Update src/streamlit_app.py

src/streamlit_app.py

@@ -6,26 +6,31 @@ import json
 from datetime import datetime
 import requests
 import difflib
+import pickle
+from keras.models import load_model
+import os
+import requests
 
 from recommendation_utils import (
     load_nn_model, load_svd_model, load_trainset,
     recommend_with_nn, recommend_with_svd
 )
-
-import pickle
-from keras.models import load_model
-import os
-import requests
+from recommendation_utils import (
+    load_nn_model, load_svd_model, load_trainset,
+    recommend_with_nn, recommend_with_svd, load_encodings
+)
 
 @st.cache_resource(show_spinner=False)
 def download_models_once():
+    DOWNLOAD_DIR = "/tmp"
+    
     files = {
         "encodings.pkl": "1EzpdpaopfUp-Tfc7YjxPVYQUwnU_BX5-",
         "recommender_model.keras": "1OwdH3RxlQfAX9UbUB7RwlS13i9uDZuG-",
         "svd_model.pkl": "1fN2biQruVjJHHv2vX1g1hLuMeJoPqyFX",
         "trainset.pkl": "1IDVVAQ57Xvf3HCAikbOSQgAigdHP7Ik7"
     }
-
+    
     def gdrive_download(file_id, destination):
         URL = "https://drive.google.com/uc?export=download"
         session = requests.Session()
@@ -40,23 +45,27 @@ def download_models_once():
             for chunk in response.iter_content(32768):
                 if chunk:
                     f.write(chunk)
-
+    
     for filename, file_id in files.items():
-        if not os.path.exists(filename):
+        full_path = os.path.join(DOWNLOAD_DIR, filename)
+        if not os.path.exists(full_path):
             with st.spinner(f"Downloading {filename}..."):
-                gdrive_download(file_id, filename)
+                gdrive_download(file_id, full_path)
+
 
 download_models_once()
 
 @st.cache_resource
 def load_models():
-    nn_model = load_nn_model("recommender_model.keras")
-    svd_model = load_svd_model("svd_model.pkl")
-    trainset = load_trainset("trainset.pkl")
+    nn_model = load_nn_model("/tmp/recommender_model.keras")
+    svd_model = load_svd_model("/tmp/svd_model.pkl")
+    trainset = load_trainset("/tmp/trainset.pkl")
     return nn_model, svd_model, trainset
 
 nn_model, svd_model, trainset = load_models()
 
+encodings = load_encodings("/tmp/encodings.pkl")
+
 st.set_page_config(layout="wide")
 
 MOVIES_PATH = os.path.join(os.path.dirname(__file__), "movies.csv")
@@ -544,7 +553,8 @@ else:
         if user_ratings_dict:
             if st.session_state["model_selection"] == "Neural Network":
                 available_movies = movie_df["movieId"].tolist()
-                recommendations = recommend_with_nn(user_ratings_dict, nn_model, available_movies, top_n=10)
+                recommendations = recommend_with_nn(user_ratings, nn_model, encodings, top_n=10)
+
             else:
                 ratings_full = pd.DataFrame(all_ratings_data)
                 ratings_full["userId"] = 999999  # Dummy user