refactor: 步骤5简化为数据生成+格式检查
- 步骤5不再提交到ParamHub,只生成产品数据 - 智能体任务模板去掉提交部分,增加格式检查要求 - 新增本地格式验证方法_validate_product_data - 验证字段类型、必填项、数值范围等 - 步骤6恢复为独立的提交审核步骤 - 更新步骤描述为'调用智能体生成产品数据并检查格式'
This commit is contained in:
@@ -17,9 +17,9 @@
|
||||
首先,请访问 ParamHub API 文档获取对应类别的字段定义:
|
||||
- API文档地址:http://192.168.2.8:12007/hz4th_coder/param-hub-python/src/branch/master/API.md
|
||||
- 根据产品类别({{category}})确定应该使用哪个API:
|
||||
- AI模型 → `/api/models`
|
||||
- GPU → `/api/gpus`
|
||||
- CPU → `/api/cpus`
|
||||
- AI模型 → `/api/models`,字段包括:name, organization, parameters, context_length, mmlu, publish_date, visible, is_pinned
|
||||
- GPU → `/api/gpus`,字段包括:name, manufacturer, memory_gb, cuda_cores, tensor_cores, price_usd, release_year, visible, is_pinned
|
||||
- CPU → `/api/cpus`,字段包括:name, manufacturer, cores, threads, base_clock, boost_clock, price_usd, visible, is_pinned
|
||||
- 其他动态分类 → `/api/items/{category_id}`
|
||||
|
||||
### 2. 从内容库获取数据内容
|
||||
@@ -28,39 +28,32 @@
|
||||
### 3. 整理产品参数
|
||||
根据获取到的内容,提取并整理产品的各项参数,严格按照API文档中定义的字段格式填充。
|
||||
|
||||
### 4. 提交到审核系统
|
||||
使用以下命令提交到ParamHub审核系统:
|
||||
|
||||
```bash
|
||||
# 先登录获取cookie
|
||||
curl -c /tmp/paramhub_cookie.txt -X POST "http://localhost:16041/login" \
|
||||
-H "Content-Type: application/json" \
|
||||
-d '{"password": "admin123"}'
|
||||
|
||||
# 然后提交产品数据(根据类别选择对应的API)
|
||||
curl -b /tmp/paramhub_cookie.txt -X POST "http://localhost:16041/api/{对应类别API}" \
|
||||
-H "Content-Type: application/json" \
|
||||
-d '{
|
||||
"name": "{{product_name}}",
|
||||
"visible": true,
|
||||
"is_pinned": false,
|
||||
// ... 其他字段根据API文档填充
|
||||
}'
|
||||
```
|
||||
### 4. 格式检查
|
||||
对生成的数据进行以下检查:
|
||||
- 必填字段是否齐全(name必须有值)
|
||||
- 字段类型是否正确(数字字段不能是字符串,布尔字段必须是true/false)
|
||||
- 字段值是否合理(如参数量应为正数,价格应为正数等)
|
||||
- 如果发现格式问题,请修正后重新输出
|
||||
|
||||
### 5. 输出要求
|
||||
请以JSON格式输出执行结果:
|
||||
请以JSON格式输出最终的产品数据(不要提交,只输出数据):
|
||||
```json
|
||||
{
|
||||
"success": true,
|
||||
"review_id": "审核ID(如果提交成功)",
|
||||
"submitted_data": {
|
||||
"product_data": {
|
||||
"name": "产品名称",
|
||||
"field1": "值1",
|
||||
"field2": "值2"
|
||||
"field2": "值2",
|
||||
"visible": true,
|
||||
"is_pinned": false
|
||||
},
|
||||
"data_sources": [数据ID列表],
|
||||
"message": "执行说明"
|
||||
"format_check": {
|
||||
"passed": true,
|
||||
"issues_found": [],
|
||||
"issues_fixed": []
|
||||
},
|
||||
"message": "数据生成说明"
|
||||
}
|
||||
```
|
||||
|
||||
@@ -68,4 +61,5 @@ curl -b /tmp/paramhub_cookie.txt -X POST "http://localhost:16041/api/{对应类
|
||||
- 严格按照API文档的字段定义填充数据
|
||||
- 不要编造或推测任何参数,只使用内容库中实际存在的信息
|
||||
- 如果某些字段无法从内容中提取,可以留空或填写默认值
|
||||
- 提交成功后记录返回的review_id
|
||||
- 不要执行任何提交操作,只生成并输出数据
|
||||
- 确保输出的JSON格式正确,可以被程序解析
|
||||
+104
-55
@@ -21,7 +21,7 @@ PROCESS_STEPS = [
|
||||
{'num': 2, 'name': '搜索互联网', 'description': '从互联网搜索最新数据'},
|
||||
