feat: update cms case api

test/swagger_cms_case.py

@@ -0,0 +1,515 @@
+# -*- coding: utf-8 -*-
+"""
+CMS 病例库 + AI 病例生成 + 病例审核 — Swagger Try-it-out 等效脚本（真实 HTTP）。
+
+覆盖 CSV 行 80~98（超级管理员/内容管理员/医院管理员 各角色拆分）：
+  CMS-CASE-1~7、CMS-CASE-AI-1、CMS-AUDIT-1~3
+
+- 详细日志：logs/test-swagger-cms-case-YYYY-MM-DD.log
+- 样例 JSON：logs/cms-case-swagger-examples.json（供回填 docx/API - Sheet1.csv）
+- PDF/AI 需本地 Django + Redis + .env 中 DEEPSEEK_API_KEY；无 Key 时加 --skip-ai 跳过
+
+运行：
+  .venv\\Scripts\\python.exe test/swagger_cms_case.py
+  .venv\\Scripts\\python.exe test/swagger_cms_case.py --update-csv
+  .venv\\Scripts\\python.exe test/swagger_cms_case.py --skip-ai --update-csv
+"""
+from __future__ import annotations
+
+import argparse
+import csv
+import io
+import json
+import os
+import subprocess
+import sys
+from datetime import datetime
+from typing import Any
+
+import requests
+
+sys.stdout.reconfigure(encoding='utf-8')
+sys.stderr.reconfigure(encoding='utf-8')
+
+BASE = os.environ.get('SWAGGER_BASE', 'http://127.0.0.1:8000')
+PYTHON = os.path.join(os.path.dirname(os.path.dirname(os.path.abspath(__file__))), '.venv', 'Scripts', 'python.exe')
+if not os.path.isfile(PYTHON):
+    PYTHON = sys.executable
+CWD = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+LOG_DIR = os.path.join(CWD, 'logs')
+CSV_PATH = os.path.join(CWD, 'docx', 'API - Sheet1.csv')
+os.makedirs(LOG_DIR, exist_ok=True)
+
+LOG_FILE = os.path.join(LOG_DIR, f'test-swagger-cms-case-{datetime.now():%Y-%m-%d}.log')
+EXAMPLES_FILE = os.path.join(LOG_DIR, 'cms-case-swagger-examples.json')
+
+SUPER_PHONE = '13700009001'
+CONTENT_PHONE = '13700009002'
+HOSPITAL_PHONE = '13700009004'
+
+_fh = open(LOG_FILE, 'a', encoding='utf-8')
+examples: dict[str, dict] = {}
+results: list[tuple[str, str, int]] = []
+
+
+def w(text=''):
+    line = str(text)
+    _fh.write(line + '\n')
+    _fh.flush()
+    print(line)
+
+
+def django_eval(code: str) -> str:
+    pre = (
+        'import django,os;'
+        'os.environ.setdefault("DJANGO_SETTINGS_MODULE","config.settings");'
+        'django.setup();'
+    )
+    p = subprocess.run(
+        [PYTHON, '-c', pre + code],
+        capture_output=True, text=True, cwd=CWD, encoding='utf-8',
+    )
+    if p.returncode != 0:
+        w(f'[django_eval stderr] {p.stderr}')
+    return (p.stdout or '').strip()
+
+
+def compact_json(obj: Any, max_len: int = 2800) -> str:
+    s = json.dumps(obj, ensure_ascii=False, separators=(',', ':'))
+    if len(s) > max_len:
+        return s[:max_len] + '...(truncated)'
+    return s
+
+
+def case_payload(title: str, institution_id=None, with_exam=False):
+    p = {
+        'title': title,
+        'case_type': 'traditional',
+        'difficulty': 'medium',
+        'chief_complaint': '发热 3 天',
+        'description': '患儿，男，4 岁，因发热 3 天就诊。',
+        'patient_age': 4,
+        'patient_gender': 'male',
+        'department_name': '儿科',
+        'estimated_minutes': 30,
+        'osce_enabled': False,
+        'tags': '儿科,发热',
+        'traditional': {
+            'standard_diagnosis': '上呼吸道感染',
+            'standard_treatment': '对症治疗，退热处理',
+            'guideline_reference': '《儿科学》第 9 版',
+        },
+        'scoring_rules': [
