'commit'

dsRag/Util/EsSearchUtil.py

@@ -1,6 +1,9 @@
-from elasticsearch import Elasticsearch
 import jieba
-import numpy as np
+from elasticsearch import Elasticsearch
+from gensim.models import KeyedVectors
+
+from Config.Config import MS_MODEL_PATH, MS_MODEL_LIMIT
+
 
 class EsSearchUtil:
     def __init__(self, es_config):
@@ -16,9 +19,18 @@ class EsSearchUtil:
         )
 
     def text_to_embedding(self, text):
-        # 当前实现为随机向量生成，后续可替换为实际模型
-        vector = np.random.rand(200).tolist()
-        return vector
+        # 加载预训练模型
+        model = KeyedVectors.load_word2vec_format(MS_MODEL_PATH, binary=False, limit=MS_MODEL_LIMIT)
+        # 对文本分词并计算平均向量
+        words = jieba.lcut(text)
+        vectors = [model[word] for word in words if word in model]
+        
+        if not vectors:
+            return [0.0] * model.vector_size
+            
+        # 计算平均向量
+        avg_vector = [sum(dim)/len(vectors) for dim in zip(*vectors)]
+        return avg_vector
 
     def vector_search(self, query, size=10):
         query_embedding = self.text_to_embedding(query)