diff_indicators_zxh.py 2.7 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778
  1. # -*- coding: utf-8 -*-
  2. """系统比对:报告实际指标行 vs v5 提取器白名单"""
  3. import fitz, re, sys
  4. sys.path.insert(0, r'C:/Users/liaoxg.DESKTOP-DGLN79N/.workbuddy/skills/bun-report-parser/scripts')
  5. import extract_full_report_v5 as v5
  6. doc = fitz.open(r'E:/cfc/docs/参考资料/560822024-张学红.pdf')
  7. # 收集全文行
  8. all_lines = []
  9. for page in doc:
  10. for l in page.get_text().split('\n'):
  11. l = l.strip()
  12. if l:
  13. all_lines.append(l)
  14. # 动态抓取「名称 + 数值(+状态) + 范围」三元组行(不看白名单)
  15. status_words = ['正常', '过多', '轻度产气', '过低', '不足', '缺乏', '偏高', '注意']
  16. found = {} # 名称 -> (值, 状态)
  17. i = 0
  18. cur = None
  19. while i < len(all_lines):
  20. line = all_lines[i]
  21. # 形态A: 名称 数值 状态 范围 (同一行)
  22. m = re.match(r'^(.+?)\s+(\d+)\s+(' + '|'.join(status_words) + r')\s+(\d+-\d+)$', line)
  23. if m:
  24. found[m.group(1)] = (m.group(2), m.group(3), m.group(4))
  25. i += 1
  26. continue
  27. # 形态B: 数值 状态 范围 单独行(名称在上一行,可能多行)
  28. m2 = re.match(r'^(\d+)\s+(' + '|'.join(status_words) + r')\s+(\d+-\d+)$', line)
  29. if m2:
  30. # 向上找名称:跳过页码/空行,取最近的非数值行
  31. name_parts = []
  32. j = i - 1
  33. while j >= 0 and len(name_parts) < 4:
  34. prev = all_lines[j]
  35. if re.match(r'^\d+/169$', prev) or not prev:
  36. j -= 1
  37. continue
  38. if re.match(r'^\d', prev) or prev in status_words:
  39. break
  40. name_parts.insert(0, prev)
  41. # 名称行通常以括号结尾或较短;遇到表头关键词停止
  42. if prev.endswith(')') or prev.endswith(')') or len(prev) < 12:
  43. break
  44. j -= 1
  45. name = ''.join(name_parts)
  46. if name:
  47. found[name] = (m2.group(1), m2.group(2), m2.group(3))
  48. i += 1
  49. continue
  50. i += 1
  51. # 与白名单比对
  52. whitelists = {
  53. 'barrier': set(v5.KNOWN_BARRIER),
  54. 'scfa': set(v5.KNOWN_SCFA),
  55. 'neuro': set(v5.KNOWN_NEUROTRANSMITTER),
  56. }
  57. all_known = set()
  58. for s in whitelists.values():
  59. all_known |= s
  60. print('===== 动态抓到的指标行(白名单外的=新增/漏采)=====')
  61. extra = []
  62. for name, (val, st, rng) in sorted(found.items()):
  63. # 名称归一化后查白名单
  64. hit = any(v5._npar(name).startswith(v5._npar(k)) or v5._npar(k).startswith(v5._npar(name)) for k in all_known)
  65. flag = 'OK ' if hit else 'NEW'
  66. if not hit:
  67. extra.append((name, val, st, rng))
  68. print(f' [{flag}] {name} = {val} {st} ({rng})')
  69. print()
  70. print('===== 白名单外指标(需要补采)=====')
  71. for name, val, st, rng in extra:
  72. print(f' {name}: {val} {st} 范围{rng}')