+            {'dimension': '诊断准确性', 'score_weight': 0.5, 'ai_auto_score': True, 'scoring_standard': '能准确诊断'},
+            {'dimension': '医患沟通', 'score_weight': 0.5, 'ai_auto_score': False, 'scoring_standard': '沟通到位'},
+        ],
+    }
+    if institution_id is not None:
+        p['institution_id'] = institution_id
+    if with_exam:
+        p['exam_items'] = [{
+            'item_code': 'blood_routine',
+            'item_name': '血常规',
+            'item_type': 'lab',
+            'category': '实验室检查',
+            'result_text': 'WBC 10×10^9/L',
+            'is_key': True,
+            'is_abnormal': False,
+            'score_weight': 1.0,
+            'display_order': 1,
+        }]
+    return p
+
+
+def shrink_list_resp(resp: dict) -> dict:
+    """列表响应保留 1 条 results 样例，便于 CSV。"""
+    if not isinstance(resp, dict):
+        return resp
+    out = dict(resp)
+    if isinstance(out.get('results'), list) and out['results']:
+        out['results'] = [out['results'][0]]
+    return out
+
+
+def shrink_full_resp(resp: dict) -> dict:
+    """完整病例响应保留关键字段。"""
+    if not isinstance(resp, dict):
+        return resp
+    c = resp.get('case') or {}
+    return {
+        'case': {k: c.get(k) for k in (
+            'id', 'title', 'case_type', 'institution', 'institution_name',
+            'department', 'department_name', 'publish_status', 'created_by_name', 'created_at',
+        )},
+        'traditional': resp.get('traditional'),
+        'scoring_rules': (resp.get('scoring_rules') or [])[:2],
+        'exam_items': (resp.get('exam_items') or [])[:1],
+    }
+
+
+def shrink_ai_resp(resp: dict) -> dict:
+    if not isinstance(resp, dict):
+        return resp
+    data = resp.get('data') or {}
+    return {
+        'parse_id': resp.get('parse_id'),
+        'case_type': resp.get('case_type'),
+        'ai_usage': resp.get('ai_usage'),
+        'prompt_version': resp.get('prompt_version'),
+        'parsing_seconds': resp.get('parsing_seconds'),
+        'generating_seconds': resp.get('generating_seconds'),
+        'data': {k: data.get(k) for k in (
+            'title', 'case_type', 'chief_complaint', 'department_name', 'patient_age', 'patient_gender',
+        ) if k in data},
+    }
+
+
+def call(
+    csv_key: str,
+    name: str,
+    method: str,
+    path: str,
+    token: str | None = None,
+    json_body=None,
+    params=None,
+    files=None,
+    data=None,
+    expect=200,
+    shrink=None,
+):
+    """发起 HTTP 请求，记录日志与 CSV 样例。"""
+    headers = {'Authorization': f'Bearer {token}'} if token else {}
+    url = BASE + path
+    r = requests.request(
+        method, url, headers=headers, json=json_body, params=params,
+        files=files, data=data, timeout=120,
+    )
+    ct = r.headers.get('content-type', '')
+    is_json = ct.startswith('application/json')
+    resp_raw = r.json() if is_json else r.text
+    resp_store = resp_raw
+    if shrink == 'list':
+        resp_store = shrink_list_resp(resp_raw if isinstance(resp_raw, dict) else {})
+    elif shrink == 'full':
+        resp_store = shrink_full_resp(resp_raw if isinstance(resp_raw, dict) else {})
+    elif shrink == 'ai':
+        resp_store = shrink_ai_resp(resp_raw if isinstance(resp_raw, dict) else {})
+    elif shrink == 'create':
