Python fitz包

pip install pymupdf

import fitz

# 打开 PDF 文件
pdf_path = "example.pdf"
pdf_document = fitz.open(pdf_path)

# 获取页面数量
num_pages = pdf_document.page_count
print(f"PDF 文件共有 {num_pages} 页")

# 提取第一页的文本
page = pdf_document[0]
text = page.get_text("text")
print(text)

# 提取第一页的图像
image_list = page.get_images(full=True)
for image_index, image in enumerate(image_list):
    xref = image[0]
    base_image = pdf_document.extract_image(xref)
    image_bytes = base_image["image"]
    image_extension = base_image["ext"]

    with open(f"image_{image_index}.{image_extension}", "wb") as image_file:
        image_file.write(image_bytes)

# 添加文本到第一页
page.insert_text((100, 100), "Hello, geek-docs.com", fontsize=12, color=(1, 0, 0))

# 添加图像到第一页
image_path = "image.jpg"
image_rect = fitz.Rect(100, 100, 200, 200)
page.insert_image(image_rect, filename=image_path)

# 保存修改后的 PDF 文件
pdf_document.save("modified_example.pdf")

# 关闭 PDF 文件
pdf_document.close()