Python实现PDF转 Markdown的完整指南

pip install Spire.PDF

from spire.pdf import *

input_file = "sample.pdf"
output_file = "output.md"

# 加载 PDF 文件
doc = PdfDocument()
doc.LoadFromFile(input_file)

# 保存为 Markdown 格式
doc.SaveToFile(output_file, FileFormat.Markdown)

doc.Close()

from spire.pdf import *

input_file = "report.pdf"
output_file = "chapter2.md"

# 加载原始 PDF
src_pdf = PdfDocument()
src_pdf.LoadFromFile(input_file)

# 创建新的 PDF 文档用于存放选定页面
partial_pdf = PdfDocument()

# 提取第 3–6 页（从 0 开始计数）
partial_pdf.InsertPageRange(src_pdf, 2, 5)

# 保存为 Markdown
partial_pdf.SaveToFile(output_file, FileFormat.Markdown)

src_pdf.Close()
partial_pdf.Close()

from spire.pdf import *

input_file = "invoice.pdf"
output_file = "extracted_tables.md"

# 加载 PDF 文件
pdf = PdfDocument()
pdf.LoadFromFile(input_file)

# 初始化表格提取器
extractor = PdfTableExtractor(pdf)

markdown_tables = []

# 遍历每一页
for page_idx in range(pdf.Pages.Count):
    table_list = extractor.ExtractTable(page_idx)

    if table_list:
        for table in table_list:
            row_count = table.RowCount
            col_count = table.ColumnCount

            for row in range(row_count):
                row_text = []

                for col in range(col_count):
                    cell_value = table.GetText(row, col).replace("\n", " ")
                    row_text.append(cell_value)

                # 转换为 Markdown 表格行
                markdown_tables.append("| " + " | ".join(row_text) + " |")

                # 添加表头分隔线
                if row == 0:
                    markdown_tables.append(
                        "| " + " | ".join(["---"] * col_count) + " |"
                    )

            # 表格之间添加空行
            markdown_tables.append("\n")

# 写入文件
with open(output_file, "w", encoding="utf-8") as f:
    f.write("\n".join(markdown_tables))

pdf.Close()

from spire.pdf import *
from spire.ocr import *

def convert_pdf_page_to_image(pdf, page_index):
    """将 PDF 指定页面转换为图片"""
    return pdf.SaveAsImage(page_index)

def recognize_text_from_image(image_path, language, model_path):
    """使用 OCR 识别图片文字"""
    scanner = OcrScanner()

    configure_options = ConfigureOptions()
    configure_options.Language = language
    configure_options.ModelPath = model_path

    scanner.ConfigureDependencies(configure_options)
    scanner.Scan(image_path)

    visualTextAligner = VisualTextAligner(scanner.Text)
    return visualTextAligner.ToString()

input_file = "ScannedDocument.pdf"
output_file = "ScannedPDF.md"

model_path = r"D:\Download\win-x64"

pdf = PdfDocument()
pdf.LoadFromFile(input_file)

with open(output_file, "w", encoding="utf-8") as writer:
    writer.write("# 扫描 PDF 的 OCR 结果\n\n")

    for page_index in range(pdf.Pages.Count):
        image = convert_pdf_page_to_image(pdf, page_index)

        image_path = f"page_{page_index + 1}.png"
        image.Save(image_path)

        recognized_text = recognize_text_from_image(
            image_path,
            "Chinese",
            model_path
        )

        writer.write(f"## 第 {page_index + 1} 页\n\n")
        writer.write(recognized_text)
        writer.write("\n\n")

pdf.Close()

print(f'扫描版 PDF 已成功转换为 "{output_file}"')

recognized_text = recognize_text_from_image(
    image_path,
    "English",
    model_path
)