+        resp_store = shrink_full_resp(resp_raw if isinstance(resp_raw, dict) else {})
+
+    if json_body is not None:
+        req_example = json.dumps(json_body, ensure_ascii=False)
+    elif files is not None:
+        fn = list(files.values())[0][0] if files else 'file.pdf'
+        extra = f', case_type={data.get("case_type")}' if data else ''
+        req_example = f'multipart/form-data: files={fn}{extra}'
+    elif params:
+        req_example = '?' + '&'.join(f'{k}={v}' for k, v in params.items())
+    elif method == 'GET' and '{' in path:
+        req_example = f'路径参数 + 请求头 Authorization: Bearer <token>'
+    elif method in ('POST',) and json_body is None and files is None:
+        req_example = '路径参数 id + 请求头 Authorization（无 Body）'
+    else:
+        req_example = '请求头 Authorization: Bearer <token>'
+
+    exp_list = expect if isinstance(expect, (list, tuple)) else [expect]
+    ok = 'PASS' if r.status_code in exp_list else 'FAIL'
+    results.append((csv_key, ok, r.status_code))
+    examples[csv_key] = {
+        'name': name,
+        'method': method,
+        'path': path.lstrip('/'),
+        'status': r.status_code,
+        'req': req_example,
+        'resp': resp_store,
+    }
+
+    w(f'\n[{ok}] {csv_key} {name}  ->  {method} {path}  (status={r.status_code})')
+    w(f'  请求: {req_example}')
+    body_str = json.dumps(resp_store, ensure_ascii=False, indent=2) if is_json else str(resp_store)
+    if len(body_str) > 1800:
+        body_str = body_str[:1800] + f' ...(截断, 共{len(body_str)}字符)'
+    w(f'  响应: {body_str}')
+    return r
+
+
+def cleanup():
+    django_eval(
+        'from apps.case.models import CaseBase; from apps.user.models import User; '
+        f'CaseBase.all_objects.filter(created_by__phone__in=["{SUPER_PHONE}","{CONTENT_PHONE}"]).hard_delete(); '
+        f'User.objects.filter(phone__in=["{SUPER_PHONE}","{CONTENT_PHONE}","{HOSPITAL_PHONE}"]).delete(); '
+        'from apps.user.models import Institution; '
+        'Institution.all_objects.filter(code__in=["SWG_CASE_A","SWG_CASE_B"]).hard_delete(); '
+        'print("cleaned")'
+    )
+
+
+def setup_tokens():
+    w('\n[准备] 建临时机构/用户并签发 token ...')
+    setup = django_eval(
+        'from apps.user.models import User, Institution, Department; '
+        'from rest_framework_simplejwt.tokens import RefreshToken; '
+        'iA,_=Institution.objects.get_or_create(code="SWG_CASE_A", defaults={"name":"病例联调A院","type":"hospital"}); '
+        'iB,_=Institution.objects.get_or_create(code="SWG_CASE_B", defaults={"name":"病例联调B院","type":"hospital"}); '
+        'd1,_=Department.objects.get_or_create(name="儿科", defaults={"category":"临床"}); '
+        'd2,_=Department.objects.get_or_create(name="内科", defaults={"category":"临床"}); '
+        f'[User.objects.filter(phone=p).delete() for p in ["{SUPER_PHONE}","{CONTENT_PHONE}","{HOSPITAL_PHONE}"]]; '
+        f'su=User.objects.create_user(username="{SUPER_PHONE}",password=None,phone="{SUPER_PHONE}",'
+        f'real_name="病例超管",role_type="super_admin",institution=iA,status=1); '
+        f'cu=User.objects.create_user(username="{CONTENT_PHONE}",password=None,phone="{CONTENT_PHONE}",'
+        f'real_name="病例内容员",role_type="content_admin",institution=iA,status=1); '
