table-extractor

table-extractor

热门

>

316Star
69Fork
更新于 2026/1/31
SKILL.md
只读
名称
table-extractor
描述

>

版本
1.0

表格提取技能

概述

本技能能够使用 camelot(PDF表格提取的黄金标准)从PDF文档中精确提取表格。支持处理包含合并单元格、无边框表格以及跨页布局的复杂表格,准确率高。

使用方法

  1. 提供包含表格的PDF文件
  2. 可选指定页面或表格检测方法
  3. 我将提取表格并返回为pandas DataFrame

示例提示:

  • "提取此PDF中的所有表格"
  • "获取这份报告第5页的表格"
  • "提取此文档中的无边框表格"
  • "将PDF表格转换为Excel格式"

领域知识

camelot 基础

import camelot

# 从PDF中提取表格
tables = camelot.read_pdf('document.pdf')

# 访问结果
print(f"找到 {len(tables)} 个表格")

# 获取第一个表格作为DataFrame
df = tables[0].df
print(df)

提取方法

方法 适用场景 描述
lattice 有边框表格 通过线条/边框检测表格
stream 无边框表格 利用文本定位
# Lattice方法(默认)——适用于有可见边框的表格
tables = camelot.read_pdf('document.pdf', flavor='lattice')

# Stream方法——适用于无边框表格
tables = camelot.read_pdf('document.pdf', flavor='stream')

页面选择

# 单页
tables = camelot.read_pdf('document.pdf', pages='1')

# 多页
tables = camelot.read_pdf('document.pdf', pages='1,3,5')

# 页码范围
tables = camelot.read_pdf('document.pdf', pages='1-5')

# 所有页面
tables = camelot.read_pdf('document.pdf', pages='all')

高级选项

Lattice选项
tables = camelot.read_pdf(
    'document.pdf',
    flavor='lattice',
    line_scale=40,              # 线条检测灵敏度
    copy_text=['h', 'v'],       # 跨合并单元格复制文本
    shift_text=['l', 't'],      # 文本对齐偏移
    split_text=True,            # 按换行符拆分文本
    flag_size=True,             # 标记上标/下标
    strip_text='\n',            # 要去除的字符
    process_background=False,   # 处理背景线条
)
Stream选项
tables = camelot.read_pdf(
    'document.pdf',
    flavor='stream',
    edge_tol=500,               # 边缘容差
    row_tol=10,                 # 行容差
    column_tol=0,               # 列容差
    strip_text='\n',            # 要去除的字符
)

表格区域指定

# 从特定区域提取(x1, y1, x2, y2)
# 坐标从左下角开始,单位为PDF点(72点 = 1英寸)
tables = camelot.read_pdf(
    'document.pdf',
    table_areas=['72,720,540,400'],  # 一个区域
)

# 多个区域
tables = camelot.read_pdf(
    'document.pdf',
    table_areas=['72,720,540,400', '72,380,540,200'],
)

列指定

# 手动指定列位置(适用于stream方法)
tables = camelot.read_pdf(
    'document.pdf',
    flavor='stream',
    columns=['100,200,300,400'],  # 列分隔线的X位置
)

处理结果

import camelot

tables = camelot.read_pdf('document.pdf')

for i, table in enumerate(tables):
    # 访问DataFrame
    df = table.df
    
    # 表格元数据
    print(f"表格 {i+1}:")
    print(f"  页面: {table.page}")
    print(f"  准确率: {table.accuracy}")
    print(f"  空白: {table.whitespace}")
    print(f"  顺序: {table.order}")
    print(f"  形状: {df.shape}")
    
    # 解析报告
    report = table.parsing_report
    print(f"  报告: {report}")

导出选项

import camelot

tables = camelot.read_pdf('document.pdf')