{'num': 3, 'name': '抓取网页内容', 'description': '抓取搜索结果网页的详细内容'},
|
||||
{'num': 4, 'name': '提取产品数据(智能体)', 'description': '调用hz4th_editor智能体提取产品相关内容'},
|
||||
{'num': 5, 'name': '填充字段(智能体)', 'description': '调用hz4th_editor智能体整理产品数据并提交审核'},
|
||||
{'num': 5, 'name': '填充字段(智能体)', 'description': '调用智能体生成产品数据并检查格式'},
|
||||
{'num': 6, 'name': '提交审核', 'description': '提交到ParamHub待审核区'},
|
||||
]
|
||||
|
||||
@@ -227,9 +227,9 @@ class ProcessMonitor:
|
||||
except Exception as e:
|
||||
self._fail_step(session_id, 4, str(e))
|
||||
|
||||
# 步骤5: 填充字段并提交审核(调用智能体执行)
|
||||
# 步骤5: 填充字段(调用智能体生成数据并检查格式)
|
||||
if not self._check_pause(session_id) and all_data['extracted_data']:
|
||||
self._start_step(session_id, product_name, 5, '填充字段并提交(智能体)')
|
||||
self._start_step(session_id, product_name, 5, '填充字段(智能体)')
|
||||
try:
|
||||
# 构建任务文本
|
||||
fill_task_text = self._build_fill_fields_task(
|
||||
@@ -243,60 +243,42 @@ class ProcessMonitor:
|
||||
fill_parsed = self._parse_fill_agent_response(fill_agent_result.get('output', ''))
|
||||
|
||||
if fill_parsed and fill_parsed.get('success'):
|
||||
review_id = fill_parsed.get('review_id')
|
||||
submitted_data = fill_parsed.get('submitted_data', {})
|
||||
product_data = fill_parsed.get('product_data', {})
|
||||
format_check = fill_parsed.get('format_check', {})
|
||||
|
||||
all_data['filled_data'] = submitted_data
|
||||
# 本地格式验证
|
||||
validation_result = self._validate_product_data(product_data, category)
|
||||
|
||||
self._complete_step(session_id, 5, {
|
||||
'filled': True,
|
||||
'agent': 'hz4th_editor',
|
||||
'task_text': fill_task_text,
|
||||
'review_id': review_id,
|
||||
'submitted_data': submitted_data,
|
||||
'agent_output': fill_agent_result.get('output', '')[:2000]
|
||||
})
|
||||
|
||||
# 如果智能体返回了review_id,直接标记步骤6完成
|
||||
if review_id:
|
||||
result['success'] = True
|
||||
result['review_id'] = review_id
|
||||
if validation_result.get('valid'):
|
||||
all_data['filled_data'] = product_data
|
||||
|
||||
db.update_session_status(session_id, 'completed',
|
||||
review_id=review_id,
|
||||
result=json.dumps(result, ensure_ascii=False))
|
||||
|
||||
db.add_process_history(
|
||||
product_name=product_name,
|
||||
category=category,
|
||||
subcategory=subcategory,
|
||||
status='submitted',
|
||||
review_id=review_id,
|
||||
details=all_data
|
||||
)
|
||||
logger.info(f"[{session_id}] 步骤5完成: 智能体提交成功, review_id={review_id}")
|
||||
else:
|
||||
self._complete_step(session_id, 5, {
|
||||
'filled': True,
|
||||
'agent': 'hz4th_editor',
|
||||
'task_text': fill_task_text,
|
||||
'submitted_data': submitted_data,
|
||||
'message': '数据已整理但未获取到review_id',
|
||||
'product_data': product_data,
|
||||
'format_check': format_check,
|
||||
'validation': validation_result,
|
||||
'agent_output': fill_agent_result.get('output', '')[:2000]