+        f'hu=User.objects.create_user(username="{HOSPITAL_PHONE}",password=None,phone="{HOSPITAL_PHONE}",'
+        f'real_name="病例院管",role_type="hospital_admin",institution=iA,status=1); '
+        'print("|".join([str(RefreshToken.for_user(su).access_token),'
+        'str(RefreshToken.for_user(cu).access_token),str(RefreshToken.for_user(hu).access_token),'
+        'str(d1.id),str(d2.id),str(iA.id),str(iB.id)]))'
+    )
+    su_tok, cu_tok, hu_tok, d1_id, d2_id, iA_id, iB_id = setup.split('|')
+    w(f'[准备] 完成 instA={iA_id} instB={iB_id} dept儿科={d1_id} dept内科={d2_id}')
+    return su_tok, cu_tok, hu_tok, int(d1_id), int(d2_id), int(iA_id), int(iB_id)
+
+
+def run_tests(skip_ai: bool):
+    su_tok, cu_tok, hu_tok, d1_id, d2_id, iA_id, iB_id = setup_tokens()
+
+    w('\n' + '=' * 90)
+    w('  超级管理员 - 病例库')
+    w('=' * 90)
+
+    # 先建病例供后续列表/查看
+    body_super = case_payload('Swagger超管病例(联调)', with_exam=True)
+    r = call('82', 'CMS-CASE-3 表单新增(超管)', 'POST', '/api/cms/cases/',
+             su_tok, json_body=body_super, expect=201, shrink='create')
+    super_case_id = (r.json().get('case') or {}).get('id')
+    r_b = call('_b', 'CMS-CASE-3 指定B院(流程用)', 'POST', '/api/cms/cases/',
+               su_tok, json_body=case_payload('Swagger B院病例', institution_id=iB_id),
+               expect=201, shrink='create')
+    b_case_id = (r_b.json().get('case') or {}).get('id')
+
+    call('80', 'CMS-CASE-1 病例列表(超管)', 'GET', '/api/cms/cases/',
+         su_tok, params={'page': 1}, expect=200, shrink='list')
+
+    pdf_bytes = (
+        b'%PDF-1.4\n1 0 obj<</Type/Catalog/Pages 2 0 R>>endobj\n'
+        b'2 0 obj<</Type/Pages/Kids[3 0 R]/Count 1>>endobj\n'
+        b'3 0 obj<</Type/Page/MediaBox[0 0 612 792]/Parent 2 0 R/Contents 4 0 R>>endobj\n'
+        b'4 0 obj<</Length 44>>stream\nBT /F1 12 Tf 100 700 Td (fever 3d) Tj ET\nendstream\nendobj\n'
+        b'xref\n0 5\ntrailer<</Root 1 0 R>>\nstartxref\n0\n%%EOF'
+    )
+    if not skip_ai:
+        w('\n[AI] PDF 导入 / AI 生成（需 DEEPSEEK_API_KEY）...')
+        call('81', 'CMS-CASE-2 PDF导入(超管)', 'POST', '/api/cms/cases/import-pdf/',
+             su_tok, files={'files': ('case.pdf', pdf_bytes, 'application/pdf')},
+             data={'case_type': 'traditional'}, expect=[200, 429, 502, 504], shrink='ai')
+        call('87', 'CMS-CASE-AI-1 AI生成(超管)', 'POST', '/api/cms/cases/ai-generate/',
+             su_tok, json_body={
+                 'case_type': 'traditional',
+                 'prompt': '请生成一个儿科急性上呼吸道感染传统病例，患儿4岁男孩，发热咳嗽3天。',
+             }, expect=[200, 429, 502, 504], shrink='ai')
+    else:
+        w('\n[跳过] PDF/AI 接口（--skip-ai）')
+
+    call('83', 'CMS-CASE-4 编辑关联(超管)', 'POST', f'/api/cms/cases/{super_case_id}/relations/',
+         su_tok, json_body={'institution_id': iB_id, 'department_id': d2_id}, expect=200)
+
+    # 单独草稿用于 submit
+    r = call('84', 'CMS-CASE-5 提交(超管)', 'POST', f'/api/cms/cases/{super_case_id}/submit/',
+            su_tok, expect=200)
+    # 再建一条用于 disable
+    r2 = call('85', 'CMS-CASE-6 停用(超管)', 'POST', f'/api/cms/cases/{b_case_id}/disable/',
+              su_tok, expect=200)
+
+    call('86', 'CMS-CASE-7 查看(超管)', 'GET', f'/api/cms/cases/{super_case_id}/full/',
+         su_tok, expect=200, shrink='full')
+
+    w('\n' + '=' * 90)
+    w('  内容管理员 - 病例库')
