PDF文档处理专家 Real Skill

一句话说明

全面的 PDF 文档处理工具，支持文本表格提取、PDF 创建、文档操作和表单填写。

PDF Processing Guide - Comprehensive PDF manipulation toolkit for extracting text and tables, creating new PDFs, merging/splitting documents, and handling forms.

---

使用场景

适用：

• 从 PDF 提取文本、表格数据并导出 Excel
• 创建新的 PDF 文档（报告、证书、发票）
• 合并、拆分、旋转 PDF 文档
• 填写 PDF 表单（政府表单、申请表）
• 批量处理多个 PDF 文件

不适用：

• 扫描版 PDF（需先 OCR 处理 / requires OCR first）
• 复杂的 PDF 编辑（用专业软件 / use professional software）

---

核心流程

用户需求 → 判断任务类型 → 选择工具 → 执行操作 → 验证输出
User request → Task type classification → Tool selection → Execute → Validate

任务类型判断 (Task Classification)：

Extracting Data (提取数据)

• 文本提取 (Text extraction) → pdfplumber.extract_text()
  • Better accuracy than pypdf for text extraction
• 表格提取 (Table extraction) → pdfplumber.extract_tables() + pandas
  • Intelligent table structure recognition
  • Export to Excel, CSV, JSON

Creating PDFs (创建 PDF)

• 简单文档 (Simple documents) → reportlab.pdfgen.canvas
  • Quick document creation with text and graphics
• 复杂布局 (Complex layouts) → reportlab.platypus
  • Paragraphs, Tables, Images with advanced styling

PDF Operations (PDF 操作)

• 合并 (Merge) → pypdf.PdfWriter
  • Combine multiple PDFs into one
• 拆分 (Split) → pypdf.PdfWriter (save each page separately)
• 旋转 (Rotate) → page.rotate()
• 元信息 (Metadata) → pypdf.PdfReader.metadata / PdfWriter.add_metadata()

Form Processing (表单处理)

• 查看字段 (View fields) → See forms.md
• 填写表单 (Fill forms) → See forms.md
• 批量处理 (Batch processing) → Loop + form filling API

---

任务完成标准

必须满足（缺一不可）：

• PDF 文件可正常打开
• 提取的数据完整准确
• 表格结构保留正确（行列对应）
• 创建的 PDF 内容清晰可读

质量评级：

• ⭐⭐⭐⭐⭐ 优秀 - 数据完整 + 格式规范 + 自动化批处理
• ⭐⭐⭐ 及格 - 数据正确 + PDF 可用
• ⭐ 失败 - 数据缺失或 PDF 损坏

---

参考资料（供 AI 使用）

类型	路径	说明
核心文档	docs/00-SKILL-完整处理指南.md	Complete PDF processing guide (~300 lines)
表单处理	docs/forms.md	PDF form filling detailed tutorial
高级参考	docs/reference.md	JavaScript libraries and advanced features

---

关键原则（AI 必读 / Critical Principles）

1. Prefer pdfplumber for Extraction (优先使用 pdfplumber)

• 文本提取: More accurate than pypdf
• 表格识别: Strong table recognition capability
• 布局保留: Preserves layout information
• 复杂 PDF: Better for complex PDF processing

2. Table Data Normalization (表格数据规范化)

# Best practice for table extraction
import pdfplumber
import pandas as pd

all_tables = []
with pdfplumber.open("document.pdf") as pdf:
    for page in pdf.pages:
        tables = page.extract_tables()
        for table in tables:
            if table:  # Check not empty
                df = pd.DataFrame(table[1:], columns=table[0])  # First row as header
                all_tables.append(df)

# Combine and clean
combined = pd.concat(all_tables, ignore_index=True)
combined.to_excel("output.xlsx", index=False)

3. Scanned PDFs Require OCR (扫描版需要 OCR)

• 识别: If text extraction returns empty or garbled text
• 处理: Must use OCR tools first (Tesseract, Adobe Acrobat, etc.)
• 告知用户: Clearly inform user that OCR pre-processing is needed

