pipeline-llm/scripts/test_llm_call_trace_io.py

#!/usr/bin/env python3
"""llm_call_trace_io 行为验证（stub harness，不依赖真实 DB/宿主）。

覆盖场景：
  1. 无 id / 超长 id → Message 提示
  2. 查无记录 → Message 提示
  3. 机构隔离：普通机构看别家行 → 拒绝；平台 org '0' → 放行；本机构 → 放行
  4. 正常行 → TabPanel 三 tab（summary/req/resp），CodeEditor readonly，摘要含
     usage 上下文 + 同批次行
  5. 路径穿越：req_path='../../etc/passwd' → [非法路径，拒绝访问]
  6. 文件缺失 → 提示无原文文件
"""
import asyncio
import json
import os
import sys
import tempfile

sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))

# ---- stub: 假 filesroot + 真实 trace 文件 ----
tmp = tempfile.mkdtemp()
FILESROOT = tmp
trace_rel = 'llm_trace/20260911/callA_000'
os.makedirs(os.path.join(FILESROOT, 'llm_trace/20260911'), exist_ok=True)
with open(os.path.join(FILESROOT, trace_rel + '.req.json'), 'w') as f:
    json.dump({'url': 'https://up/v1/chat', 'method': 'POST', 'kind': 'chat',
               'headers': {'Authorization': '[REDACTED]'},
               'body': {'model': 'm1', 'messages': [{'role': 'user', 'content': 'hi <b>'}]}}, f)
with open(os.path.join(FILESROOT, trace_rel + '.resp.json'), 'w') as f:
    json.dump({'status_code': 200, 'elapsed_ms': 123, 'error': '',
               'body': {'choices': [{'message': {'content': 'hello'}}]}}, f)

# ---- stub: appPublic.jsonConfig.getConfig 返回假 filesroot ----
import types
fake_appPublic = types.ModuleType('appPublic')
fake_jc = types.ModuleType('appPublic.jsonConfig')
class _Cfg:
    filesroot = FILESROOT
fake_jc.getConfig = lambda *a, **k: _Cfg()
sys.modules['appPublic.jsonConfig'] = fake_jc

# ---- stub: DBPools / sqlor ----
TRACE_ROW = {
    'id': 'tid001', 'call_id': 'callA', 'seq': 0, 'kind': 'chat',
    'url': 'https://up/v1/chat/completions', 'method': 'POST',
    'status_code': 200, 'elapsed_ms': 123,
    'req_path': trace_rel + '.req.json', 'resp_path': trace_rel + '.resp.json',
    'org_id': 'org9', 'note': '', 'created_at': '2026-09-11 10:00:00',
}
BATCH_ROWS = [TRACE_ROW, {**TRACE_ROW, 'id': 'tid002', 'seq': 1, 'kind': 'query',
                           'status_code': 200, 'note': ''}]
USAGE_ROWS = [{'status': 'SUCCEEDED', 'accounting_status': 'accounted', 'cost': '0.01',
               'charge': '0.02', 'task_ref': 'sess1', 'project_id': '0', 'model_name': 'qwen-x'}]

class FakeSor:
    def __init__(self, db): self.db = db
    async def sqlExe(self, sql, ns):
        self.db.calls.append((sql, ns))
        if 'FROM llm_call_trace WHERE id=' in sql:
            return [dict(TRACE_ROW)] if not self.db.no_row else []
        if 'FROM llm_call_trace WHERE call_id=' in sql:
            return [dict(r) for r in BATCH_ROWS]
        if 'FROM llm_usage u LEFT JOIN llm_model' in sql:
            return [dict(r) for r in USAGE_ROWS]
        return []
    async def __aenter__(self): return self
    async def __aexit__(self, *a): return False

class FakeDB:
    def __init__(self): self.databases = {}; self.no_row = False; self.calls = []
    def sqlorContext(self, dbname): return FakeSor(self)

_FAKE_DB = FakeDB()
import sqlor.dbpools as dbpools
dbpools.DBPools = lambda *a, **k: _FAKE_DB

# ---- stub: ahserver.serverenv.ServerEnv ----
class FakeEnv(dict):
    _inst = None
    def __new__(cls):
        if FakeEnv._inst is None: FakeEnv._inst = super().__new__(cls)
        return FakeEnv._inst
    def __getattr__(self, k):
        try: return self[k]
        except KeyError: raise AttributeError(k)
import ahserver.serverenv as serverenv
serverenv.ServerEnv = FakeEnv