+    w('=' * 90)
+
+    body_content = case_payload('Swagger内容员病例(联调)')
+    r = call('90', 'CMS-CASE-3 表单新增(内容员)', 'POST', '/api/cms/cases/',
+             cu_tok, json_body=body_content, expect=201, shrink='create')
+    content_case_id = (r.json().get('case') or {}).get('id')
+
+    call('88', 'CMS-CASE-1 病例列表(内容员)', 'GET', '/api/cms/cases/',
+         cu_tok, params={'page': 1}, expect=200, shrink='list')
+
+    if not skip_ai:
+        call('89', 'CMS-CASE-2 PDF导入(内容员)', 'POST', '/api/cms/cases/import-pdf/',
+             cu_tok, files={'files': ('case.pdf', pdf_bytes, 'application/pdf')},
+             data={'case_type': 'traditional'}, expect=[200, 429, 502, 504], shrink='ai')
+        call('95', 'CMS-CASE-AI-1 AI生成(内容员)', 'POST', '/api/cms/cases/ai-generate/',
+             cu_tok, json_body={
+                 'case_type': 'traditional',
+                 'prompt': '请生成一个儿科发热传统病例，4岁男孩，发热3天。',
+             }, expect=[200, 429, 502, 504], shrink='ai')
+    else:
+        w('\n[跳过] 内容员 PDF/AI')
+
+    call('91', 'CMS-CASE-4 编辑关联(内容员)', 'POST', f'/api/cms/cases/{content_case_id}/relations/',
+         cu_tok, json_body={'department_id': d2_id}, expect=200)
+
+    call('92', 'CMS-CASE-5 提交(内容员)', 'POST', f'/api/cms/cases/{content_case_id}/submit/',
+         cu_tok, expect=200)
+
+    r_dis = call('_dis', '待停用(流程用)', 'POST', '/api/cms/cases/',
+                 cu_tok, json_body=case_payload('Swagger待停用病例'), expect=201, shrink='create')
+    disable_id = (r_dis.json().get('case') or {}).get('id')
+    call('93', 'CMS-CASE-6 停用(内容员)', 'POST', f'/api/cms/cases/{disable_id}/disable/',
+         cu_tok, expect=200)
+
+    call('94', 'CMS-CASE-7 查看(内容员)', 'GET', f'/api/cms/cases/{content_case_id}/full/',
+         cu_tok, expect=200, shrink='full')
+
+    w('\n' + '=' * 90)
+    w('  医院管理员 - 病例审核')
+    w('=' * 90)
+
+    call('96', 'CMS-AUDIT-1 待审核列表', 'GET', '/api/cms/cases/',
+         hu_tok, params={'publish_status': 1, 'page': 1}, expect=200, shrink='list')
+
+    call('97', 'CMS-AUDIT-2 查看病例', 'GET', f'/api/cms/cases/{content_case_id}/full/',
+         hu_tok, expect=200, shrink='full')
+
+    call('98', 'CMS-AUDIT-3 发布', 'POST', f'/api/cms/cases/{content_case_id}/publish/',
+         hu_tok, expect=200)
+
+
+# 各行「说明」原文（CSV 第 9 列，update 时保持不变）
+ROW_NOTES = {
+    '80': '仅超级管理员。返回全平台病例(可用?institution=按机构过滤)。已软删(下架)不返回。不可发布(归医院管理员 CMS-AUDIT-3)。未登录401；非超管403',
+    '81': '仅超级管理员。解析预览不落库,前端审核后调 CMS-CASE-3 入库时可传 institution_id 指定任意机构。case_type非法→400;非multipart→415;限流429;AI异常502/504;非超管403',
+    '82': '仅超级管理员。落库为草稿(publish_status=0),可指定任意 institution_id。错误：CASE_TYPE_NOT_SUPPORTED/CASE_SUBTYPE_REQUIRED/CASE_SUBTYPE_CONFLICT/CASE_VALIDATION_ERROR/CASE_FIELD_NOT_ALLOWED/CASE_INSTITUTION_NOT_FOUND/CASE_DEPARTMENT_NOT_FOUND/CASE_EXAM_ITEM_DUPLICATE;非超管403',
+    '83': '仅超级管理员。可改全平台任意病例的所属机构/科室,不改病例内容。三者都没传→400 CASE_VALIDATION_ERROR；机构/科室不存在→400；病例不存在/已软删→404;非超管403',
+    '84': '仅超级管理员。全平台任意病例:草稿(0)→正常(1),进入对应机构医院管理员审核队列。非草稿→400 CASE_NOT_SUBMITTABLE；病例不存在/已软删→404;非超管403',
+    '85': '仅超级管理员。全平台任意病例停用=下架=软删除(is_deleted=1,不物删)。下架后默认列表/查看不返回；病例不存在/已软删→404;非超管403',
+    '86': '仅超级管理员。可查看全平台任意病例完整结构。病例不存在/已软删→404;非超管403',
+    '87': '仅超级管理员。DeepSeek按病例模板(prompts/case_{type}_full.md)生成,不落库;前端审核后走 CMS-CASE-3 入库时可指定 institution_id。prompt空→400;限流429;AI异常500/502/504;非超管403',
