AiDOPWarehouse/doc/_bid_docx_kit.py

# -*- coding: utf-8 -*-
"""技术响应文件（附件五，格式自拟）Word 排版工具。

统一版式：A4、宋体小四正文、黑体各级标题、图表居中并自动编号。
内容侧只描述结构（h2/p/ul/fig/table），排版细节全部收敛在这里。
"""
import os
import re

from docx import Document
from docx.enum.section import WD_ORIENT, WD_SECTION
from docx.enum.table import WD_TABLE_ALIGNMENT
from docx.enum.text import WD_ALIGN_PARAGRAPH
from docx.oxml import OxmlElement
from docx.oxml.ns import qn
from docx.shared import Cm, Pt, RGBColor
from PIL import Image

HERE = os.path.dirname(os.path.abspath(__file__))
ASSETS = os.path.join(HERE, "_bid_assets")

SONG = "宋体"
HEI = "黑体"
LATIN = "Times New Roman"

BODY_PT = 12.0      # 小四
CAP_PT = 10.5       # 五号
TBL_PT = 9.0

# 版面几何随当前节（纵向／横向）切换，图与表都按当前正文宽度排版
PORTRAIT = {"w": 14.66, "max_fig_h": 19.0, "full_fig_h": 22.5}
LANDSCAPE = {"w": 26.70, "max_fig_h": 12.6, "full_fig_h": 17.2}
GEO = dict(PORTRAIT)

HEAD_FILL = "DDEBF7"
TITLE_C = RGBColor(0x1F, 0x4E, 0x79)
BLACK = RGBColor(0, 0, 0)


# ------------------------------------------------------------------ 基础
def _font(run, name_cn, size_pt, bold=False, color=None, italic=False):
    run.font.size = Pt(size_pt)
    run.font.bold = bold
    run.font.italic = italic
    run.font.name = LATIN
    if color is not None:
        run.font.color.rgb = color
    rPr = run._element.get_or_add_rPr()
    rf = rPr.find(qn("w:rFonts"))
    if rf is None:
        rf = OxmlElement("w:rFonts")
        rPr.insert(0, rf)
    rf.set(qn("w:ascii"), LATIN)
    rf.set(qn("w:hAnsi"), LATIN)
    rf.set(qn("w:eastAsia"), name_cn)


def _spacing(p, before=0, after=0, line=1.5, first_indent=0.0):
    pf = p.paragraph_format
    pf.space_before = Pt(before)
    pf.space_after = Pt(after)
    pf.line_spacing = line
    if first_indent:
        pf.first_line_indent = Pt(first_indent)


def new_doc():
    doc = Document()
    sec = doc.sections[0]
    sec.page_width = Cm(21.0)
    sec.page_height = Cm(29.7)
    sec.top_margin = Cm(2.54)
    sec.bottom_margin = Cm(2.54)
    sec.left_margin = Cm(3.17)
    sec.right_margin = Cm(3.17)
    st = doc.styles["Normal"]
    st.font.size = Pt(BODY_PT)
    st.font.name = LATIN
    st.element.rPr.rFonts.set(qn("w:eastAsia"), SONG)
    _add_page_footer(sec)
    return doc


def _add_page_footer(sec):
    # 新节默认沿用上一节页脚，若不断开会把页码域重复追加到同一段落
    sec.footer.is_linked_to_previous = False
    ftr = sec.footer
    for extra in list(ftr.paragraphs[1:]):
        extra._p.getparent().remove(extra._p)
    p = ftr.paragraphs[0] if ftr.paragraphs else ftr.add_paragraph()
    for r in list(p.runs):
        r._element.getparent().remove(r._element)
    p.alignment = WD_ALIGN_PARAGRAPH.CENTER
    for kind, payload in (("begin", None), ("instr", " PAGE  \\* MERGEFORMAT "),
                          ("separate", None), ("text", "1"), ("end", None)):
        r = OxmlElement("w:r")
        if kind == "instr":
            it = OxmlElement("w:instrText")
            it.set(qn("xml:space"), "preserve")
            it.text = payload
            r.append(it)
        elif kind == "text":
            t = OxmlElement("w:t")
            t.text = payload
            r.append(t)
        else:
            fc = OxmlElement("w:fldChar")
            fc.set(qn("w:fldCharType"), kind)
            r.append(fc)
        p._p.append(r)
    for run in p.runs:
        _font(run, SONG, CAP_PT)