4. Memory Management for Large Files (大文件内存管理)

• Batch processing for multiple PDFs
• Process pages sequentially rather than loading entire document
• Use generators when possible

---

快速命令参考 (Quick Commands)

Text Extraction (文本提取)

import pdfplumber

with pdfplumber.open("document.pdf") as pdf:
    for page in pdf.pages:
        text = page.extract_text()
        print(text)

Table to Excel (表格转 Excel)

import pdfplumber
import pandas as pd

all_tables = []
with pdfplumber.open("document.pdf") as pdf:
    for page in pdf.pages:
        tables = page.extract_tables()
        for table in tables:
            if table:
                df = pd.DataFrame(table[1:], columns=table[0])
                all_tables.append(df)

combined = pd.concat(all_tables, ignore_index=True)
combined.to_excel("extracted_tables.xlsx", index=False)

Merge PDFs (合并 PDF)

from pypdf import PdfWriter, PdfReader

writer = PdfWriter()
for pdf_file in ["doc1.pdf", "doc2.pdf", "doc3.pdf"]:
    reader = PdfReader(pdf_file)
    for page in reader.pages:
        writer.add_page(page)

with open("merged.pdf", "wb") as output:
    writer.write(output)

Split PDF (拆分 PDF)

from pypdf import PdfReader, PdfWriter

reader = PdfReader("large_document.pdf")
for i, page in enumerate(reader.pages):
    writer = PdfWriter()
    writer.add_page(page)
    with open(f"page_{i+1}.pdf", "wb") as output:
        writer.write(output)

Create PDF (创建 PDF)

from reportlab.lib.pagesizes import letter
from reportlab.pdfgen import canvas

c = canvas.Canvas("hello.pdf", pagesize=letter)
width, height = letter
c.drawString(100, height - 100, "Hello World!")
c.save()

Extract Metadata (提取元信息)

from pypdf import PdfReader

reader = PdfReader("document.pdf")
meta = reader.metadata
print(f"Title: {meta.title}")
print(f"Author: {meta.author}")
print(f"Pages: {len(reader.pages)}")

---

依赖安装 (Dependencies Installation)

Python Dependencies

pip install -r requirements.txt
# Includes: pypdf, pdfplumber, reportlab, pandas, openpyxl

---

常见场景示例 (Common Scenarios)

Scenario 1: Batch Extract Invoice Tables (批量提取发票表格)

import pdfplumber
import pandas as pd
import glob

all_data = []
for pdf_file in glob.glob("invoices/*.pdf"):
    with pdfplumber.open(pdf_file) as pdf:
        for page in pdf.pages:
            tables = page.extract_tables()
            for table in tables:
                if table:
                    df = pd.DataFrame(table[1:], columns=table[0])
                    df['source_file'] = pdf_file
                    all_data.append(df)

combined = pd.concat(all_data, ignore_index=True)
combined.to_excel("all_invoices.xlsx", index=False)

Scenario 2: Create Report with Table (生成带表格的报告)

from reportlab.lib.pagesizes import letter
from reportlab.platypus import SimpleDocTemplate, Table, TableStyle, Paragraph
from reportlab.lib.styles import getSampleStyleSheet

doc = SimpleDocTemplate("report.pdf", pagesize=letter)
styles = getSampleStyleSheet()
story = []

# Title
story.append(Paragraph("Sales Report", styles['Title']))

# Table
data = [
    ['Product', 'Q1', 'Q2', 'Q3', 'Q4'],
    ['Product A', '1000', '1200', '1100', '1300'],
    ['Product B', '800', '900', '950', '1000']
]
t = Table(data)
t.setStyle(TableStyle([
    ('BACKGROUND', (0, 0), (-1, 0), 'grey'),
    ('TEXTCOLOR', (0, 0), (-1, 0), 'white'),
    ('ALIGN', (0, 0), (-1, -1), 'CENTER'),
    ('GRID', (0, 0), (-1, -1), 1, 'black')
]))
story.append(t)

doc.build(story)