+    '88': '仅内容管理员。仅返回本院(institution=登录用户所属机构)病例;他院不可见。?institution 参数无效(后端强制本院)。已软删不返回。不可审核发布(归医院管理员)。未登录401；非内容管理员403',
+    '89': '仅内容管理员。解析预览不落库,前端审核后调 CMS-CASE-3 入库时 institution 强制本院(忽略 institution_id)。case_type非法→400;非multipart→415;限流429;AI异常502/504;非内容管理员403',
+    '90': '仅内容管理员。新建强制 institution=本院(传 institution_id 亦忽略)。落库为草稿(publish_status=0)。错误同 CMS-CASE-3;非内容管理员403',
+    '91': '仅内容管理员。仅能操作本院病例;可改科室,机构锁定本院(传 institution_id 无效或不可改他院)。不改病例内容。他院病例→404;非内容管理员403',
+    '92': '仅内容管理员。仅本院病例:草稿(0)→正常(1),进入本院医院管理员审核队列。非草稿→400 CASE_NOT_SUBMITTABLE；他院/不存在/已软删→404;非内容管理员403',
+    '93': '仅内容管理员。仅本院病例停用=下架=软删除(is_deleted=1)。下架后默认列表/查看不返回；他院/不存在/已软删→404;非内容管理员403',
+    '94': '仅内容管理员。仅可查看本院病例完整结构。他院/不存在/已软删→404;非内容管理员403',
+    '95': '仅内容管理员。DeepSeek按病例模板生成,不落库;前端审核后走 CMS-CASE-3 入库时强制落本院。prompt空→400;限流429;AI异常500/502/504;非内容管理员403',
+    '96': '仅医院管理员。本院 publish_status=1(正常/待审核)病例列表,即 CMS-CASE-1 加 publish_status=1 且后端强制本院。超管/内容管理员请用各自病例列表接口。未登录401；非医院管理员403',
+    '97': '仅医院管理员。审核前查看本院待审核病例完整内容。他院/不存在/已软删→404;非医院管理员403',
+    '98': '仅医院管理员(超管/内容管理员均403)。本院病例:正常(1)→已发布(2),发布后对本院移动端医学生可见。非正常→400 CASE_NOT_PUBLISHABLE；他院/不存在/已软删→404',
+}
+
+# 各行 params 字段说明前缀（不含「实际示例」）
+ROW_PARAMS_BASE = {
+    '80': '查询参数(均可选)：search(标题/主诉/标签/ICD)、case_type、publish_status(0草稿/1正常/2已发布)、institution(机构ID，可按任意机构过滤)、department(科室ID)、status、osce_enabled、ordering、page；请求头：Authorization',
+    '81': '请求体(multipart/form-data)：files=1~5份.pdf、case_type=traditional|teaching；请求头：Authorization',
+    '82': '请求体(JSON)：title*、case_type*(traditional|teaching)、institution_id(可选，指定落库机构；缺省落创建者机构)、department_name、子表traditional|teaching*、scoring_rules*(≥1，dimension必填、score_weight∈(0,1])、exam_items(可选，item_code不重复)、difficulty/chief_complaint/description/patient_age/patient_gender/tags/icd_codes/estimated_minutes/osce_enabled…；不接收stages；请求头：Authorization',
+    '83': '路径参数：id；请求体(JSON，至少一项)：institution_id(机构ID，null清空，可改任意机构)、department_id(科室ID，null清空) 或 department_name；请求头：Authorization',
+    '84': '路径参数：id(无Body)；请求头：Authorization',
+    '85': '路径参数：id(无Body)；请求头：Authorization',
+    '86': '路径参数：id；请求头：Authorization',
+    '87': '请求体(JSON)：prompt*(病例长描述)、case_type*(traditional|teaching)；请求头：Authorization',
+    '88': '查询参数(均可选)：search、case_type、publish_status(0草稿/1正常/2已发布)、department(科室ID)、status、osce_enabled、ordering、page；请求头：Authorization',
+    '89': '请求体(multipart/form-data)：files=1~5份.pdf、case_type=traditional|teaching；请求头：Authorization',
+    '90': '请求体(JSON)：title*、case_type*(traditional|teaching)、department_name、子表traditional|teaching*、scoring_rules*(≥1，dimension必填、score_weight∈(0,1])、exam_items(可选，item_code不重复)、difficulty/chief_complaint/description/patient_age/patient_gender/tags/icd_codes/estimated_minutes/osce_enabled…；不接收institution_id/stages；请求头：Authorization',
+    '91': '路径参数：id；请求体(JSON，至少一项)：department_id(科室ID，null清空) 或 department_name；请求头：Authorization',
+    '92': '路径参数：id(无Body)；请求头：Authorization',
+    '93': '路径参数：id(无Body)；请求头：Authorization',
+    '94': '路径参数：id；请求头：Authorization',
+    '95': '请求体(JSON)：prompt*(病例长描述)、case_type*(traditional|teaching)；请求头：Authorization',
+    '96': '查询参数：publish_status=1(正常=待审核)、search、department、page；请求头：Authorization',
