Python利用PyMuPDF实现高效提取PDF文本与图片

pip install PyMuPDF

import fitz  # 正确导入方式
# import pymupdf   # ❌ 错误！不能这样导入

# Document（文档类）- 代表整个PDF文件
doc = fitz.open("document.pdf")

# Page（页面类）- 代表PDF中的单个页面
page = doc[0]  # 获取第一页
# 或使用 load_page 方法
page = doc.load_page(0)

# Rect（矩形类）- 用于定义页面区域或裁剪范围
rect = fitz.Rect(100, 100, 500, 700)  # (x0, y0, x1, y1)

import fitz

# 1. 打开PDF文档
doc = fitz.open("example.pdf")

# 2. 获取文档基本信息
print(f"总页数: {doc.page_count}")
print(f"元数据: {doc.metadata}")  # 标题、作者、创建日期等

# 3. 加载第一页并提取文本
page = doc[0]
text = page.get_text()
print("第一页文本内容:")
print(text[:200])  # 打印前200个字符

# 4. 关闭文档（释放资源）
doc.close()

import fitz

doc = fitz.open("document.pdf")
text = ""

for page_num in range(doc.page_count):
    page = doc[page_num]
    text += page.get_text("text") + "\n"

print(f"提取了 {len(text)} 个字符")
doc.close()

import fitz

doc = fitz.open("layout.pdf")
page = doc[0]

# 提取带坐标的文本块
blocks = page.get_text("blocks")

for block in blocks:
    # block格式: (x0, y0, x1, y1, "文本内容", 块编号, 块类型)
    x0, y0, x1, y1, content, block_no, block_type = block
    print(f"块{block_no}: 位置[{x0:.2f},{y0:.2f}]到[{x1:.2f},{y1:.2f}]")
    print(f"内容: {content.strip()}")
    print("-" * 50)

doc.close()

import fitz

doc = fitz.open("two_column_article.pdf")
page = doc[0]

# 定义感兴趣的区域（矩形框，单位为点points，1pt ≈ 1/72英寸）
# 例如：截取页面左半部分
left_half = fitz.Rect(0, 0, page.rect.width / 2, page.rect.height)

# 按区域提取文本
region_text = page.get_text("text", clip=left_half)
print("左半区域文本:")
print(region_text)

# 按区域提取文本块（保留位置信息）
region_blocks = page.get_text("blocks", clip=left_half)
for block in region_blocks:
    print(block[4])  # 文本内容

doc.close()

import fitz
import os

def extract_images_from_pdf(pdf_path, output_dir="extracted_images"):
    """
    从PDF中提取所有图片，保存到指定目录
    
    Args:
        pdf_path: PDF文件路径
        output_dir: 图片输出目录
    """
    # 创建输出目录
    os.makedirs(output_dir, exist_ok=True)
    
    # 打开PDF
    doc = fitz.open(pdf_path)
    
    img_count = 0
    for page_num in range(doc.page_count):
        page = doc[page_num]
        
        # 获取当前页的所有图片
        image_list = page.get_images(full=True)
        
        for img_index, img in enumerate(image_list):
            xref = img[0]  # 图片对象的引用ID
            
            # 提取图片数据
            base_image = doc.extract_image(xref)
            image_bytes = base_image["image"]  # 图片原始字节数据
            image_ext = base_image["ext"]      # 图片扩展名
            image_width = base_image["width"]  # 图片宽度
            image_height = base_image["height"]  # 图片高度
            
            # 构造保存文件名
            img_filename = f"page{page_num+1}_img{img_index+1}.{image_ext}"
            img_path = os.path.join(output_dir, img_filename)
            
            # 保存图片
            with open(img_path, "wb") as img_file:
                img_file.write(image_bytes)
            
            img_count += 1
            print(f"保存: {img_filename} (尺寸: {image_width}x{image_height})")
    
    doc.close()
    print(f"提取完成！共提取 {img_count} 张图片，保存在 {output_dir}")

# 使用示例
extract_images_from_pdf("report.pdf", "images_output")

import fitz

def is_scanned_pdf(pdf_path):
    """
    判断PDF是否为扫描件（无文字层）
    
    Args:
        pdf_path: PDF文件路径
    
    Returns:
        True: 扫描件（无文字层）
        False: 文字型PDF（有文字层）
    """
    doc = fitz.open(pdf_path)
    
    for page_num in range(doc.page_count):
        page = doc[page_num]
        text = page.get_text("text")
        
