PDF转Markdown，TextIn文档解析工具全体验，兼容多种格式！（附独家免费权益）

pip install TextInParseX

pip3 install TextInParseX -i http://mirrors.aliyun.com/pypi/simple/ --trusted-host=mirrors.aliyun.com

import TextInParseX as px# 初始化解析器app_id = "#############################"     #填入你的textin的api_id和secret——codesecret_code = "#############################"parseX_client = px.ParseXClient(app_id, secret_code)pdf_file_path = "example.pdf" #你的本地文件路径#通过ParseX直接调用url获取解析对象result = parseX_client.begin_analyze_document_from_url(pdf_file_path)

import TextInParseX as pximport jsonjson_file = 'test_json/example.json'with open(json_file, 'r') as fr:    json_result = json.load(fr)parseX_client = px.ParseXClient()result = parseX_client.begin_analyze_document_from_json(json_result)#或者直接输入json文件result = parseX_client.begin_analyze_document_from_file(json_file)

print('Markdown:')print(result.all_markdown)print("\n")print("All text in document:")#为可视化方便, 输出0-1000个字符parseX_client.print_all_elements(result.all_text, 0, 1000)print("\n")print(f"Total tables in document: {len(result.all_tables)}")for index, table in enumerate(result.all_tables):    print(f"Table {index}:")    parseX_client.print_all_elements(table)    print("\n")print(f"Total paragraphs in document: {len(result.all_paragraphs)}")for p_idx, each_paragraph in enumerate(result.all_paragraphs):    print(f"\n--- Paragraph {p_idx}/{len(result.all_paragraphs)} ---")    print(f"Paragraph position: {each_paragraph.pos}")for l_idx, each_line in enumerate(each_paragraph.lines):        print(f"  Line {l_idx}/{len(each_paragraph.lines)}")        print(f"    Line positions: {each_line.pos}")        print(f"  Line text: {each_line.text}")print(f"Total images in document: {len(result.all_images)}")for index, image in enumerate(result.all_images):    print(f"Image {index}:")    parseX_client.print_all_elements(image)    print("\n")all_images_cv_mat = result.get_all_images_cv_mat()print(f"Total images (as cv::Mat) in document: {len(all_images_cv_mat)}")for idx, mat in enumerate(all_images_cv_mat):    print(f"Image {idx} (cv::Mat) shape: {mat.shape}")

#页的索引指向pdf和文档的页，按照页数的规则，从1开始；table等版面元素的索引默认程序读取的规则，从0开始for page in result.pages:    print(f"=== Page {page.page_id} ===")    print("\n")        for index, table in enumerate(page.tables):        print(f"Table {index}:")        parseX_client.print_all_elements(table)        print("\n")    for index, image in enumerate(page.images):        print(f"Image {index}:")        parseX_client.print_all_elements(image)         print("\n")        images_cv_mat = page.get_images_cv_mat()    print(f"Total images (as cv::Mat) in page {page.page_id}: {len(images_cv_mat)}")    for idx, mat in enumerate(images_cv_mat):        print(f"Image {idx} (cv::Mat) shape: {mat.shape}")    print("\n")       print("Text:")    # 限定只能打印前1000个字符    parseX_client.print_all_elements(page.paragraph_text, 0, 1000)      print("\n")        # 获取当前页的段落    print(f"Total paragraphs: {len(page.paragraphs)}")    for p_idx, each_paragraph in enumerate(page.paragraphs):        print(f"\n--- Paragraph {p_idx}/{len(page.paragraphs)} ---")        print(f"Paragraph position: {each_paragraph.pos}")        for l_idx, each_line in enumerate(each_paragraph.lines):            print(f"  Line {l_idx}/{len(each_paragraph.lines)}")            print(f"    Line positions: {each_line.pos}")            print(f"  Line text: {each_line.text}")    print('Finished getting paragraphs')    print("\n\n")

# 获取当前页的段落    print(f"Total paragraphs: {len(page.paragraphs)}")for p_idx, each_paragraph in enumerate(page.paragraphs):        print(f"\n--- Paragraph {p_idx}/{len(page.paragraphs)} ---")        print(f"Paragraph position: {each_paragraph.pos}")for l_idx, each_line in enumerate(each_paragraph.lines):            print(f"  Line {l_idx}/{len(each_paragraph.lines)}")            print(f"    Line positions: {each_line.pos}")            print(f"  Line text: {each_line.text}")    print('Finished getting paragraphs')    print("\n\n")