diff_indicators_zxh2.py 2.6 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162636465666768
  1. # -*- coding: utf-8 -*-
  2. """系统比对 v2:适配「名称\\n数值 状态\\n范围」布局"""
  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. all_lines = []
  8. for page in doc:
  9. for l in page.get_text().split('\n'):
  10. l = l.strip()
  11. if l:
  12. all_lines.append(l)
  13. status_words = ['正常', '过多', '轻度产气', '过低', '不足', '缺乏', '偏⾼', '偏低', '注意']
  14. val_re = re.compile(r'^(\d+)\s*(' + '|'.join(status_words) + r')$')
  15. range_re = re.compile(r'^\d+\s*-\s*\d+$')
  16. found = []
  17. i = 0
  18. while i < len(all_lines):
  19. m = val_re.match(all_lines[i])
  20. if m:
  21. val, st = m.group(1), m.group(2)
  22. rng = ''
  23. if i + 1 < len(all_lines) and range_re.match(all_lines[i + 1]):
  24. rng = all_lines[i + 1]
  25. # 向上收集名称(可能 1-3 行)
  26. name_parts = []
  27. j = i - 1
  28. while j >= 0 and len(name_parts) < 3:
  29. prev = all_lines[j]
  30. if re.match(r'^\d+/169$', prev):
  31. j -= 1
  32. continue
  33. if val_re.match(prev) or range_re.match(prev) or prev in status_words:
  34. break
  35. if prev in ('名称', '评估值', '正常范围', '指标范围', '过量 / 缺乏相关症状',
  36. '过量 /缺乏相关症状', '致病菌', '丰度', '评估', '相关症状'):
  37. break
  38. name_parts.insert(0, prev)
  39. if prev.endswith(')') or prev.endswith(')') or len(prev) <= 14:
  40. break
  41. j -= 1
  42. name = ''.join(name_parts)
  43. if name:
  44. found.append((name, val, st, rng))
  45. i += 1
  46. whitelists = {'barrier': v5.KNOWN_BARRIER, 'scfa': v5.KNOWN_SCFA, 'neuro': v5.KNOWN_NEUROTRANSMITTER}
  47. all_known = set()
  48. for s in whitelists.values():
  49. all_known |= set(s)
  50. all_known |= set(v5.KNOWN_AMINO) if hasattr(v5, 'KNOWN_AMINO') else set()
  51. print('===== 全部动态指标行 %d 条 =====' % len(found))
  52. extra = []
  53. for name, val, st, rng in found:
  54. nname = v5._npar(name)
  55. hit = any(nname.startswith(v5._npar(k)) or v5._npar(k).startswith(nname) or v5._npar(k) in nname for k in all_known)
  56. if not hit:
  57. extra.append((name, val, st, rng))
  58. print(f" [{'OK ' if hit else 'NEW'}] {name} = {val} {st} 范围{rng}")
  59. print()
  60. print('===== 白名单外(漏采)%d 条 =====' % len(extra))
  61. for name, val, st, rng in extra:
  62. print(f' >>> {name}: {val} {st} 范围{rng}')