'commit'

AI/WxMini/Start.py

@@ -1,72 +1,24 @@
-import os
+import asyncio
-import uuid
+import logging
 import time
-import jieba
+import uuid
-from fastapi import FastAPI, Form, HTTPException
-from openai import AsyncOpenAI  # 使用异步客户端
-from gensim.models import KeyedVectors
 from contextlib import asynccontextmanager
-from WxMini.Utils.OssUtil import upload_mp3_to_oss, upload_mp3_to_oss_from_memory
+
-from WxMini.Utils.TtsUtil import TTS
+from fastapi import FastAPI, Form, HTTPException
+from openai import AsyncOpenAI
+
+from WxMini.Milvus.Config.MulvusConfig import *
 from WxMini.Milvus.Utils.MilvusCollectionManager import MilvusCollectionManager
 from WxMini.Milvus.Utils.MilvusConnectionPool import *
-from WxMini.Milvus.Config.MulvusConfig import *
+from WxMini.Utils.OssUtil import upload_mp3_to_oss_from_memory
-import asyncio  # 引入异步支持
+from WxMini.Utils.TtsUtil import TTS
-import logging  # 增加日志记录
+from WxMini.Utils.MySQLUtil import init_mysql_pool, save_chat_to_mysql
-
+from WxMini.Utils.EmbeddingUtil import text_to_embedding
-import jieba.analyse
 
 # 配置日志
 logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s")
 logger = logging.getLogger(__name__)
 
-from aiomysql import create_pool
-
-# MySQL 配置
-MYSQL_CONFIG = {
-    "host": MYSQL_HOST,
-    "port": MYSQL_PORT,
-    "user": MYSQL_USER,
-    "password": MYSQL_PASSWORD,
-    "db": MYSQL_DB_NAME,
-    "minsize": 1,
-    "maxsize": 20,
-}
-
-
-# 保存聊天记录到 MySQL
-async def save_chat_to_mysql(mysql_pool, session_id, prompt, result):
-    async with mysql_pool.acquire() as conn:
-        async with conn.cursor() as cur:
-            await cur.execute(
-                "INSERT INTO t_chat_log (session_id, user_input, model_response, create_time) VALUES (%s, %s, %s, NOW())",
-                (session_id, prompt, result)
-            )
-            await conn.commit()
-
-
-# 初始化 MySQL 连接池
-async def init_mysql_pool():
-    return await create_pool(**MYSQL_CONFIG)
-
-
-# 提取用户输入的关键词
-def extract_keywords(text, topK=3):
-    """
-    提取用户输入的关键词
-    :param text: 用户输入的文本
-    :param topK: 返回的关键词数量
-    :return: 关键词列表
-    """
-    keywords = jieba.analyse.extract_tags(text, topK=topK)
-    return keywords
-
-
-# 初始化 Word2Vec 模型
-model_path = MS_MODEL_PATH
-model = KeyedVectors.load_word2vec_format(model_path, binary=False, limit=MS_MODEL_LIMIT)
-logger.info(f"模型加载成功，词向量维度: {model.vector_size}")
-
 # 初始化 Milvus 连接池
 milvus_pool = MilvusConnectionPool(host=MS_HOST, port=MS_PORT, max_connections=MS_MAX_CONNECTIONS)
 
@@ -74,22 +26,6 @@ milvus_pool = MilvusConnectionPool(host=MS_HOST, port=MS_PORT, max_connections=M
 collection_name = MS_COLLECTION_NAME
 collection_manager = MilvusCollectionManager(collection_name)
 
-
-# 将文本转换为嵌入向量
-def text_to_embedding(text):
-    words = jieba.lcut(text)  # 使用 jieba 分词
-    logger.info(f"文本: {text}, 分词结果: {words}")
-    embeddings = [model[word] for word in words if word in model]
-    logger.info(f"有效词向量数量: {len(embeddings)}")
-    if embeddings:
-        avg_embedding = sum(embeddings) / len(embeddings)
-        logger.info(f"生成的平均向量: {avg_embedding[:5]}...")  # 打印前 5 维
-        return avg_embedding
-    else:
-        logger.warning("未找到有效词，返回零向量")
-        return [0.0] * model.vector_size
-
-
 # 使用 Lifespan Events 处理应用启动和关闭逻辑
 @asynccontextmanager
 async def lifespan(app: FastAPI):
@@ -106,7 +42,6 @@ async def lifespan(app: FastAPI):
     await app.state.mysql_pool.wait_closed()
     logger.info("Milvus 和 MySQL 连接池已关闭。")
 
