使用Python实现按卷内目录拆分PDF归档册

pip install spire.pdf.free

# -*- coding: utf-8 -*-
import os
import re
from spire.pdf import (pdfdocument, pdftextextractor, pdftextextractoptions,
                       pdfmargins, pointf, fileformat, pdfcompressionlevel,
                       pdfimagehelper)

archive = "项目资料归档册_2026q3.pdf"
outdir = "output"
os.makedirs(outdir, exist_ok=true)
row = re.compile(r"^\s*(\d+)\s+((?:ht|fp)-\s+)\s+(\s+)\s+(\d+(?:-\d+)?)\s+(\s+)\s*$")
contract = "合同"


def read_catalog_text(path):
    """目录页固定在第 1 页，这也是唯一需要抽取文本的页面。"""
    doc = pdfdocument()
    doc.loadfromfile(path)
    text = pdftextextractor(doc.pages[0]).extracttext(pdftextextractoptions())
    total = doc.pages.count
    doc.close()
    return text, total


def parse_catalog(text):
    items = []
    for raw in text.splitlines():
        m = row.match(raw)
        if not m:
            continue
        a, _, b = m.group(4).partition("-")          # 页段 "2-3" 拆成起始页与结束页
        items.append({
            "no": m.group(2),                         # 文件编号
            "type": m.group(3),                       # 文件类型
            "pages": list(range(int(a), int(b or a) + 1)),
            "parent": none if m.group(5) == "-" else m.group(5),   # 关联编号
        })
    return items

def collect(src, phys_pages, phys_to_target, out_path):
    """按物理页序取出页段，在内存中重排到目标页序后保存。"""
    doc = pdfdocument()
    doc.loadfromfile(src)
    out = pdfdocument()
    for n in phys_pages:
        page = out.pages.add(doc.pages[n - 1].size, pdfmargins(0.0))
        doc.pages[n - 1].createtemplate().draw(page, pointf(0.0, 0.0))
    out.pages.rearrange([phys_pages.index(n) for n in phys_to_target])
    out.compressionlevel = pdfcompressionlevel.best
    out.fileinfo.incrementalupdate = false
    helper = pdfimagehelper()
    for i in range(out.pages.count):
        for info in helper.getimagesinfo(out.pages[i]):
            info.trycompressimage()
    out.savetofile(out_path, fileformat.pdf)
    out.close()
    doc.close()

def page_count(path):
    d = pdfdocument()
    d.loadfromfile(path)
    n = d.pages.count
    d.close()
    return n


text, total_pages = read_catalog_text(archive)
catalog = parse_catalog(text)
print("归档册：%s（共 %d 页）" % (os.path.basename(archive), total_pages))
print("卷内目录解析到 %d 条著录：" % len(catalog))
print("件号 | 文件编号 | 文件类型 | 页段 | 关联编号")
for idx, it in enumerate(catalog, 1):
    seg = str(it["pages"][0]) if len(it["pages"]) == 1 else "%d-%d" % (it["pages"][0], it["pages"][-1])
    print("%d | %s | %s | %s | %s" % (idx, it["no"], it["type"], seg, it["parent"] or "-"))

contracts = [i for i in catalog if i["parent"] is none and contract in i["type"]]
print("\n共 %d 份合同，逐份归集：\n" % len(contracts))

made, missing = [], []
for c in contracts:
    invoices = [i for i in catalog if i["parent"] == c["no"]]
    related = sorted([c] + invoices, key=lambda x: x["pages"][0])
    phys = [p for i in related for p in i["pages"]]          # 原册中的物理页序
    target = list(c["pages"]) + [p for i in related if i is not c for p in i["pages"]]
    out = os.path.join(outdir, "归档件_%s.pdf" % c["no"])
    collect(archive, phys, target, out)
    made.append(out)
    if invoices:
        status, note = "完整", "合同 %d 页 + 发票 %d 页" % (len(c["pages"]), len(invoices))
    else:
        status, note = "缺件", "合同 %d 页，未匹配到发票" % len(c["pages"])
        missing.append(c["no"])
    order = "已按合同在前重排" if phys != target else "物理页序即目标页序"
    print("  %-13s %s  %s（%s）→ %s" % (c["no"], status, note, order, os.path.basename(out)))
    print("      物理页序 %s → 目标页序 %s，输出 %d 页" % (phys, target, page_count(out)))

book = os.path.join(outdir, "季度归档合订本_2026q3.pdf")
merged = pdfdocument()
toc = pdfdocument()
toc.loadfromfile(archive)
pg = merged.pages.add(toc.pages[0].size, pdfmargins(0.0))
toc.pages[0].createtemplate().draw(pg, pointf(0.0, 0.0))
parts = [toc]
for out in sorted(made):
    part = pdfdocument()
    part.loadfromfile(out)
    merged.appendpage(part)
    parts.append(part)
merged.compressionlevel = pdfcompressionlevel.best
merged.fileinfo.incrementalupdate = false
merged.savetofile(book, fileformat.pdf)
merged.close()
for part in parts:
    part.close()

print("\n季度归档合订本：%s（%d 页）" % (os.path.basename(book), page_count(book)))
print("缺件提醒：%s" % ("、".join(missing) if missing else "无"))

def first_line(path):
    d = pdfdocument()
    d.loadfromfile(path)
    t = pdftextextractor(d.pages[0]).extracttext(pdftextextractoptions())
    d.close()
    for line in t.splitlines():
        if line.strip():
            return line.strip()
    return ""
lines = ["项目资料归档册拆分结果 2026q3",
         "归档册：%s（%d 页）" % (os.path.basename(archive), total_pages),
         "目录著录 %d 条，合同 %d 份" % (len(catalog), len(contracts)),
         ""]
for out in made:
    lines.append("%-26s %d 页  首页：%s" % (os.path.basename(out), page_count(out), first_line(out)))
lines.append("%-26s %d 页" % (os.path.basename(book), page_count(book)))
lines.append("")
lines.append("缺发票的合同：%s" % ("、".join(missing) if missing else "无"))
open(os.path.join(outdir, "归档处理报告.txt"), "w", encoding="utf-8").write("\n".join(lines))
print("\n处理报告已写出。")

归档册：项目资料归档册_2026q3.pdf（共 8 页）
卷内目录解析到 5 条著录：
件号 | 文件编号 | 文件类型 | 页段 | 关联编号
1 | ht-2026-0142 | 设备采购合同 | 2-3 | -
2 | fp-2026-08851 | 增值税专用发票 | 4 | ht-2026-0142
3 | fp-2026-08907 | 增值税专用发票 | 5 | ht-2026-0157
4 | ht-2026-0157 | 技术服务合同 | 6-7 | -
5 | ht-2026-0173 | 设备采购合同 | 8 | -
共 3 份合同，逐份归集：
  ht-2026-0142  完整  合同 2 页 + 发票 1 页（物理页序即目标页序）→ 归档件_ht-2026-0142.pdf
      物理页序 [2, 3, 4] → 目标页序 [2, 3, 4]，输出 3 页
  ht-2026-0157  完整  合同 2 页 + 发票 1 页（已按合同在前重排）→ 归档件_ht-2026-0157.pdf
      物理页序 [5, 6, 7] → 目标页序 [6, 7, 5]，输出 3 页
  ht-2026-0173  缺件  合同 1 页，未匹配到发票（物理页序即目标页序）→ 归档件_ht-2026-0173.pdf
      物理页序 [8] → 目标页序 [8]，输出 1 页
季度归档合订本：季度归档合订本_2026q3.pdf（8 页）
缺件提醒：ht-2026-0173
处理报告已写出。