LWLT-AIBOT/tools/build_business_instruction_docx.py

#!/usr/bin/env python3
"""Build the versioned business-instruction DOCX from its Markdown source.

The existing DOCX is used only as a layout/style template. The document body is
regenerated in full, so the Markdown file remains the single editable source.
"""

from __future__ import annotations

import argparse
import re
from pathlib import Path

from docx import Document
from docx.enum.text import WD_ALIGN_PARAGRAPH, WD_BREAK
from docx.oxml import OxmlElement
from docx.oxml.ns import qn
from docx.shared import Pt, RGBColor


HEADING_STYLE = {
    1: "Doc Title",
    2: "Heading 1",
    3: "Heading 2",
    4: "Heading 3",
}
EAST_ASIA_FONT = "Arial Unicode MS"
LATIN_FONT = "Arial Unicode MS"
CODE_FONT = "Arial Unicode MS"


def parse_args() -> argparse.Namespace:
    parser = argparse.ArgumentParser()
    parser.add_argument("--source", type=Path, required=True)
    parser.add_argument("--template", type=Path, required=True)
    parser.add_argument("--output", type=Path, required=True)
    parser.add_argument("--version", required=True)
    return parser.parse_args()


def set_font(element, latin: str = LATIN_FONT, east_asia: str = EAST_ASIA_FONT) -> None:
    r_pr = element.get_or_add_rPr()
    r_fonts = r_pr.rFonts
    if r_fonts is None:
        r_fonts = OxmlElement("w:rFonts")
        r_pr.insert(0, r_fonts)
    r_fonts.set(qn("w:ascii"), latin)
    r_fonts.set(qn("w:hAnsi"), latin)
    r_fonts.set(qn("w:eastAsia"), east_asia)
    r_fonts.set(qn("w:cs"), latin)


def configure_styles(doc: Document) -> None:
    for style_name in (
        "Normal",
        "Subtitle",
        "Doc Title",
        "Heading 1",
        "Heading 2",
        "Heading 3",
        "Instruction List",
        "Reference Bullet",
        "Code Block",
        "Header",
        "Footer",
    ):
        style = doc.styles[style_name]
        latin = CODE_FONT if style_name == "Code Block" else LATIN_FONT
        style.font.name = latin
        set_font(style.element, latin=latin)

    doc.styles["Normal"].font.size = Pt(10.5)
    doc.styles["Subtitle"].font.size = Pt(9)
    doc.styles["Subtitle"].font.color.rgb = RGBColor(89, 99, 110)
    doc.styles["Doc Title"].font.size = Pt(22)
    doc.styles["Code Block"].font.size = Pt(9)
    for style_name in (
        "Normal",
        "Subtitle",
        "Doc Title",
        "Heading 1",
        "Heading 2",
        "Heading 3",
        "Instruction List",
        "Reference Bullet",
        "Code Block",
    ):
        doc.styles[style_name].paragraph_format.alignment = WD_ALIGN_PARAGRAPH.LEFT


def clear_body(doc: Document) -> None:
    body = doc._element.body
    for child in list(body):
        if child.tag != qn("w:sectPr"):
            body.remove(child)


def normalize_inline(text: str) -> str:
    return text.replace("`", "").replace("  ", " ").strip()


def add_text(doc: Document, text: str, style: str) -> None:
    paragraph = doc.add_paragraph(style=style)
    run = paragraph.add_run(normalize_inline(text))
    latin = CODE_FONT if style == "Code Block" else LATIN_FONT
    run.font.name = latin
    set_font(run._element, latin=latin)


def add_code_block(doc: Document, lines: list[str]) -> None:
    paragraph = doc.add_paragraph(style="Code Block")
    for index, line in enumerate(lines):
        run = paragraph.add_run(line)
        run.font.name = CODE_FONT
        set_font(run._element, latin=CODE_FONT)
        if index < len(lines) - 1:
            run.add_break(WD_BREAK.LINE)


def build_document(source: Path, template: Path, output: Path, version: str) -> None:
    markdown = source.read_text(encoding="utf-8")
    doc = Document(template)
    clear_body(doc)
    configure_styles(doc)

    lines = markdown.splitlines()
    code_lines: list[str] = []
    in_code = False

    for raw_line in lines:
        stripped = raw_line.strip()

        if stripped.startswith("```"):
            if in_code:
                add_code_block(doc, code_lines)
                code_lines = []
            in_code = not in_code
            continue

        if in_code:
            code_lines.append(raw_line.rstrip())
            continue

        if not stripped:
            continue

        heading = re.match(r"^(#{1,4})\s+(.+)$", stripped)
        if heading:
            level = len(heading.group(1))
            title = heading.group(2)
            if level == 1:
                title = f"{title} {version}"
            add_text(doc, title, HEADING_STYLE[level])
            continue

        if stripped.startswith("交付版本：") or stripped.startswith("适用范围："):
            add_text(doc, stripped, "Subtitle")
            continue

        if re.match(r"^\d+\.\s+", stripped):
            add_text(doc, stripped, "Instruction List")
            continue

        if stripped.startswith("- "):
            add_text(doc, f"• {stripped[2:]}", "Reference Bullet")
            continue

        add_text(doc, stripped, "Normal")

    if in_code:
        raise ValueError("Unclosed fenced code block in Markdown source")

    doc.core_properties.title = f"运营业务 AI 输入模板总表 {version}"
    doc.core_properties.subject = "老挝联泰 ERP 业务指令模板与示例"
    doc.core_properties.comments = (
        f"Generated from {source.as_posix()} by tools/build_business_instruction_docx.py"
    )
    output.parent.mkdir(parents=True, exist_ok=True)
    doc.save(output)


def main() -> None:
    args = parse_args()
    build_document(args.source, args.template, args.output, args.version)


if __name__ == "__main__":
    main()