|
||||
})
|
||||
result['message'] = '数据已整理但提交状态未知'
|
||||
logger.info(f"[{session_id}] 步骤5完成: 数据生成成功,格式验证通过")
|
||||
else:
|
||||
# 格式验证失败,记录问题
|
||||
self._fail_step(session_id, 5, f"数据格式验证失败: {validation_result.get('errors', [])}")
|
||||
result['message'] = '数据格式验证失败'
|
||||
else:
|
||||
self._fail_step(session_id, 5, f"智能体执行失败: {fill_parsed.get('message', '未知错误') if fill_parsed else '解析失败'}")
|
||||
result['message'] = f'智能体执行失败'
|
||||
error_msg = fill_parsed.get('message', '未知错误') if fill_parsed else '解析失败'
|
||||
self._fail_step(session_id, 5, f"智能体执行失败: {error_msg}")
|
||||
result['message'] = f'智能体执行失败: {error_msg}'
|
||||
else:
|
||||
self._fail_step(session_id, 5, f"智能体调用失败: {fill_agent_result.get('error', '未知错误')}")
|
||||
result['message'] = f'智能体调用失败: {fill_agent_result.get("error")}'
|
||||
except Exception as e:
|
||||
self._fail_step(session_id, 5, str(e))
|
||||
|
||||
# 步骤6: 确认提交结果(如果步骤5未获取到review_id,尝试本地提交)
|
||||
if not self._check_pause(session_id) and all_data['filled_data'] and not result.get('review_id'):
|
||||
self._start_step(session_id, product_name, 6, '确认提交结果')
|
||||
# 步骤6: 提交审核
|
||||
if not self._check_pause(session_id) and all_data['filled_data']:
|
||||
self._start_step(session_id, product_name, 6, '提交审核')
|
||||
try:
|
||||
category_type = self._get_category_type(category)
|
||||
success, review_id_or_error = paramhub_client.submit_for_review(
|
||||
@@ -322,7 +304,7 @@ class ProcessMonitor:
|
||||
review_id=review_id_or_error,
|
||||
details=all_data
|
||||
)
|
||||
logger.info(f"[{session_id}] 步骤6完成: 本地提交成功")
|
||||
logger.info(f"[{session_id}] 步骤6完成: 提交成功")
|
||||
else:
|
||||
self._fail_step(session_id, 6, review_id_or_error)
|
||||
db.update_session_status(session_id, 'failed')
|
||||
@@ -616,26 +598,93 @@ class ProcessMonitor:
|
||||
if parsed_data:
|
||||
return {
|
||||
'success': parsed_data.get('success', False),
|
||||
'review_id': parsed_data.get('review_id'),
|
||||
'submitted_data': parsed_data.get('submitted_data', {}),
|
||||
'product_data': parsed_data.get('product_data', {}),
|
||||
'data_sources': parsed_data.get('data_sources', []),
|
||||
'format_check': parsed_data.get('format_check', {}),
|
||||
'message': parsed_data.get('message', ''),
|
||||
'raw_output': output
|
||||
}
|
||||
|
||||
# 尝试从文本中提取review_id
|
||||
review_match = re.search(r'review[_-]?id[\s:]*([\w-]+)', output, re.I)
|
||||
if review_match:
|
||||
return {
|
||||
'success': True,
|
||||
'review_id': review_match.group(1),
|
||||
'submitted_data': {},
|
||||
'message': '从输出中提取到review_id',
|
||||
'raw_output': output
|
||||
}
|
||||
|
||||
return None
|
||||
|
||||
def _validate_product_data(self, product_data, category):
|
||||
"""本地验证产品数据格式"""
|
||||
errors = []
|
||||
warnings = []
|
||||
|
||||
if not product_data:
|
||||
