parent
7ff869e4a5
commit
8decf74959
@ -1,40 +0,0 @@
|
||||
from pymilvus import FieldSchema, DataType, utility
|
||||
|
||||
from WxMini.Milvus.Config.MulvusConfig import *
|
||||
from WxMini.Milvus.Utils.MilvusCollectionManager import MilvusCollectionManager
|
||||
from WxMini.Milvus.Utils.MilvusConnectionPool import *
|
||||
|
||||
# 1. 使用连接池管理 Milvus 连接
|
||||
milvus_pool = MilvusConnectionPool(host=MS_HOST, port=MS_PORT, max_connections=MS_MAX_CONNECTIONS)
|
||||
|
||||
# 2. 从连接池中获取一个连接
|
||||
connection = milvus_pool.get_connection()
|
||||
|
||||
# 3. 初始化集合管理器
|
||||
collection_name = MS_COLLECTION_NAME
|
||||
collection_manager = MilvusCollectionManager(collection_name)
|
||||
|
||||
# 4. 判断集合是否存在,存在则删除
|
||||
if utility.has_collection(collection_name):
|
||||
print(f"集合 '{collection_name}' 已存在,正在删除...")
|
||||
utility.drop_collection(collection_name)
|
||||
print(f"集合 '{collection_name}' 已删除。")
|
||||
|
||||
# 5. 定义集合的字段和模式
|
||||
fields = [
|
||||
FieldSchema(name="id", dtype=DataType.INT64, is_primary=True, auto_id=True), # 主键字段,自动生成 ID
|
||||
FieldSchema(name="text", dtype=DataType.VARCHAR, max_length=500), # 存储对话文本
|
||||
FieldSchema(name="embedding", dtype=DataType.FLOAT_VECTOR, dim=MS_DIMENSION) # 向量字段,维度为 200
|
||||
]
|
||||
schema_description = "Simple demo collection"
|
||||
|
||||
# 6. 创建集合
|
||||
print(f"正在创建集合 '{collection_name}'...")
|
||||
collection_manager.create_collection(fields, schema_description)
|
||||
print(f"集合 '{collection_name}' 创建成功。")
|
||||
|
||||
# 7. 释放连接
|
||||
milvus_pool.release_connection(connection)
|
||||
|
||||
# 8. 关闭连接池
|
||||
milvus_pool.close()
|
@ -1,27 +0,0 @@
|
||||
from WxMini.Milvus.Utils.MilvusCollectionManager import MilvusCollectionManager
|
||||
from WxMini.Milvus.Utils.MilvusConnectionPool import *
|
||||
from WxMini.Milvus.Config.MulvusConfig import *
|
||||
|
||||
# 1. 使用连接池管理 Milvus 连接
|
||||
milvus_pool = MilvusConnectionPool(host=MS_HOST, port=MS_PORT, max_connections=MS_MAX_CONNECTIONS)
|
||||
|
||||
# 2. 从连接池中获取一个连接
|
||||
connection = milvus_pool.get_connection()
|
||||
|
||||
# 3. 初始化集合管理器
|
||||
collection_name = MS_COLLECTION_NAME
|
||||
collection_manager = MilvusCollectionManager(collection_name)
|
||||
|
||||
# 4. 创建索引
|
||||
index_params = {
|
||||
"index_type": "IVF_FLAT", # 使用 IVF_FLAT 索引类型
|
||||
"metric_type": "L2", # 使用 L2 距离度量方式
|
||||
"params": {"nlist": 128} # 设置 IVF_FLAT 的 nlist 参数
|
||||
}
|
||||
collection_manager.create_index("embedding", index_params)
|
||||
|
||||
# 5. 释放连接
|
||||
milvus_pool.release_connection(connection)
|
||||
|
||||
# 6. 关闭连接池
|
||||
milvus_pool.close()
|
@ -1,45 +0,0 @@
|
||||
from WxMini.Milvus.Utils.MilvusCollectionManager import MilvusCollectionManager
|
||||
from WxMini.Milvus.Utils.MilvusConnectionPool import *
|
||||
from WxMini.Milvus.Config.MulvusConfig import *
|
||||
|
||||
# 1. 使用连接池管理 Milvus 连接
|
||||
milvus_pool = MilvusConnectionPool(host=MS_HOST, port=MS_PORT, max_connections=MS_MAX_CONNECTIONS)
|
||||
|
||||
# 2. 从连接池中获取一个连接
|
||||
connection = milvus_pool.get_connection()
|
||||
|
||||
# 3. 初始化集合管理器
|
||||
collection_name = MS_COLLECTION_NAME
|
||||
collection_manager = MilvusCollectionManager(collection_name)
|
||||
|
||||
# 4. 加载集合到内存
|
||||
collection_manager.load_collection()
|
||||
print(f"集合 '{collection_name}' 已加载到内存。")
|
||||
|
||||
# 5. 查询所有数据
|
||||
try:
|
||||
