Update GlobalMovies.py

GlobalMovies.py

@@ -1,63 +1,94 @@
-# assignment3_analysis.py
-# AI Movie Recommender System – Foreign Films Edition
-# Author: Karley
-# Description: This script collects foreign film data using the TMDB API,
-# processes text using NLP (TF-IDF), and recommends similar movies using content-based filtering.
+# GlobalMovies.py
+# Backend logic for WorldFlix – Global Movie Recommender
 
 import requests
 from sklearn.feature_extraction.text import TfidfVectorizer
 from sklearn.metrics.pairwise import cosine_similarity
 
-API_KEY = "3859ec79f3bda0615907794345ff1f17"  # Replace with your actual TMDB API key
+# -----------------------------
+# TMDB API Setup
+# -----------------------------
+API_KEY = "your_tmdb_api_key_here"  # 🔁 Replace with your real key
 
 # -----------------------------
-# 1. TMDB API DATA COLLECTION
+# Region + Language Mapping
+# -----------------------------
+region_languages = {
+    "European Films": {
+        "fr": "French",
+        "it": "Italian",
+        "es": "Spanish",
+        "de": "German",
+        "pt": "Portuguese"
+    },
+    "Asian Films": {
+        "ja": "Japanese",
+        "ko": "Korean",
+        "hi": "Hindi",
+        "zh": "Chinese"
+    },
+    "Latin American Films": {
+        "es": "Spanish",
+        "pt": "Portuguese"
+    },
+    "Middle Eastern Films": {
+        "fa": "Persian",
+        "ar": "Arabic",
+        "tr": "Turkish"
+    },
+    "African Films": {
+        "sw": "Swahili",
+        "ar": "Arabic",
+        "fr": "French"
+    }
+}
+
+# -----------------------------
+# TMDB Data Fetching
 # -----------------------------
 def get_foreign_movies(original_language="fr", page=1):
     """
-    Retrieves a list of foreign language movies from TMDB by language.
+    Retrieves a list of foreign language movies from TMDB.
     """
     url = f"https://api.themoviedb.org/3/discover/movie?api_key={API_KEY}&with_original_language={original_language}&sort_by=popularity.desc&page={page}"
     response = requests.get(url)
-    return response.json()["results"]
+    return response.json().get("results", [])
 
-def build_movie_dataset(language="fr", pages=2):
+def build_movie_dataset(languages=["fr", "es", "it"], pages=2):
     """
-    Builds a dataset of movie titles, descriptions, and original languages.
+    Builds a dataset of movies from a list of language codes.
+    Returns: titles, descriptions, language codes
     """
-    titles = []
-    descriptions = []
-    languages = []
-    for page in range(1, pages + 1):
-        movies = get_foreign_movies(language, page)
-        for movie in movies:
-            if movie.get("overview"):
-                titles.append(movie["title"])
-                descriptions.append(movie["overview"])
-                languages.append(movie["original_language"])
-    return titles, descriptions, languages
+    titles, descriptions, langs = [], [], []
+    for lang in languages:
+        for page in range(1, pages + 1):
+            movies = get_foreign_movies(original_language=lang, page=page)
+            for movie in movies:
+                if movie.get("overview") and movie.get("title"):
+                    titles.append(movie["title"])
+                    descriptions.append(movie["overview"])
+                    langs.append(movie["original_language"])
+    return titles, descriptions, langs
 
-# Build your dataset
-movie_titles, movie_descriptions, movie_languages = build_movie_dataset("fr")
+# 🔁 Default: Pull movies for all region languages (adjust pages if needed)
+all_region_langs = {lang for group in region_languages.values() for lang in group}
+movie_titles, movie_descriptions, movie_languages = build_movie_dataset(list(all_region_langs), pages=2)
 
 # -----------------------------
-# 2. NLP PROCESSING (TF-IDF)
+# NLP: TF-IDF Matrix
 # -----------------------------
 def create_tfidf_matrix(descriptions):
-    """
-    Transforms movie descriptions into a TF-IDF matrix for similarity comparison.
-    """
     tfidf = TfidfVectorizer(stop_words="english")
     return tfidf.fit_transform(descriptions)
 
 tfidf_matrix = create_tfidf_matrix(movie_descriptions)
 
 # -----------------------------
-# 3. CONTENT-BASED RECOMMENDER
+# Recommender Function
 # -----------------------------
-def recommend_foreign(movie_title, tfidf_matrix, movie_titles, movie_languages, lang="fr"):
+def recommend_foreign(movie_title, tfidf_matrix, movie_titles, movie_languages, lang="fr", top_n=5):
     """
-    Returns 5 similar foreign-language movies using cosine similarity.
+    Returns top_n similar movies in the same language as movie_title.
     """
     if movie_title not in movie_titles:
         return ["Movie not found in dataset."]
@@ -70,38 +101,7 @@ def recommend_foreign(movie_title, tfidf_matrix, movie_titles, movie_languages,
     for i in indices:
         if movie_languages[i] == lang and i != idx:
             recommendations.append(movie_titles[i])
-        if len(recommendations) == 5:
-            break
-    return recommendations
-# --- NLP: TF-IDF Vectorization ---
-from sklearn.feature_extraction.text import TfidfVectorizer
-
-def create_tfidf_matrix(descriptions):
-    """
-    Transforms text descriptions into TF-IDF vectors.
-    """
-    tfidf = TfidfVectorizer(stop_words='english')
-    return tfidf.fit_transform(descriptions)
-
-tfidf_matrix = create_tfidf_matrix(movie_descriptions)
-# --- Content-Based Filtering using Cosine Similarity ---
-from sklearn.metrics.pairwise import cosine_similarity
-
-def recommend_movies(movie_title, tfidf_matrix, titles, languages, lang="fr", top_n=5):
-    """
-    Recommends movies similar to the given title, filtered by language.
-    """
-    if movie_title not in titles:
-        return ["Movie not found in dataset."]
-    
-    idx = titles.index(movie_title)
-    cosine_sim = cosine_similarity(tfidf_matrix[idx], tfidf_matrix).flatten()
-    sim_indices = cosine_sim.argsort()[::-1]
-
-    recommendations = []
-    for i in sim_indices:
-        if i != idx and languages[i] == lang:
-            recommendations.append(titles[i])
         if len(recommendations) == top_n:
             break
     return recommendations
+