# ---------------------------------------------------------- 合并入册时的改号
# 各章单独成文时编号自成一套（第 1 章 → 1.1、图 1-1）；并入响应文件后要挂到
# 「六（三）」下重新排序，故提供整章改号与大纲级别下移两个开关。
RENUM = None     # (旧章号, 新章号)
OUTLINE = None   # {标题级别: 大纲级别值}，值为 0 起算


def set_renum(old, new):
    global RENUM
    RENUM = None if old is None else (int(old), int(new))


def set_outline(mapping):
    global OUTLINE
    OUTLINE = mapping


def _renum_head(text):
    """小节标题的首段编号：1.2.3 → 4.2.3。只改开头，不碰标题里的其它数字。"""
    if not RENUM:
        return text
    old, new = RENUM
    m = re.match(rf"^{old}((?:\.\d+)*)(?=[\s　]|$)", text)
    return f"{new}{m.group(1)}{text[m.end():]}" if m else text


def _renum_body(text):
    """正文里的交叉引用「详见 1.8.2 节」→「详见 4.8.2 节」。

    只在「见」字之后替换，避免误伤版本号、倍数、金额等小数。
    """
    if not RENUM:
        return text
    old, new = RENUM
    return re.sub(rf"(?<=见)(\s*){old}((?:\.\d+)+)",
                  lambda m: f"{m.group(1)}{new}{m.group(2)}", text)


# ------------------------------------------------------------------ 段落
def _heading(doc, text, level, name_cn, size_pt, color, bold=False,
             before=0, after=0, center=False):
    """单章成文时套内置 Heading 样式，供 Word 目录域识别；

    外观全部用行内格式锁死，避免继承样式自带的蓝色、斜体与字体。
    并入总册时（OUTLINE 生效）改用 Normal 样式 + 显式大纲级别：目录域的 \\o
    开关按「内置标题样式」取条目并以样式级别为准，会无视行内的大纲级别覆盖，
    致使章内三级标题也被收进总目录；去掉样式后全册统一由大纲级别决定层级。
    """
    p = doc.add_paragraph(style="Normal" if OUTLINE else f"Heading {level}")
    if center:
        p.alignment = WD_ALIGN_PARAGRAPH.CENTER
    _spacing(p, before=before, after=after, line=1.5)
    p.paragraph_format.keep_with_next = True
    if OUTLINE and level in OUTLINE:
        pPr = p._p.get_or_add_pPr()
        old = pPr.find(qn("w:outlineLvl"))
        if old is not None:
            pPr.remove(old)
        ol = OxmlElement("w:outlineLvl")
        ol.set(qn("w:val"), str(OUTLINE[level]))
        pPr.append(ol)
    _font(p.add_run(_renum_head(text)), name_cn, size_pt, bold=bold,
          color=color, italic=False)
    return p


def h1(doc, text):
    return _heading(doc, text, 1, HEI, 16, TITLE_C,
                    before=6, after=16, center=True)


def h2(doc, text):
    return _heading(doc, text, 2, HEI, 14, TITLE_C, before=14, after=8)


def h3(doc, text):
    return _heading(doc, text, 3, HEI, 12, BLACK, before=10, after=6)


def h4(doc, text):
    return _heading(doc, text, 4, SONG, BODY_PT, BLACK, bold=True,
                    before=8, after=4)


