|
|
|
@ -1,24 +1,29 @@
|
|
|
|
|
# pip install gensim
|
|
|
|
|
# pip install gensim jieba
|
|
|
|
|
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
|
|
|
|
|
|
|
|
|
|
# 加载预训练的 Word2Vec 模型
|
|
|
|
|
model_path = "D:/Tencent_AILab_ChineseEmbedding/Tencent_AILab_ChineseEmbedding.txt" # 替换为你的 Word2Vec 模型路径
|
|
|
|
|
# 参考文档:使用gensim之KeyedVectors操作词向量模型
|
|
|
|
|
# https://www.cnblogs.com/bill-h/p/14655224.html
|
|
|
|
|
# 读取词向量模型(限定前10000个词)
|
|
|
|
|
model = KeyedVectors.load_word2vec_format(model_path, binary=False, limit=10000)
|
|
|
|
|
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 = text.split()
|
|
|
|
|
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]
|
|
|
|
|
print(f"有效词向量数量: {len(embeddings)}")
|
|
|
|
|
if embeddings:
|
|
|
|
|
return sum(embeddings) / len(embeddings) # 取词向量的平均值
|
|
|
|
|
avg_embedding = sum(embeddings) / len(embeddings)
|
|
|
|
|
print(f"生成的平均向量: {avg_embedding[:5]}...") # 打印前 5 维
|
|
|
|
|
return avg_embedding
|
|
|
|
|
else:
|
|
|
|
|
return [0.0] * model.vector_size # 如果文本中没有有效词,返回零向量
|
|
|
|
|
print("未找到有效词,返回零向量")
|
|
|
|
|
return [0.0] * model.vector_size
|
|
|
|
|
|
|
|
|
|
# 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)
|
|
|
|
@ -45,7 +50,7 @@ search_params = {
|
|
|
|
|
"params": {"nprobe": 100} # 设置 IVF_FLAT 的 nprobe 参数
|
|
|
|
|
}
|
|
|
|
|
start_time = time.time()
|
|
|
|
|
results = collection_manager.search(current_embedding, search_params, limit=5) # 返回 5 条结果
|
|
|
|
|
results = collection_manager.search(current_embedding, search_params, limit=2) # 返回 2 条结果
|
|
|
|
|
end_time = time.time()
|
|
|
|
|
|
|
|
|
|
# 8. 输出查询结果
|
|
|
|
|