        # 如果某一页有文字（去除空白后长度>0），则不是扫描件
        if len(text.strip()) > 50:  # 至少50个非空白字符
            doc.close()
            return False
    
    doc.close()
    return True

# 使用示例
if is_scanned_pdf("scanned_document.pdf"):
    print("⚠️ 这是一个扫描件PDF，需要先进行OCR处理")
else:
    print("✅ 这是一个文字型PDF，可直接提取文本")

def analyze_page_text_status(pdf_path):
    """分析PDF各页的文字层状态"""
    doc = fitz.open(pdf_path)
    
    results = []
    for page_num in range(doc.page_count):
        page = doc[page_num]
        text = page.get_text("text")
        char_count = len(text.strip())
        
        status = "扫描件" if char_count == 0 else f"文字型 ({char_count}字符)"
        results.append({"page": page_num + 1, "status": status, "char_count": char_count})
    
    doc.close()
    return results

# 输出分析结果
analysis = analyze_page_text_status("mixed_document.pdf")
for item in analysis:
    print(f"第{item['page']}页: {item['status']}")

import fitz
import os

def batch_extract_to_txt(input_folder, output_folder=None):
    """
    批量提取文件夹中所有PDF的文本，每个PDF生成一个TXT文件
    
    Args:
        input_folder: PDF文件所在文件夹
        output_folder: 输出TXT文件夹（默认为input_folder下的extracted_txt）
    """
    if output_folder is None:
        output_folder = os.path.join(input_folder, "extracted_txt")
    
    os.makedirs(output_folder, exist_ok=True)
    
    # 获取所有PDF文件
    pdf_files = [f for f in os.listdir(input_folder) if f.lower().endswith('.pdf')]
    
    print(f"找到 {len(pdf_files)} 个PDF文件")
    
    for idx, pdf_name in enumerate(pdf_files, 1):
        pdf_path = os.path.join(input_folder, pdf_name)
        txt_name = pdf_name.replace('.pdf', '.txt')
        txt_path = os.path.join(output_folder, txt_name)
        
        print(f"[{idx}/{len(pdf_files)}] 处理: {pdf_name}")
        
        doc = fitz.open(pdf_path)
        all_text = []
        
        for page_num in range(doc.page_count):
            page = doc[page_num]
            text = page.get_text("text")
            all_text.append(f"=== 第{page_num+1}页 ===\n{text}")
        
        doc.close()
        
        with open(txt_path, 'w', encoding='utf-8') as f:
            f.write("\n\n".join(all_text))
        
        print(f"    -> 已保存到: {txt_path}")
    
    print(f"批量导出完成！共处理 {len(pdf_files)} 个PDF")

# 使用示例
batch_extract_to_txt("./pdf_folder", "./output_txt")

import fitz
import os
import openpyxl
from openpyxl.styles import Font, Alignment

def batch_extract_to_excel(input_folder, output_excel="pdf_extract_result.xlsx"):
    """
    批量提取文件夹中所有PDF的文本，汇总到单个Excel文件
    
    Args:
        input_folder: PDF文件所在文件夹
        output_excel: 输出Excel文件名
    """
    # 创建Excel工作簿
    wb = openpyxl.Workbook()
    ws = wb.active
    ws.title = "PDF文本提取结果"
    
    # 设置表头
    headers = ["序号", "文件名", "页数", "文本预览", "完整文本"]
    for col, header in enumerate(headers, 1):
        cell = ws.cell(row=1, column=col, value=header)
        cell.font = Font(bold=True)
        cell.alignment = Alignment(horizontal="center")
    
    pdf_files = [f for f in os.listdir(input_folder) if f.lower().endswith('.pdf')]
    
    print(f"找到 {len(pdf_files)} 个PDF文件")
    
    for idx, pdf_name in enumerate(pdf_files, 1):
        pdf_path = os.path.join(input_folder, pdf_name)
        
        print(f"[{idx}/{len(pdf_files)}] 处理: {pdf_name}")
        
        doc = fitz.open(pdf_path)
        page_count = doc.page_count
        
        # 提取所有页的文本
        full_text = []
        for page_num in range(page_count):
            page = doc[page_num]
            full_text.append(f"【第{page_num+1}页】\n{page.get_text('text')}")
        
        doc.close()
        
        full_text_str = "\n\n".join(full_text)
        text_preview = full_text_str[:200] + "..." if len(full_text_str) > 200 else full_text_str
        
