A more comprehensive large model analysis setup, front-end interface optimization.

This commit is contained in:
戒酒的李白
2025-02-11 22:30:52 +08:00
parent 8ec05efe27
commit a60a0f3296
3 changed files with 357 additions and 51 deletions
+89 -35
View File
@@ -14,9 +14,26 @@ class AIAnalyzer:
openai.api_key = self.api_key
# 系统提示词,限制AI的输出格式
self.system_prompt = """你是一个专业的舆情分析助手。你的任务是分析每条消息的情感倾向、关键词和潜在影响。
请严格按照以下JSON格式返回分析结果:
# 不同深度的分析提示词
self.prompt_templates = {
'basic': """你是一个专业的舆情分析助手。请对每条消息进行基础的情感分析。
请按以下JSON格式返回:
{
"analysis_results": [
{
"message_id": "消息ID",
"sentiment": "情感倾向 (积极/消极/中性)",
"sentiment_score": "情感分数 (0-1)",
"keywords": ["关键词1", "关键词2"],
"key_points": "简要概述",
"influence_analysis": "基础影响分析",
"risk_level": "风险等级 (低/中/高)",
"timestamp": "分析时间戳"
}
]
}""",
'standard': """你是一个专业的舆情分析助手。请对每条消息进行标准深度的分析。
请按以下JSON格式返回:
{
"analysis_results": [
{
@@ -30,41 +47,69 @@ class AIAnalyzer:
"timestamp": "分析时间戳"
}
]
}
请确保每个字段都有值,并保持JSON格式的一致性。"""
}""",
'deep': """你是一个专业的舆情分析助手。请对每条消息进行深度分析。
请按以下JSON格式返回:
{
"analysis_results": [
{
"message_id": "消息ID",
"sentiment": "情感倾向 (积极/消极/中性)",
"sentiment_score": "情感分数 (0-1)",
"keywords": ["关键词1", "关键词2", "关键词3", "关键词4", "关键词5"],
"key_points": "详细的核心观点分析",
"influence_analysis": "深度影响分析,包括短期和长期影响",
"risk_factors": ["风险因素1", "风险因素2", "风险因素3"],
"risk_level": "风险等级 (低/中/高)",
"suggestions": ["建议1", "建议2", "建议3"],
"timestamp": "分析时间戳"
}
]
}"""
}
async def analyze_messages(self, messages: List[Dict]) -> List[Dict]:
async def analyze_messages(self, messages: List[Dict], batch_size: int = 50,
model_type: str = "gpt-3.5-turbo",
analysis_depth: str = "standard") -> List[Dict]:
"""分析一批消息并返回分析结果"""
try:
# 构建输入消息
formatted_messages = []
for msg in messages:
formatted_messages.append(f"消息ID: {msg['id']}\n内容: {msg['content']}")
all_results = []
messages_text = "\n---\n".join(formatted_messages)
# 调用OpenAI API
response = await openai.ChatCompletion.acreate(
model="gpt-3.5-turbo",
messages=[
{"role": "system", "content": self.system_prompt},
{"role": "user", "content": f"请分析以下消息:\n{messages_text}"}
],
temperature=0.3, # 降低随机性
max_tokens=2000,
n=1
)
# 解析返回结果
try:
result = json.loads(response.choices[0].message.content)
# 验证结果格式
if not isinstance(result, dict) or 'analysis_results' not in result:
raise ValueError("AI返回格式不正确")
return result['analysis_results']
except json.JSONDecodeError:
logging.error("AI返回结果解析失败")
return []
# 分批处理消息
for i in range(0, len(messages), batch_size):
batch = messages[i:i + batch_size]
formatted_messages = []
for msg in batch:
formatted_messages.append(f"消息ID: {msg['id']}\n内容: {msg['content']}")
messages_text = "\n---\n".join(formatted_messages)
# 获取对应深度的提示词
system_prompt = self.prompt_templates.get(analysis_depth, self.prompt_templates['standard'])
# 调用OpenAI API
response = await openai.ChatCompletion.acreate(
model=model_type,
messages=[
{"role": "system", "content": system_prompt},
{"role": "user", "content": f"请分析以下消息:\n{messages_text}"}
],
temperature=0.3, # 降低随机性
max_tokens=2000 if analysis_depth != 'deep' else 3000,
n=1
)
try:
result = json.loads(response.choices[0].message.content)
if isinstance(result, dict) and 'analysis_results' in result:
all_results.extend(result['analysis_results'])
else:
logging.error(f"API返回格式不正确: {response.choices[0].message.content}")
except json.JSONDecodeError as e:
logging.error(f"JSON解析失败: {e}")
continue
return all_results
except Exception as e:
logging.error(f"AI分析过程出错: {e}")
@@ -72,7 +117,7 @@ class AIAnalyzer:
def format_analysis_for_display(self, analysis: Dict) -> Dict:
"""将分析结果格式化为前端显示格式"""
return {
base_result = {
'id': analysis['message_id'],
'sentiment': analysis['sentiment'],
'sentiment_score': f"{float(analysis['sentiment_score']):.2%}",
@@ -84,6 +129,15 @@ class AIAnalyzer:
float(analysis['timestamp'])
).strftime('%Y-%m-%d %H:%M:%S')
}
# 如果是深度分析,添加额外信息
if 'risk_factors' in analysis:
base_result.update({
'risk_factors': analysis['risk_factors'],
'suggestions': analysis['suggestions']
})
return base_result
# 创建全局AI分析器实例
ai_analyzer = AIAnalyzer()