# 导出为CSV
tables[0].to_csv('table.csv')

# 导出为Excel
tables[0].to_excel('table.xlsx')

# 导出为JSON
tables[0].to_json('table.json')

# 导出为HTML
tables[0].to_html('table.html')

# 导出所有表格
for i, table in enumerate(tables):
    table.to_excel(f'table_{i+1}.xlsx')

可视化调试

import camelot

# 启用可视化调试
tables = camelot.read_pdf('document.pdf')

# 绘制检测到的表格区域
camelot.plot(tables[0], kind='contour').show()

# 绘制表格上的文本
camelot.plot(tables[0], kind='text').show()

# 绘制检测到的线条(仅lattice)
camelot.plot(tables[0], kind='joint').show()
camelot.plot(tables[0], kind='line').show()

# 保存绘图
fig = camelot.plot(tables[0])
fig.savefig('debug.png')

处理跨页表格

import camelot
import pandas as pd

def extract_multipage_table(pdf_path, pages='all'):
    """提取并合并跨页表格。"""
    
    tables = camelot.read_pdf(pdf_path, pages=pages)
    
    # 按相似结构(列)分组表格
    table_groups = {}
    
    for table in tables:
        cols = tuple(table.df.columns)
        if cols not in table_groups:
            table_groups[cols] = []
        table_groups[cols].append(table.df)
    
    # 合并相似表格
    combined = []
    for cols, dfs in table_groups.items():
        if len(dfs) > 1:
            # 合并并去重表头行
            combined_df = pd.concat(dfs, ignore_index=True)
            combined.append(combined_df)
        else:
            combined.append(dfs[0])
    
    return combined

最佳实践

  1. 尝试两种方法:有边框用lattice,无边框用stream
  2. 检查准确率分数:通常90%以上为良好
  3. 使用可视化调试:理解提取结果
  4. 指定区域:适用于包含多种表格类型的PDF
  5. 处理表头:第一行通常需要特殊处理

常见模式

批量表格提取

import camelot
from pathlib import Path
import pandas as pd

def batch_extract_tables(input_dir, output_dir):
    """提取目录中所有PDF的表格。"""
    
    input_path = Path(input_dir)
    output_path = Path(output_dir)
    output_path.mkdir(exist_ok=True)
    
    results = []
    
    for pdf_file in input_path.glob('*.pdf'):
        try:
            tables = camelot.read_pdf(str(pdf_file), pages='all')
            
            for i, table in enumerate(tables):
                # 跳过低准确率表格
                if table.accuracy < 80:
                    continue
                
                output_file = output_path / f"{pdf_file.stem}_table_{i+1}.xlsx"
                table.to_excel(str(output_file))
                
                results.append({
                    'source': str(pdf_file),
                    'table': i + 1,
                    'page': table.page,
                    'accuracy': table.accuracy,
                    'output': str(output_file)
                })
        
        except Exception as e:
            results.append({
                'source': str(pdf_file),
                'error': str(e)
            })
    
    return results

自动检测表格方法

import camelot

def smart_extract_tables(pdf_path, pages='1'):
    """尝试两种方法并返回最佳结果。"""
    
    # 先尝试lattice
    lattice_tables = camelot.read_pdf(pdf_path, pages=pages, flavor='lattice')
    
    # 再尝试stream
    stream_tables = camelot.read_pdf(pdf_path, pages=pages, flavor='stream')
    
    # 比较并返回最佳
    results = []
    
    if lattice_tables and lattice_tables[0].accuracy > 70:
        results.extend(lattice_tables)
    elif stream_tables:
        results.extend(stream_tables)
    
    return results

示例

示例1:财务报表提取

import camelot
import pandas as pd

def extract_financial_tables(pdf_path):
    """从年报中提取财务表格。"""
    
    # 提取所有表格
    tables = camelot.read_pdf(pdf_path, pages='all', flavor='lattice')
    
    financial_data = {
        'income_statement': None,
        'balance_sheet': None,
        'cash_flow': None,
        'other_tables': []
    }
    
    for table in tables:
        df = table.df
        text = df.to_string().lower()
        
        # 识别表格类型
        if 'revenue' in text or 'sales' in text:
            if 'operating income' in text or 'net income' in text:
                financial_data['income_statement'] = df
        elif 'asset' in text and 'liabilities' in text:
            financial_data['balance_sheet'] = df
        elif 'cash flow' in text or 'operating activities' in text:
            financial_data['cash_flow'] = df
        else:
            financial_data['other_tables'].append({
                'page': table.page,
                'data': df,
                'accuracy': table.accuracy
            })
    
    return financial_data

financials = extract_financial_tables('annual_report.pdf')
if financials['income_statement'] is not None:
    print("找到利润表:")
    print(financials['income_statement'])

