md_to_gost/md2gost/renderable/heading.py

from copy import copy
from typing import Generator
import re

from docx.text.paragraph import Paragraph as DocxParagraph
from docx.shared import Parented, Length

from .paragraph_sizer import ParagraphSizer
from ..layout_tracker import LayoutState
from .paragraph import Paragraph
from ..rendered_info import RenderedInfo
from ..sub_renderable import SubRenderable
from ..util import create_element


# Leading section numbers written in markdown: "1 ", "1.1 ", "1.2.3. "
_LEADING_NUMBER_RE = re.compile(r"^\d+(?:\.\d+)*\.?\s+")


class Heading(Paragraph):
    def __init__(self, parent: Parented, level: int, numbered: bool,
                 numbering_mode: str = "manual"):
        super().__init__(parent)

        self._numbered = numbered
        self._numbering_mode = numbering_mode  # auto | manual
        self._parent = parent
        self._level = level

        if not 1 <= level <= 9:
            raise ValueError("Heading level must be in range from 1 to 9")

        self.style = f"Heading {level}"

        if not numbered:
            self._remove_numbering()
            # Alignment decided later in Renderer: only СОДЕРЖАНИЕ / СПИСОК centered.
            # ВВЕДЕНИЕ / ЗАКЛЮЧЕНИЕ / ПРИЛОЖЕНИЯ stay left like Heading 1 (1.25 cm indent).
        elif numbering_mode == "manual":
            # Digits already in markdown text — kill Word list numbering to avoid "1 1 …"
            self._remove_numbering()

        self._rendered_page = 0
        self._title_stripped = False

    @property
    def is_numbered(self) -> bool:
        return self._numbered

    @property
    def numbering_mode(self) -> str:
        return self._numbering_mode

    @numbering_mode.setter
    def numbering_mode(self, value: str):
        self._numbering_mode = value
        if self._numbered and value == "manual":
            self._remove_numbering()

    @property
    def rendered_page(self) -> int:
        return self._rendered_page

    @property
    def level(self) -> int:
        return self._level

    @property
    def text(self) -> str:
        return self._docx_paragraph.text

    def apply_numbering_mode(self, mode: str):
        """Apply auto/manual after runs are filled."""
        self._numbering_mode = mode
        if not self._numbered:
            return
        if mode == "manual":
            self._remove_numbering()
        elif mode == "auto":
            self._strip_leading_number_from_runs()

    def _strip_leading_number_from_runs(self):
        """Remove '1 ', '1.1 ' etc. from text so Word list provides the number."""
        if self._title_stripped or not self._numbered:
            return
        full = self._docx_paragraph.text or ""
        m = _LEADING_NUMBER_RE.match(full)
        if not m:
            return
        # Rewrite runs: put remaining text into first run, clear the rest
        remaining = full[m.end():]
        runs = self._docx_paragraph.runs
        if not runs:
            return
        runs[0].text = remaining
        for r in runs[1:]:
            r.text = ""
        self._title_stripped = True

    def _remove_numbering(self):
        pPr = self._docx_paragraph._p.get_or_add_pPr()
        # Remove existing numPr if any, then set numId=0
        for child in list(pPr):
            if child.tag.endswith("}numPr"):
                pPr.remove(child)
        pPr.append(
            create_element("w:numPr", [
                create_element("w:ilvl", {
                    "w:val": "0"
                }),
                create_element("w:numId", {
                    "w:val": "0"
                })
            ])
        )

    def exclude_from_toc(self):
        """Keep Heading N look, but outline level = Body Text so Word TOC skips it."""
        pPr = self._docx_paragraph._p.get_or_add_pPr()
        for child in list(pPr):
            if child.tag.endswith("}outlineLvl"):
                pPr.remove(child)
        # 9 = body text (not outline levels 1–3 used by TOC \o "1-3")
        pPr.append(create_element("w:outlineLvl", {"w:val": "9"}))

    def render(self, previous_rendered: RenderedInfo, layout_state: LayoutState)\
            -> Generator[RenderedInfo | SubRenderable, None, None]:
        remaining_height = layout_state.remaining_page_height

        if self._level == 1 and layout_state.page != 1 and\
                not (isinstance(previous_rendered.docx_element, DocxParagraph)
                     and previous_rendered.docx_element.text == "\n"):
            self.page_break_before = True

        height_data = ParagraphSizer(
            self._docx_paragraph,
            previous_rendered.docx_element
            if previous_rendered and isinstance(previous_rendered.docx_element, DocxParagraph) else None,
            layout_state.max_width).calculate_height()

        if layout_state.current_page_height == 0 and layout_state.page != 1:
            height_data.before = 0

        # if a heading + 3 lines don't fit to the page, they go to the next page
        if ((height_data.lines + 3 - 1) * height_data.line_spacing + 1) * height_data.line_height\
                > layout_state.remaining_page_height:
            self._docx_paragraph.paragraph_format.space_before = 0  # libreoffice fix
            height = height_data.full - height_data.before

            # force this behaviour as there could be a table or an image instead of text
            self.page_break_before = True

        else:
            height = height_data.full

        if self.page_break_before:
            height += remaining_height

        layout_state.add_height(height)
        self._rendered_page = layout_state.page

        yield RenderedInfo(self._docx_paragraph, Length(height))