+    '97': '路径参数：id；请求头：Authorization',
+    '98': '路径参数：id(无Body)；请求头：Authorization',
+}
+
+
+def update_csv():
+    """用 examples 回填 CSV 第 80~98 行的 params / response 列。"""
+    if not os.path.isfile(EXAMPLES_FILE):
+        w(f'样例文件不存在: {EXAMPLES_FILE}')
+        return False
+    with open(EXAMPLES_FILE, encoding='utf-8') as f:
+        ex = json.load(f)
+
+    rows = []
+    with open(CSV_PATH, encoding='utf-8', newline='') as f:
+        rows = list(csv.reader(f))
+
+    updated = 0
+    for i, row in enumerate(rows):
+        if not row or row[0] not in ex:
+            continue
+        row_num = row[0]
+        e = ex[row_num]
+        method = e.get('method', row[6] if len(row) > 6 else 'GET')
+        base_params = ROW_PARAMS_BASE.get(row_num, row[7] if len(row) > 7 else '')
+        req = e.get('req', '')
+
+        if method == 'GET' and req.startswith('?'):
+            new_params = f'{base_params} ｜ 实际示例：{method} /{e["path"]}{req}'
+        elif req.startswith('multipart'):
+            new_params = f'{base_params} ｜ 实际示例：{req}'
+        elif req.startswith('{') or req.startswith('['):
+            new_params = f'{base_params} ｜ 实际示例：{req}'
+        elif '路径参数' in req or '无 Body' in req or 'Bearer' in req:
+            new_params = f'{base_params} ｜ 实际示例：{method} /{e["path"]}'
+        else:
+            new_params = f'{base_params} ｜ 实际示例：{req}'
+
+        new_resp = f'实际返回：{compact_json(e.get("resp", ""))}'
+        note = ROW_NOTES.get(row_num, row[9] if len(row) > 9 else '')
+
+        rows[i] = row[:7] + [new_params, new_resp, note, '1', row[11] if len(row) > 11 else '0']
+        updated += 1
+        w(f'  CSV 行 #{row_num} 已更新 ({e.get("name")})')
+
+    with open(CSV_PATH, 'w', encoding='utf-8', newline='') as f:
+        csv.writer(f, quoting=csv.QUOTE_MINIMAL).writerows(rows)
+    w(f'\nCSV 更新完成：{updated} 行 -> {CSV_PATH}')
+    return updated > 0
+
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('--skip-ai', action='store_true', help='跳过 PDF/AI（无 DeepSeek Key 时）')
+    parser.add_argument('--update-csv', action='store_true', help='测试后回填 docx/API - Sheet1.csv')
+    parser.add_argument('--csv-only', action='store_true', help='仅从已有 examples JSON 更新 CSV')
+    args = parser.parse_args()
+
+    w('=' * 90)
+    w(f'  CMS 病例 Swagger 测试  {datetime.now():%Y-%m-%d %H:%M:%S}')
+    w(f'  BASE={BASE}')
+    w('=' * 90)
+
+    if args.csv_only:
+        update_csv()
+        _fh.close()
+        return 0
+
+    try:
+        run_tests(skip_ai=args.skip_ai)
+    finally:
+        w('\n[清理] 删除测试病例与用户 ...')
+        cleanup()
+
+    with open(EXAMPLES_FILE, 'w', encoding='utf-8') as f:
+        json.dump(examples, f, ensure_ascii=False, indent=2)
+
+    w('\n' + '=' * 90)
+    total = len(results)
+    passed = sum(1 for _, ok, _ in results if ok == 'PASS')
+    w(f'  总计 {total}  |  通过 {passed}  |  失败 {total - passed}')
+    fails = [(c, s) for c, ok, s in results if ok == 'FAIL']
+    if fails:
+        w('  失败: ' + ', '.join(f'{c}({s})' for c, s in fails))
+    w(f'  日志: {LOG_FILE}')
+    w(f'  样例: {EXAMPLES_FILE}')
+
+    if args.update_csv:
+        update_csv()
+
+    _fh.close()
+    return 0 if not fails else 1
+
+
+if __name__ == '__main__':
+    raise SystemExit(main())