-
 # 初始化 FastAPI 应用
 app = FastAPI(lifespan=lifespan)
 
@@ -116,7 +51,6 @@ client = AsyncOpenAI(
     base_url="https://dashscope.aliyuncs.com/compatible-mode/v1",
 )
 
-
 @app.post("/reply")
 async def reply(session_id: str = Form(...), prompt: str = Form(...)):
     """
@@ -220,8 +154,7 @@ async def reply(session_id: str = Form(...), prompt: str = Form(...)):
             audio_data = await asyncio.to_thread(t.generate_audio, result)  # 假设 TTS 类有一个 generate_audio 方法返回音频数据
 
             # 将音频数据直接上传到 OSS
-            await asyncio.to_thread(upload_mp3_to_oss_from_memory, tts_file,
+            await asyncio.to_thread(upload_mp3_to_oss_from_memory, tts_file, audio_data)
-                                    audio_data)  # 假设 upload_mp3_to_oss_from_memory 支持从内存上传
             logger.info(f"TTS 文件已直接上传到 OSS: {tts_file}")
 
             # 完整的 URL
@@ -241,9 +174,8 @@ async def reply(session_id: str = Form(...), prompt: str = Form(...)):
         # 释放连接
         milvus_pool.release_connection(connection)
 
-
 # 运行 FastAPI 应用
 if __name__ == "__main__":
     import uvicorn
 
-    uvicorn.run("Start:app", host="0.0.0.0", port=5800, workers=1)
+    uvicorn.run("Start:app", host="0.0.0.0", port=5800, workers=1)

AI/WxMini/Utils/EmbeddingUtil.py

@@ -0,0 +1,26 @@
+import logging
+import jieba
+from gensim.models import KeyedVectors
+
+# 配置日志
+logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s")
+logger = logging.getLogger(__name__)
+
+# 初始化 Word2Vec 模型
+model_path = r"D:\Tencent_AILab_ChineseEmbedding\Tencent_AILab_ChineseEmbedding.txt"
+model = KeyedVectors.load_word2vec_format(model_path, binary=False, limit=10000)
+logger.info(f"模型加载成功，词向量维度: {model.vector_size}")
+
+# 将文本转换为嵌入向量
+def text_to_embedding(text):
+    words = jieba.lcut(text)  # 使用 jieba 分词
+    logger.info(f"文本: {text}, 分词结果: {words}")
+    embeddings = [model[word] for word in words if word in model]
+    logger.info(f"有效词向量数量: {len(embeddings)}")
+    if embeddings:
+        avg_embedding = sum(embeddings) / len(embeddings)
+        logger.info(f"生成的平均向量: {avg_embedding[:5]}...")  # 打印前 5 维
+        return avg_embedding
+    else:
+        logger.warning("未找到有效词，返回零向量")
+        return [0.0] * model.vector_size

AI/WxMini/Utils/MySQLUtil.py

@@ -0,0 +1,36 @@
+import asyncio
+import logging
+from aiomysql import create_pool
+from WxMini.Milvus.Config.MulvusConfig import *
+
+# 配置日志
+logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s")
+logger = logging.getLogger(__name__)
+
+# MySQL 配置
+MYSQL_CONFIG = {
+    "host": MYSQL_HOST,
+    "port": MYSQL_PORT,
+    "user": MYSQL_USER,
+    "password": MYSQL_PASSWORD,
+    "db": MYSQL_DB_NAME,
+    "minsize": 1,
+    "maxsize": 20,
+}
+
+
+# 初始化 MySQL 连接池
+async def init_mysql_pool():
+    return await create_pool(**MYSQL_CONFIG)
+
+
+# 保存聊天记录到 MySQL
+async def save_chat_to_mysql(mysql_pool, session_id, prompt, result):
+    async with mysql_pool.acquire() as conn:
+        async with conn.cursor() as cur:
+            await cur.execute(
+                "INSERT INTO t_chat_log (session_id, user_input, model_response, create_time) VALUES (%s, %s, %s, NOW())",
+                (session_id, prompt, result)
+            )
+            await conn.commit()
+            logger.info("用户输入和大模型反馈已记录到 MySQL 数据库。")