last iupdate with the recommandation system

projet_en_groupe/algorithme_netflix.py

@@ -474,32 +474,12 @@ def save_to_csv(data, default_filename='output.csv'):
 # début de l'algorithme de recommandation 
 
 # Load the CSV file
-"""
-def print_matrix(data_1):
-    # Initialize the category matrix with zeros
-    unique_categories = set(category for categories in data_1['listed_in'].str.split(', ') for category in categories)
-    category_matrix = {category: {other_category: 0 for other_category in unique_categories} for category in unique_categories}
+categories = []
 
-    # Fill the matrix with the number of movies/series for each category pair
-    for categories in data_1['listed_in'].str.split(', '):
-        for category1 in categories:
-            for category2 in categories:
-                category_matrix[category1][category2] += 1
 
-    category_df = pd.DataFrame(category_matrix)
-
-    # Save the DataFrame to an Excel file
-    output_file = 'category_matrix.xlsx'
-    category_df.to_excel(output_file, index_label='Categories')
-
-    print(f"Category matrix saved to {output_file}")
-    return category_matrix"""
-
-
-
-def read_movie_series_info(file_path_1):
+def read_movie_series_info(file_path):
     catalog = {}
-    with open(file_path_1, 'r') as info_file:
+    with open(file_path, 'r', encoding='utf-8') as info_file:
         info_reader = csv.reader(info_file)
         next(info_reader)  # Skip header row
         for row in info_reader:
@@ -507,57 +487,106 @@ def read_movie_series_info(file_path_1):
             catalog[show_id] = [title, listed_in.split(', ')]
     return catalog
 
-def read_user_ratings(file_path_2):
+def read_user_ratings(file_path):
     ratings = {}
-    user_ids = []  # Assurez-vous d'initialiser la liste user_ids
-    with open(file_path_2, 'r') as ratings_file:
+    with open(file_path, 'r', encoding='utf-8') as ratings_file:
         ratings_reader = csv.reader(ratings_file)
-        header = next(ratings_reader)  # Lire la première ligne pour obtenir les ID des utilisateurs
-        user_ids = header[1:]  # Les ID des utilisateurs sont dans la première ligne, sauf la première colonne (show_id)
+        header = next(ratings_reader)  # Skip header row
+        user_ids = list(map(int, header[1:]))
+        
         for row in ratings_reader:
             show_id = row[0]
             user_ratings = list(map(int, row[1:]))
-            ratings[show_id] = user_ratings
-    return ratings, user_ids  
+            ratings[show_id] = dict(zip(user_ids, user_ratings))
+            
+    return ratings
 
-def create_category_matrix(catalog, ratings, categories):
-    category_matrix = [[0 for _ in range(len(categories))] for _ in range(len(categories))]
+def create_category_matrix(catalog, categories):
+    # Créez la matrice avec une rangée et une colonne supplémentaires pour les noms de catégories
+    category_matrix = [[0 for _ in range(len(categories) + 1)] for _ in range(len(categories) + 1)]
 
+    # Ajoutez les noms de catégories à la première ligne et à la première colonne
+    for i in range(len(categories)):
+        category_matrix[0][i + 1] = categories[i]  # Ajoutez les noms de catégories à la première ligne
+        category_matrix[i + 1][0] = categories[i]  # Ajoutez les noms de catégories à la première colonne
+    
+    # Remplissez la matrice avec les données
     for show_id, movie_categories in catalog.items():
         for i in range(len(categories)):
             if categories[i] in movie_categories[1]:
                 for j in range(len(categories)):
                     if categories[j] in movie_categories[1]:
-                        category_matrix[i][j] += ratings[show_id][j]
-
+                        category_matrix[i + 1][j + 1] += 1  # Commencez à remplir à partir de la deuxième ligne et de la deuxième colonne
+    # Ajoutez les noms de catégories à la première colonne et les données de la matrice
     return category_matrix
 
-def calculate_user_preferences(category_matrix, user_preferences, categories):
-    user_vector = []
 
-    print("Length of user_preferences:", len(user_preferences))
-    print("Length of categories:", len(categories))
+def recommend_movies(user_id, catalog, user_ratings, category_matrix, threshold=0.5):
+    global categories  # Déclarer categories en tant que variable globale
+    categories = list(set(category for _, movie_info in catalog.items() for category in movie_info[1])) # permet d'actualiser la variable catégories global au niveau local
 
-    # Check if user_preferences is not empty and has the same length as categories
-    if len(user_preferences) != len(categories) or not user_preferences:
-        print("Error: Invalid user preferences")
-        return []
+    user_id = int(user_id)  # Convertir user_id en entier
 
-    for i in range(len(categories)):
-        user_vector.append(sum(category_matrix[i][j] * user_preferences[j] for j in range(len(categories))))
+    suggestions = {}
+    category_index = {}
+
+    # Créer le dictionnaire pour stocker les indices des catégories
+    category_index = {category: i + 1 for i, category in enumerate(categories)}  
+    for show_id, categories in catalog.items():
+        # Check if the user has rated the show
+        if show_id in user_ratings and user_id in user_ratings[show_id] and user_ratings[show_id][user_id] == 0:
+            # Liste des catégories communes entre le film/série et les films/séries notés par l'utilisateur
+            common_categories = [category for category in categories[1] if category in catalog[show_id][1]]
+            if common_categories:
+                # Calculez la similarité entre le film/série et les films/séries notés par l'utilisateur
+                similarity = sum(
+                    min(category_matrix[category_index[category]][category_index[user_category]] for user_category in common_categories)
+                    for category in categories[1]
+                )
+
+                # Ne recommandez que des films/séries dont la similarité dépasse le seuil spécifié
+                if similarity > threshold:
+                    suggestions[show_id] = {'title': catalog[show_id][0], 'similarity': similarity}
+
+    # Triez les suggestions par similarité décroissante
+    sorted_suggestions = sorted(suggestions.items(), key=lambda x: x[1]['similarity'], reverse=True)
+
+    return sorted_suggestions[:5]
+
+def recommandation_algorithm() :
+
+    # Replace file_path_1 and file_path_2 with the actual file paths
+    file_path_1 = "/Users/adrien/vscodeworkspace/coding-project/projet_en_groupe/data_cp_2023/netflix_titles-2.csv"
+    file_path_2 = "/Users/adrien/vscodeworkspace/coding-project/projet_en_groupe/data_cp_2023/ratings.csv"
+
+    user_id = input("quel est ton user ? ")
+
+    try:
+        user_id = int(user_id)
+    except ValueError:
+        print("Veuillez entrer un identifiant d'utilisateur valide.")
+        exit()
 