        # 写入Excel行
        ws.cell(row=idx+1, column=1, value=idx)
        ws.cell(row=idx+1, column=2, value=pdf_name)
        ws.cell(row=idx+1, column=3, value=page_count)
        ws.cell(row=idx+1, column=4, value=text_preview)
        ws.cell(row=idx+1, column=5, value=full_text_str)
        
        # 设置自动换行
        for col in range(4, 6):
            ws.cell(row=idx+1, column=col).alignment = Alignment(wrap_text=True)
    
    # 调整列宽
    ws.column_dimensions['A'].width = 8
    ws.column_dimensions['B'].width = 35
    ws.column_dimensions['C'].width = 8
    ws.column_dimensions['D'].width = 60
    ws.column_dimensions['E'].width = 80
    
    output_path = os.path.join(input_folder, output_excel)
    wb.save(output_path)
    print(f"批量导出完成！共处理 {len(pdf_files)} 个PDF，结果保存在 {output_path}")

# 使用示例
batch_extract_to_excel("./pdf_folder", "pdf_text_summary.xlsx")

import fitz
import os
from datetime import datetime

class PDFProcessor:
    """PDF处理器，集成文本提取、图片提取、扫描件检测功能"""
    
    def __init__(self, pdf_path):
        self.pdf_path = pdf_path
        self.doc = None
    
    def open(self):
        self.doc = fitz.open(self.pdf_path)
        return self
    
    def close(self):
        if self.doc:
            self.doc.close()
    
    def get_info(self):
        """获取PDF基本信息"""
        return {
            "filename": os.path.basename(self.pdf_path),
            "page_count": self.doc.page_count,
            "metadata": self.doc.metadata,
            "is_scanned": self._check_is_scanned()
        }
    
    def _check_is_scanned(self):
        """检测是否为扫描件"""
        text_chars = 0
        for page in self.doc:
            text_chars += len(page.get_text("text").strip())
        return text_chars < 100
    
    def extract_text_with_layout(self):
        """提取带布局的文本"""
        result = []
        for page_num, page in enumerate(self.doc):
            blocks = page.get_text("blocks")
            page_text = []
            for block in blocks:
                content = block[4].strip()
                if content:
                    page_text.append(content)
            result.append({
                "page": page_num + 1,
                "text": "\n".join(page_text),
                "block_count": len([b for b in blocks if b[4].strip()])
            })
        return result
    
    def extract_all_images(self, output_dir="images"):
        """提取所有图片"""
        os.makedirs(output_dir, exist_ok=True)
        images = []
        for page_num, page in enumerate(self.doc):
            img_list = page.get_images(full=True)
            for img_idx, img in enumerate(img_list):
                xref = img[0]
                base_img = self.doc.extract_image(xref)
                img_bytes = base_img["image"]
                img_ext = base_img["ext"]
                img_path = os.path.join(output_dir, f"p{page_num+1}_i{img_idx+1}.{img_ext}")
                with open(img_path, "wb") as f:
                    f.write(img_bytes)
                images.append(img_path)
        return images
    
    def process(self):
        """执行完整处理流程"""
        self.open()
        try:
            info = self.get_info()
            text_data = self.extract_text_with_layout()
            images = self.extract_all_images()
            
            return {
                "info": info,
                "text_data": text_data,
                "images": images,
                "total_pages": self.doc.page_count
            }
        finally:
            self.close()


# 使用示例
if __name__ == "__main__":
    processor = PDFProcessor("sample.pdf")
    result = processor.process()
    
    print("=" * 50)
    print(f"文件名: {result['info']['filename']}")
    print(f"总页数: {result['info']['page_count']}")
    print(f"扫描件: {result['info']['is_scanned']}")
    print(f"提取图片数: {len(result['images'])}")
    print("=" * 50)
    
    for page_data in result['text_data']:
        print(f"\n第{page_data['page']}页 (共{page_data['block_count']}个文本块):")
        print(page_data['text'][:300])