示例2:科研数据提取

import camelot
import pandas as pd

def extract_research_data(pdf_path, pages='all'):
    """从研究论文中提取数据表格。"""
    
    # 尝试lattice提取有边框表格
    tables = camelot.read_pdf(pdf_path, pages=pages, flavor='lattice')
    
    if not tables or all(t.accuracy < 70 for t in tables):
        # 回退到stream处理无边框表格
        tables = camelot.read_pdf(pdf_path, pages=pages, flavor='stream')
    
    extracted_data = []
    
    for table in tables:
        df = table.df
        
        # 清理DataFrame
        # 如果第一行看起来像表头,则将其设为列名
        if not df.iloc[0].str.contains(r'\d').any():
            df.columns = df.iloc[0]
            df = df[1:]
            df = df.reset_index(drop=True)
        
        extracted_data.append({
            'page': table.page,
            'accuracy': table.accuracy,
            'data': df
        })
    
    return extracted_data

data = extract_research_data('research_paper.pdf')
for i, item in enumerate(data):
    print(f"表格 {i+1}(第{item['page']}页,准确率:{item['accuracy']}%):")
    print(item['data'].head())

示例3:发票行项目

import camelot

def extract_invoice_items(pdf_path):
    """从发票中提取行项目。"""
    
    # 发票通常有边框表格
    tables = camelot.read_pdf(pdf_path, flavor='lattice')
    
    line_items = []
    
    for table in tables:
        df = table.df
        
        # 查找包含典型发票列的表格
        header_text = ' '.join(df.iloc[0].astype(str)).lower()
        
        if any(term in header_text for term in ['quantity', 'qty', 'amount', 'price', 'description']):
            # 这看起来像行项目表格
            df.columns = df.iloc[0]
            df = df[1:]
            
            for _, row in df.iterrows():
                item = {}
                for col in df.columns:
                    col_lower = str(col).lower()
                    value = row[col]
                    
                    if 'desc' in col_lower or 'item' in col_lower:
                        item['description'] = value
                    elif 'qty' in col_lower or 'quantity' in col_lower:
                        item['quantity'] = value
                    elif 'price' in col_lower or 'rate' in col_lower:
                        item['unit_price'] = value
                    elif 'amount' in col_lower or 'total' in col_lower:
                        item['amount'] = value
                
                if item:
                    line_items.append(item)
    
    return line_items

items = extract_invoice_items('invoice.pdf')
for item in items:
    print(item)

示例4:表格比较

import camelot
import pandas as pd

def compare_pdf_tables(pdf1_path, pdf2_path):
    """比较两个PDF版本之间的表格。"""
    
    tables1 = camelot.read_pdf(pdf1_path)
    tables2 = camelot.read_pdf(pdf2_path)
    
    comparisons = []
    
    # 按形状和位置匹配表格
    for t1 in tables1:
        best_match = None
        best_score = 0
        
        for t2 in tables2:
            if t1.df.shape == t2.df.shape:
                # 计算相似度
                try:
                    similarity = (t1.df == t2.df).mean().mean()
                    if similarity > best_score:
                        best_score = similarity
                        best_match = t2
                except:
                    pass
        
        if best_match:
            comparisons.append({
                'page1': t1.page,
                'page2': best_match.page,
                'similarity': best_score,
                'identical': best_score == 1.0,
                'diff': pd.DataFrame(t1.df != best_match.df)
            })
    
    return comparisons

comparison = compare_pdf_tables('report_v1.pdf', 'report_v2.pdf')

限制

  • 不支持加密PDF
  • 基于图像的PDF需要OCR预处理
  • 非常复杂的合并单元格可能需要调参
  • 旋转的表格需要预处理
  • 大型PDF可能需要逐页处理

安装

pip install camelot-py[cv]

# 额外依赖
# macOS
brew install ghostscript tcl-tk

# Ubuntu
apt-get install ghostscript python3-tk

资源