-    return user_vector
+    # Read data from CSV files
+    catalog = read_movie_series_info(file_path_1)
+    ratings = read_user_ratings(file_path_2)
+    # Create category matrix
+    categories = list(set(category for _, movie_info in catalog.items() for category in movie_info[1]))
+    category_matrix = create_category_matrix(catalog, categories)
 
-def recommend_movies(catalog, ratings, category_matrix, user_preferences, categories):
-    recommended_movies = []
+    # Display movies already viewed by the user
+    print("Films déjà vus par l'utilisateur:")
+    for show_id, user_rating in ratings.items():
+        if user_id in user_rating and user_rating[user_id] > 0:
+            print(f"- {catalog[show_id][0]}")
 
-    for show_id, (title, movie_categories) in catalog.items():
-        if show_id not in ratings:
-            movie_score = sum(category_matrix[i][j] * user_preferences[j] for i, j in enumerate(categories) if categories[i] in movie_categories[1])
-            recommended_movies.append((title, movie_score))
+    # Recommend movies
+    recommended_movies = recommend_movies(user_id, catalog, ratings, category_matrix, threshold=0.5)
 
-    recommended_movies.sort(key=lambda x: x[1], reverse=True)
-    return recommended_movies[:5]
+    # Display top 5 recommendations
+    print("\nTop 5 recommandations:")
+    for show_id, info in recommended_movies:
+        print(f"Title: {info['title']}, Similarity: {info['similarity']}")
 
 
 # Création du menu
@@ -618,55 +647,8 @@ def action() :
     directors_nationality(data_1)
   elif command == "17" :
     basic_statistics(data_1)
-  elif command == "test" :
-    # Read data from CSV files
-    catalog = read_movie_series_info(file_path_1)
-    ratings, user_ids = read_user_ratings(file_path_2)
-
-    transposed_ratings = {user_id: {} for user_id in user_ids}
-    for show_id, user_ratings in ratings.items():
-        for i, user_id in enumerate(user_ids):
-            transposed_ratings[user_id][show_id] = user_ratings[i]
-
-    print("User IDs in ratings:", transposed_ratings.keys())
-    # Check if the user ID exists in the ratings dictionary
-    user_id_to_recommend = input("Enter your user ID: ")
-    user_id_to_recommend = str(user_id_to_recommend)
-
-    # Check if the user ID exists in the ratings dictionary
-    if user_id_to_recommend not in transposed_ratings:
-        print(f"User with ID '{user_id_to_recommend}' not found in ratings.")
-    else:
-        # Get categories that exist in both catalog and user ratings
-        unique_types = file_path_1['listed_in'].unique()
-
-        print("Common Categories:", unique_types)
-
-        # Create category matrix
-        categories = list(common_categories)
-        print("Categories:", categories)
-
-        category_matrix = create_category_matrix(catalog, transposed_ratings, categories)
-
-
-        # Use user ratings as preferences
-        user_preferences = [transposed_ratings[user_id_to_recommend][category] for category in categories]
-        print("User Preferences:", user_preferences)
-
-        # Calculate user preferences
-        user_vector = calculate_user_preferences(category_matrix, user_preferences, categories)
-
-        # Recommend movies
-        recommended_movies = recommend_movies(catalog, transposed_ratings, category_matrix, user_vector, categories)
-
-        # Display top 5 recommendations
-        print(recommended_movies)
-
   elif command == "18" :
-    user_id_to_recommend = input("Enter your user ID: ")
-    # Assuming 'catalog', 'ratings', 'category_matrix', 'categories' are available from your data loading
-    recommended_movies = recommend_movies(catalog, ratings, category_matrix, user_preferences, categories)
-    print(recommended_movies)
+    recommandation_algorithm()
   elif command == "..." :
     return False
 

projet_en_groupe/recom_sys.py

@@ -56,6 +56,7 @@ def recommend_movies(user_id, catalog, user_ratings, category_matrix, threshold=
 
     # Créer le dictionnaire pour stocker les indices des catégories
     category_index = {category: i + 1 for i, category in enumerate(categories)}
+    print(category_index)
 
     for show_id, categories in catalog.items():
         # Check if the user has rated the show
@@ -100,10 +101,10 @@ if __name__ == "__main__":
     category_matrix = create_category_matrix(catalog, categories)
 
     # Display movies already viewed by the user
-    print("Films déjà vus par l'utilisateur:")
+    """print("Films déjà vus par l'utilisateur:")
     for show_id, user_rating in ratings.items():
         if user_id in user_rating and user_rating[user_id] > 0:
-            print(f"- {catalog[show_id][0]}")
+            print(f"- {catalog[show_id][0]}")"""
 
     # Recommend movies
     recommended_movies = recommend_movies(user_id, catalog, ratings, category_matrix, threshold=0.5)