def para(doc, text, indent=True):
    p = doc.add_paragraph()
    _spacing(p, after=4, line=1.5, first_indent=BODY_PT * 2 if indent else 0)
    _font(p.add_run(_renum_body(text)), SONG, BODY_PT)
    return p


def bullets(doc, items):
    """无编号要点：用「—」引导，左缩进 2 字符，避免依赖 Word 列表样式。"""
    for it in items:
        p = doc.add_paragraph()
        _spacing(p, after=3, line=1.5)
        p.paragraph_format.left_indent = Pt(BODY_PT * 2)
        p.paragraph_format.first_line_indent = Pt(-BODY_PT)
        _font(p.add_run("— " + it), SONG, BODY_PT)


def numbered(doc, items):
    for i, it in enumerate(items, 1):
        p = doc.add_paragraph()
        _spacing(p, after=3, line=1.5)
        p.paragraph_format.left_indent = Pt(BODY_PT * 2)
        p.paragraph_format.first_line_indent = Pt(-BODY_PT * 2)
        _font(p.add_run(f"（{i}）{it}"), SONG, BODY_PT)


# ------------------------------------------------------------------ 图
class Counter:
    def __init__(self, chapter):
        self.chapter = chapter
        self.fig = 0
        self.tbl = 0

    def next_fig(self):
        self.fig += 1
        return f"图 {self.chapter}-{self.fig}"

    def next_tbl(self):
        self.tbl += 1
        return f"表 {self.chapter}-{self.tbl}"


def figure(doc, cnt, name, caption, full_page=False, max_h=None):
    """full_page：图另起一页单独放，用于纵横比接近正方、缩到正文流里会看不清的图。

    max_h：单独指定本图高度上限（厘米），用于界面截图这类必须放大才可辨认的图。
    """
    path = os.path.join(ASSETS, name if name.endswith(".png") else name + ".png")
    if not os.path.exists(path):
        raise FileNotFoundError(path)
    if full_page:
        doc.add_page_break()
    w_px, h_px = Image.open(path).size
    limit = max_h or (GEO["full_fig_h"] if full_page else GEO["max_fig_h"])
    w_cm = GEO["w"]
    h_cm = w_cm * h_px / w_px
    if h_cm > limit:
        h_cm = limit
        w_cm = h_cm * w_px / h_px
    p = doc.add_paragraph()
    p.alignment = WD_ALIGN_PARAGRAPH.CENTER
    _spacing(p, before=5, after=1, line=1.0)
    p.add_run().add_picture(path, width=Cm(w_cm))
    cp = doc.add_paragraph()
    cp.alignment = WD_ALIGN_PARAGRAPH.CENTER
    _spacing(cp, after=8, line=1.0)
    _font(cp.add_run(f"{cnt.next_fig()}　{caption}"), SONG, CAP_PT)


# ------------------------------------------------------------------ 表
def _shade(cell, hexcolor):
    tcPr = cell._tc.get_or_add_tcPr()
    sh = OxmlElement("w:shd")
    sh.set(qn("w:val"), "clear")
    sh.set(qn("w:fill"), hexcolor)
    tcPr.append(sh)


def _fixed_layout(tb):
    tblPr = tb._tbl.tblPr
    el = OxmlElement("w:tblLayout")
    el.set(qn("w:type"), "fixed")
    tblPr.append(el)


def _repeat_header(row):
    """跨页时自动重复表头行。"""
    trPr = row._tr.get_or_add_trPr()
    el = OxmlElement("w:tblHeader")
    el.set(qn("w:val"), "true")
    trPr.append(el)


def table(doc, cnt, caption, headers, rows, widths=None, align=None):
    """widths：各列占比（会归一化到正文宽度）；align：每列 'l'/'c'。"""
    cp = doc.add_paragraph()
    cp.alignment = WD_ALIGN_PARAGRAPH.CENTER
    _spacing(cp, before=10, after=3, line=1.0)
    _font(cp.add_run(f"{cnt.next_tbl()}　{caption}"), SONG, CAP_PT)