# 使用 Milvus 的 query 方法查询所有数据
|
||||
results = collection_manager.collection.query(
|
||||
expr="", # 空表达式表示查询所有数据
|
||||
output_fields=["id", "text", "embedding"], # 指定返回的字段
|
||||
limit=1000 # 设置最大返回记录数
|
||||
)
|
||||
print("查询结果:")
|
||||
if results:
|
||||
for result in results:
|
||||
try:
|
||||
text = result["text"] # 获取 text 字段
|
||||
embedding = result["embedding"] # 获取 embedding 字段
|
||||
print(f"ID: {result['id']}, Text: {text}, Embedding: {embedding[:5]}...") # 只打印前 5 维向量
|
||||
except Exception as e:
|
||||
print(f"查询失败: {e}")
|
||||
else:
|
||||
print("未找到相关数据,请检查查询参数或数据。")
|
||||
except Exception as e:
|
||||
print(f"查询失败: {e}")
|
||||
|
||||
# 6. 释放连接
|
||||
milvus_pool.release_connection(connection)
|
||||
|
||||
# 7. 关闭连接池
|
||||
milvus_pool.close()
|
@ -1,78 +0,0 @@
|
||||
import time
|
||||
import jieba # 导入 jieba 分词库
|
||||
from WxMini.Milvus.Utils.MilvusCollectionManager import MilvusCollectionManager
|
||||
from WxMini.Milvus.Utils.MilvusConnectionPool import *
|
||||
from WxMini.Milvus.Config.MulvusConfig import *
|
||||
from gensim.models import KeyedVectors
|
||||
|
||||
# 1. 加载预训练的 Word2Vec 模型
|
||||
model_path = MS_MODEL_PATH # 替换为你的 Word2Vec 模型路径
|
||||
model = KeyedVectors.load_word2vec_format(model_path, binary=False, limit=MS_MODEL_LIMIT)
|
||||
print(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
|
||||
print(f"文本: {text}, 分词结果: {words}")
|
||||
embeddings = [model[word] for word in words if word in model] # 获取有效词向量
|
||||
print(f"有效词向量数量: {len(embeddings)}")
|
||||
if embeddings:
|
||||
avg_embedding = sum(embeddings) / len(embeddings) # 计算平均向量
|
||||
print(f"生成的平均向量: {avg_embedding[:5]}...") # 打印前 5 维
|
||||
return avg_embedding
|
||||
else:
|
||||
print("未找到有效词,返回零向量")
|
||||
return [0.0] * model.vector_size
|
||||
|
||||
|
||||
# 2. 使用连接池管理 Milvus 连接
|
||||
milvus_pool = MilvusConnectionPool(host=MS_HOST, port=MS_PORT, max_connections=MS_MAX_CONNECTIONS)
|
||||
|
||||
# 3. 从连接池中获取一个连接
|
||||
connection = milvus_pool.get_connection()
|
||||
|
||||
# 4. 初始化集合管理器
|
||||
collection_name = MS_COLLECTION_NAME
|
||||
collection_manager = MilvusCollectionManager(collection_name)
|
||||
|
||||
# 5. 加载集合到内存
|
||||
collection_manager.load_collection()
|
||||
|
||||
# 6. 输入一句话
|
||||
input_text = input("请输入一句话:") # 例如:“我今天心情不太好”
|
||||
|
||||
# 7. 将文本转换为嵌入向量
|
||||
current_embedding = text_to_embedding(input_text)
|
||||
print(f"当前文本的向量: {current_embedding[:5]}...") # 打印前 5 维
|
||||
|
||||
# 8. 查询与当前对话最相关的历史对话
|
||||
search_params = {
|
||||
"metric_type": "L2", # 使用 L2 距离度量方式
|
||||
"params": {"nprobe": MS_NPROBE} # 设置 IVF_FLAT 的 nprobe 参数
|
||||
}
|
||||
start_time = time.time()
|
||||
results = collection_manager.search(current_embedding, search_params, limit=2) # 返回 2 条结果
|
||||
end_time = time.time()
|
||||
|
||||
# 9. 输出查询结果
|
||||
print("最相关的历史对话:")
|
||||
if results:
|
||||
for hits in results:
|
||||
for hit in hits:
|
||||
try:
|
||||
text = collection_manager.query_text_by_id(hit.id)
|
||||
print(f"- {text} (距离: {hit.distance})")
|
||||
except Exception as e:
|
||||
print(f"查询失败: {e}")
|
||||
else:
|
||||
print("未找到相关历史对话,请检查查询参数或数据。")
|
||||
|
||||
# 10. 输出查询耗时
|
||||
print(f"查询耗时: {end_time - start_time:.4f} 秒")
|
||||
|
||||
# 11. 释放连接
|
||||
milvus_pool.release_connection(connection)
|
||||
|
||||
# 12. 关闭连接池
|
||||
milvus_pool.close()
|
Loading…
Reference in new issue