr(self, visitor_name)(elem)
                else:
                    print(elem.tag, file=sys.stderr)

    def visit_span(self, elem: Span):
        self._current_entry+= elem.text
        yield from self.process_children(elem.children, True)


def main():
    doc = Document(Path(sys.argv[1]).expanduser())

    for entry in Parser().visit_body(doc.body):
        with open(Path(sys.argv[2] if len(sys.argv)>1 else '').expanduser()/f'{entry.date.strftime("%Y_%m_%d")}.md', 'a') as f:
            f.write(entry.text)



if __name__ == "__main__":
    main()