File size: 2,452 Bytes
b4a0526
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
from train_model import train_triplet_model
from embed_data import embed_product_data, embed_user_data
from calculate_similarity import calculate_cosine_similarity
from pymongo import MongoClient

# MongoDB 연결
client = MongoClient("mongodb+srv://waseoke:[email protected]/test?retryWrites=true&w=majority")
db = client["two_tower_model"]
product_collection = db["product_tower"]
user_collection = db["user_tower"]
product_embedding_collection = db["product_embeddings"]
user_embedding_collection = db["user_embeddings"]

# 모델 학습
def train_model_and_embed():
    product_model = None  # Define or load your model
    anchor_data, positive_data, negative_data = load_training_data()
    trained_model = train_triplet_model(product_model, anchor_data, positive_data, negative_data)

    return trained_model

# 데이터 임베딩 및 저장
def embed_and_save():
    all_products = list(product_collection.find())
    all_users = list(user_collection.find())

    for product_data in all_products:
        embedding = embed_product_data(product_data)
        product_embedding_collection.update_one(
            {"product_id": product_data["product_id"]},
            {"$set": {"embedding": embedding.tolist()}},
            upsert=True
        )

    for user_data in all_users:
        embedding = embed_user_data(user_data)
        user_embedding_collection.update_one(
            {"user_id": user_data["user_id"]},
            {"$set": {"embedding": embedding.tolist()}},
            upsert=True
        )

# 추천 실행
def recommend(user_id, top_n=5):
    user_embedding_data = user_embedding_collection.find_one({"user_id": user_id})
    if not user_embedding_data:
        print(f"No embedding found for user_id: {user_id}")
        return []

    user_embedding = np.array(user_embedding_data["embedding"])
    all_products = list(product_embedding_collection.find())
    product_ids = [prod["product_id"] for prod in all_products]
    product_embeddings = [prod["embedding"] for prod in all_products]

    recommendations = calculate_cosine_similarity(user_embedding, product_embeddings, product_ids, top_n)
    print(f"Recommendations for user {user_id}: {recommendations}")
    return recommendations

# 실행
if __name__ == "__main__":
    # Train and embed data
    train_model_and_embed()
    embed_and_save()

    # Recommend products for a user
    user_id = "정우석"
    recommend(user_id, top_n=3)