AiDOPWarehouse/doc/_tmp_check_sealed.py

# -*- coding: utf-8 -*-
"""校验盖章版：与原件逐段比对文字，并渲染有章的页面便于目视确认。"""
import io
import os
import sys
import time

from docx import Document

sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding="utf-8")

HERE = os.path.dirname(os.path.abspath(__file__))
BID = r"C:\Users\skygu\OneDrive\Projects\AIDOP\项目\项目招投标"
NAME = "响应文件（完整版）-2026年河南产互联制造业数据智能运营平台研发项目.docx"
SRC = os.path.join(BID, NAME)
OUT = os.path.join(BID, NAME[:-5] + "（已签章）.docx")
PDF = os.path.join(HERE, "_tmp_inspect", "sealed.pdf")
DIR = os.path.join(HERE, "_tmp_inspect")
os.makedirs(DIR, exist_ok=True)

PAGES = [1, 5, 6, 7, 8, 9, 12, 46, 61, 62, 65, 76, 79, 91]


def compare():
    a, b = Document(SRC), Document(OUT)
    pa = [p.text for p in a.paragraphs]
    pb = [p.text for p in b.paragraphs]
    print(f"段落数：原件 {len(pa)}　盖章版 {len(pb)}")
    diff = 0
    for i, (x, y) in enumerate(zip(pa, pb)):
        if x != y:
            diff += 1
            if diff <= 8:
                print(f"  × P{i}\n      原「{x[:56]}」\n      章「{y[:56]}」")
    print(f"段落文字差异 {diff} 处")

    ta = [c.text for t in a.tables for r in t.rows for c in r.cells]
    tb = [c.text for t in b.tables for r in t.rows for c in r.cells]
    print(f"表格数：原件 {len(a.tables)}　盖章版 {len(b.tables)}；"
          f"单元格 {len(ta)} / {len(tb)}，"
          f"文字差异 {sum(1 for x, y in zip(ta, tb) if x != y)} 处")

    ia = len(a.inline_shapes)
    ib = len(b.inline_shapes)
    print(f"嵌入式图片：原件 {ia}　盖章版 {ib}（应相等，印章走浮动图形）")


def export():
    import win32com.client as win32
    word = None
    for i in range(4):
        try:
            word = win32.gencache.EnsureDispatch("Word.Application")
            word.Visible = False
            word.DisplayAlerts = 0
            break
        except Exception as e:
            print(f"  重试 {i + 1}/4：{e}")
            time.sleep(4)
    doc = word.Documents.Open(OUT, ReadOnly=True)
    try:
        print(f"浮动图形数：{doc.Shapes.Count}")
        doc.ExportAsFixedFormat(PDF, 17)
    finally:
        doc.Close(False)
        word.Quit()
    print(f"导出 {PDF}  {os.path.getsize(PDF) / 1024 / 1024:.1f}MB")


compare()
if not os.path.exists(PDF) or "--re" in sys.argv:
    export()

import fitz

d = fitz.open(PDF)
print(f"\nPDF 共 {d.page_count} 页，渲染 {PAGES}")
for p in PAGES:
    if p <= d.page_count:
        d[p - 1].get_pixmap(dpi=100).save(os.path.join(DIR, f"s{p:03d}.png"))
print("完成")
d.close()