    n = len(headers)
    tb = doc.add_table(rows=1, cols=n)
    tb.style = "Table Grid"
    tb.alignment = WD_TABLE_ALIGNMENT.CENTER
    tb.autofit = False
    _fixed_layout(tb)

    widths = widths or [1] * n
    total = float(sum(widths))
    cols_cm = [GEO["w"] * w / total for w in widths]
    align = align or (["c"] + ["l"] * (n - 1))

    def put(cell, text, is_head, col):
        cell.width = Cm(cols_cm[col])
        p = cell.paragraphs[0]
        _spacing(p, before=2, after=2, line=1.15)
        p.alignment = (WD_ALIGN_PARAGRAPH.CENTER if is_head or align[col] == "c"
                       else WD_ALIGN_PARAGRAPH.LEFT)
        _font(p.add_run(str(text)), HEI if is_head else SONG, TBL_PT,
              bold=False)
        if is_head:
            _shade(cell, HEAD_FILL)

    for j, htxt in enumerate(headers):
        put(tb.rows[0].cells[j], htxt, True, j)
    _repeat_header(tb.rows[0])
    for r in rows:
        cells = tb.add_row().cells
        for j, v in enumerate(r):
            put(cells[j], v, False, j)
    for j, cm in enumerate(cols_cm):
        for row in tb.rows:
            row.cells[j].width = Cm(cm)
    doc.add_paragraph().paragraph_format.space_after = Pt(6)
    return tb


def page_break(doc):
    doc.add_page_break()


def landscape_section(doc):
    """切换到横向节：架构图信息密度高，纵向 A4 上字号会小到看不清。"""
    global GEO
    sec = doc.add_section(WD_SECTION.NEW_PAGE)
    sec.orientation = WD_ORIENT.LANDSCAPE
    sec.page_width, sec.page_height = Cm(29.7), Cm(21.0)
    sec.top_margin = sec.bottom_margin = Cm(1.5)
    sec.left_margin = sec.right_margin = Cm(1.5)
    _add_page_footer(sec)
    GEO = dict(LANDSCAPE)
    return sec


def portrait_section(doc):
    global GEO
    sec = doc.add_section(WD_SECTION.NEW_PAGE)
    sec.orientation = WD_ORIENT.PORTRAIT
    sec.page_width, sec.page_height = Cm(21.0), Cm(29.7)
    sec.top_margin = sec.bottom_margin = Cm(2.54)
    sec.left_margin = sec.right_margin = Cm(3.17)
    _add_page_footer(sec)
    GEO = dict(PORTRAIT)
    return sec


def render(doc, cnt, blocks):
    """按内容块列表渲染。块形如 ("h2", "标题") / ("fig", (name, caption)) 等。"""
    for kind, payload in blocks:
        if kind == "h2":
            h2(doc, payload)
        elif kind == "h3":
            h3(doc, payload)
        elif kind == "h4":
            h4(doc, payload)
        elif kind == "p":
            para(doc, payload)
        elif kind == "ul":
            bullets(doc, payload)
        elif kind == "ol":
            numbered(doc, payload)
        elif kind == "fig":
            max_h = payload[2] if len(payload) > 2 else None
            figure(doc, cnt, payload[0], payload[1], max_h=max_h)
        elif kind == "figfull":
            figure(doc, cnt, payload[0], payload[1], full_page=True)
        elif kind == "table":
            cap, headers, rows = payload[0], payload[1], payload[2]
            widths = payload[3] if len(payload) > 3 else None
            align = payload[4] if len(payload) > 4 else None
            table(doc, cnt, cap, headers, rows, widths, align)
        elif kind == "pagebreak":
            page_break(doc)
        elif kind == "landscape":
            landscape_section(doc)
        elif kind == "portrait":
            portrait_section(doc)
        else:
            raise ValueError(f"未知内容块：{kind}")