"""重新生成轻量版OCR预览 — 图片外链""" import sys, os, re, json sys.path.insert(0, os.path.expanduser("~/gov_crawler/venv/lib/python3.9/site-packages")) import pymupdf from rapidocr_onnxruntime import RapidOCR PDF_DIR = os.path.expanduser("~/Nutstore Files/Downloads_nuc") OUTPUT_DIR = "/tmp/cceup_ocr_lite" def html_escape(s): return s.replace('&','&').replace('<','<').replace('>','>').replace('"','"') def process_one(pdf_path, out_dir): fname = os.path.basename(pdf_path).replace('.pdf', '') safe = re.sub(r'[^\w\u4e00-\u9fff]', '_', fname)[:50] doc = pymupdf.open(pdf_path) ocr = RapidOCR() pages_html = [] contact_names = [] contact_phones = [] for i, page in enumerate(doc): pix = page.get_pixmap(dpi=200) img_path = os.path.join(out_dir, f"{safe}_p{i}.png") pix.save(img_path) # OCR result, _ = ocr(img_path) ocr_lines = [] if result: for _, text, score in result: if score > 0.8: ocr_lines.append(text) page_text = '\n'.join(ocr_lines) pages_html.append(f'''

第 {i+1} 页

第{i+1}页
{"
".join(ocr_lines[:60]) if ocr_lines else "(无识别结果)"}
''') # 提取联系信息 phones = re.findall(r'1[3-9]\d{9}', page_text) contact_phones.extend(phones) doc.close() # 联系人归纳 blacklist = {'手机','代表','单位注册地址','为外资','姓名及号码','姓名'} names = re.findall( r'(?:联系人|联系人|项目负责人|法人|法定代表人|代表|业主|经办人|' r'负责人|项目经理|采购人|招标人|姓名)[：:\s]*([\u4e00-\u9fff·]{2,6})', '\n'.join(pages_html) ) names = [n for n in names if n not in blacklist and len(n) >= 2] contacts = '' if names or contact_phones: contacts = f'''

📞 提取到的联系方式

联系人: {"、".join(set(names)) or "未识别"}

电话: {"、".join(set(contact_phones)) or "未识别"}

''' html = f''' 📄 {html_escape(fname)}
← 返回列表

📄 {html_escape(fname)}

{contacts} {"".join(pages_html)}
''' return html, safe # 主流程 os.makedirs(OUTPUT_DIR, exist_ok=True) all_pdfs = sorted([f for f in os.listdir(PDF_DIR) if f.endswith('.pdf')]) samples = [all_pdfs[2], all_pdfs[10], all_pdfs[20], all_pdfs[35], all_pdfs[55]] index_items = [] for fname in samples: fpath = os.path.join(PDF_DIR, fname) print(f"处理: {fname}...") html, safe = process_one(fpath, OUTPUT_DIR) with open(os.path.join(OUTPUT_DIR, f"{safe}.html"), 'w', encoding='utf-8') as f: f.write(html) index_items.append((safe, fname)) print(f" OK") # 索引页 index_html = ''' 📄 CCEUP PDF OCR 预览

📄 CCEUP PDF OCR 预览

5 个样本 — 左侧为PDF截图，右侧为RapidOCR识别结果

' with open(os.path.join(OUTPUT_DIR, "index.html"), 'w', encoding='utf-8') as f: f.write(index_html) print(f"\n完成! 输出: {OUTPUT_DIR}/") print(f"文件列表: {os.listdir(OUTPUT_DIR)}")