# ---- stub: gateway._get_db（trace.py/accounting 用）----
import pipeline_llm.gateway as gw
gw._get_db = lambda: (_FAKE_DB, 'pipeline')

import pipeline_llm.init as m

FAILS = []

def check(name, cond, detail=''):
    print(('PASS' if cond else 'FAIL'), name, detail if not cond else '')
    if not cond: FAILS.append(name)

def find_widget(w, wtype):
    if isinstance(w, dict):
        if w.get('widgettype') == wtype: return w
        for v in w.values():
            r = find_widget(v, wtype)
            if r: return r
    elif isinstance(w, list):
        for v in w:
            r = find_widget(v, wtype)
            if r: return r
    return None

async def main():
    # 1. 缺 id / 超长 id
    r = await m.llm_call_trace_io({}, '0')
    check('缺id → Message', r['widgettype'] == 'Message')
    r = await m.llm_call_trace_io({'id': 'x' * 40}, '0')
    check('超长id → Message', r['widgettype'] == 'Message')

    # 2. 查无记录
    _FAKE_DB.no_row = True
    r = await m.llm_call_trace_io({'id': 'tid001'}, '0')
    check('查无 → Message', r['widgettype'] == 'Message')
    _FAKE_DB.no_row = False

    # 3. 机构隔离
    r = await m.llm_call_trace_io({'id': 'tid001'}, 'other_org')
    check('他机构 → 拒绝', r['widgettype'] == 'Message' and '无权' in r['options']['message'])
    r = await m.llm_call_trace_io({'id': 'tid001'}, '0')
    check("平台org'0' → 放行", r['widgettype'] == 'TabPanel')
    r = await m.llm_call_trace_io({'id': 'tid001'}, 'org9')
    check('本机构 → 放行', r['widgettype'] == 'TabPanel')

    # 4. TabPanel 结构
    items = r['options']['items']
    check('3个tab', len(items) == 3)
    check('tab名', [t['name'] for t in items] == ['summary', 'req', 'resp'])
    ce_req = items[1]['content']
    check('req是CodeEditor', ce_req['widgettype'] == 'CodeEditor')
    check('readonly', ce_req['options']['readonly'] is True)
    check('mode null', ce_req['options']['mode'] == 'null')
    val = json.loads(ce_req['options']['value'])
    check('req原文含REDACTED头', val['headers']['Authorization'] == '[REDACTED]')
    check('req原文含body', val['body']['messages'][0]['content'] == 'hi <b>')
    resp_val = json.loads(items[2]['content']['options']['value'])
    check('resp原文body', resp_val['body']['choices'][0]['message']['content'] == 'hello')
    summary_html = find_widget(items[0]['content'], 'Html')['options']['html']
    check('摘要含usage模型名', 'qwen-x' in summary_html)
    check('摘要含批次往返数', '同批次往返（2 次）' in summary_html)
    # HTML 转义：摘要走 Html widget，正文里任何 < > 必须被 escape（防注入）
    TRACE_ROW['note'] = '<script>alert(1)</script>'
    r2 = await m.llm_call_trace_io({'id': 'tid001'}, '0')
    s2 = find_widget(r2['options']['items'][0]['content'], 'Html')['options']['html']
    check('note<script>被转义', '<script>' not in s2 and '&lt;script&gt;' in s2)
    TRACE_ROW['note'] = ''
    # CodeEditor value 是 JSON 字符串（非 HTML），req body 里的 <b> 原样保留不转义
    check('req原文<b>原样', 'hi <b>' in items[1]['content']['options']['value'])

    # 5. 路径穿越
    bad = dict(TRACE_ROW)
    bad['req_path'] = '../../etc/passwd'
    txt, err = m._read_trace_file('../../etc/passwd')
    check('穿越拒绝', err == '[非法路径，拒绝访问]' and txt == '')
    txt, err = m._read_trace_file('llm_trace/20260911/../../etc/passwd')
    check('穿越拒绝2(中间..)', err == '[非法路径，拒绝访问]')

    # 6. 文件缺失
    txt, err = m._read_trace_file('llm_trace/20260911/nonexist.req.json')
    check('缺文件报错', '文件读取失败' in err)
    txt, err = m._read_trace_file('')
    check('空路径提示', '无原文文件' in err)

    print()
    if FAILS:
        print('FAILED:', FAILS); sys.exit(1)
    print('ALL PASS')

asyncio.run(main())