return {'valid': False, 'errors': ['数据为空'], 'warnings': []}
|
||||
|
||||
# 检查必填字段
|
||||
if not product_data.get('name'):
|
||||
errors.append('缺少必填字段: name')
|
||||
|
||||
# 检查字段类型
|
||||
category_type = self._get_category_type(category)
|
||||
|
||||
if category_type == 'model':
|
||||
# AI模型字段验证
|
||||
if 'parameters' in product_data and product_data['parameters']:
|
||||
params = product_data['parameters']
|
||||
if not isinstance(params, str) or not params.endswith('B'):
|
||||
warnings.append('parameters应为字符串格式如"70B"')
|
||||
|
||||
if 'context_length' in product_data and product_data['context_length']:
|
||||
ctx = product_data['context_length']
|
||||
if not isinstance(ctx, int) or ctx <= 0:
|
||||
errors.append('context_length应为正整数')
|
||||
|
||||
if 'mmlu' in product_data and product_data['mmlu']:
|
||||
mmlu = product_data['mmlu']
|
||||
if not isinstance(mmlu, (int, float)) or mmlu < 0 or mmlu > 100:
|
||||
warnings.append('mmlu应为0-100之间的数值')
|
||||
|
||||
elif category_type == 'gpu':
|
||||
# GPU字段验证
|
||||
if 'memory_gb' in product_data and product_data['memory_gb']:
|
||||
mem = product_data['memory_gb']
|
||||
if not isinstance(mem, (int, float)) or mem <= 0:
|
||||
errors.append('memory_gb应为正数')
|
||||
|
||||
if 'cuda_cores' in product_data and product_data['cuda_cores']:
|
||||
cores = product_data['cuda_cores']
|
||||
if not isinstance(cores, int) or cores <= 0:
|
||||
errors.append('cuda_cores应为正整数')
|
||||
|
||||
if 'price_usd' in product_data and product_data['price_usd']:
|
||||
price = product_data['price_usd']
|
||||
if not isinstance(price, (int, float)) or price <= 0:
|
||||
warnings.append('price_usd应为正数')
|
||||
|
||||
elif category_type == 'cpu':
|
||||
# CPU字段验证
|
||||
if 'cores' in product_data and product_data['cores']:
|
||||
cores = product_data['cores']
|
||||
if not isinstance(cores, int) or cores <= 0:
|
||||
errors.append('cores应为正整数')
|
||||
|
||||
if 'threads' in product_data and product_data['threads']:
|
||||
threads = product_data['threads']
|
||||
if not isinstance(threads, int) or threads <= 0:
|
||||
errors.append('threads应为正整数')
|
||||
|
||||
if 'base_clock' in product_data and product_data['base_clock']:
|
||||
clock = product_data['base_clock']
|
||||
if not isinstance(clock, (int, float)) or clock <= 0:
|
||||
errors.append('base_clock应为正数')
|
||||
|
||||
# 检查布尔字段
|
||||
for bool_field in ['visible', 'is_pinned']:
|
||||
if bool_field in product_data:
|
||||
if not isinstance(product_data[bool_field], bool):
|
||||
warnings.append(f'{bool_field}应为布尔值')
|
||||
|
||||
return {
|
||||
'valid': len(errors) == 0,
|
||||
'errors': errors,
|
||||
'warnings': warnings
|
||||
}
|
||||
|
||||
def _extract_data(self, product_name, all_data):
|
||||
"""提取产品数据(备用,已被智能体替代)"""
|
||||
all_content = []
|
||||
|
||||
Reference in New Issue
Block a user