Optimizated search

components/gradio_ui.py

@@ -1,9 +1,7 @@
 import gradio as gr
 
-
 def create_interface(engine):
     def get_recommendations_text(query):
-        """Wrapper function to safely get only the text result"""
         try:
             result = engine.get_recommendations(query)
             if isinstance(result, tuple) and len(result) >= 1:
@@ -13,8 +11,68 @@ def create_interface(engine):
         except Exception as e:
             return f"❌ Error: {str(e)}"
 
+    def get_thumbnails_html(query):
+        try:
+            result = engine.get_recommendations(query)
+            if isinstance(result, tuple) and len(result) >= 1:
+                search_results = engine.get_recommendations(query)
+                
+
+                thumbnails_html = []
+                thumbnails_html.append("""
+                <div style="display: grid; grid-template-columns: repeat(auto-fill, minmax(150px, 1fr)); gap: 15px; padding: 20px; max-height: 600px; overflow-y: auto;">
+                """)
+
+                thumbnails_html.append("""
+                <div style="grid-column: 1 / -1; text-align: center; padding: 20px; color: #666;">
+                    Thumbnails will appear here when poster URLs are available
+                </div>
+                """)
+                
+                thumbnails_html.append("</div>")
+                return "".join(thumbnails_html)
+                
+        except Exception as e:
+            return f"<div style='color: red; padding: 20px;'>❌ Error: {str(e)}</div>"
+
+    def get_thumbnails_from_results(query):
+        """Get thumbnails from search results"""
+        try:
+            formatted_results, df_results = engine.get_recommendations(query)
+        
+            html_parts = []
+            html_parts.append("""
+            <div style="display: grid; grid-template-columns: repeat(auto-fill, minmax(150px, 1fr)); gap: 15px; padding: 20px; max-height: 600px; overflow-y: auto; background: #f8f9fa; border-radius: 8px;">
+            """)
+            
+            for i in range(10): 
+                html_parts.append(f"""
+                <div style="position: relative; border-radius: 8px; overflow: hidden; box-shadow: 0 2px 8px rgba(0,0,0,0.1); transition: transform 0.2s; cursor: pointer;" 
+                     onmouseover="this.style.transform='scale(1.05)'" 
+                     onmouseout="this.style.transform='scale(1)'">
+                    <div style="width: 100%; height: 200px; background: #ddd; display: flex; align-items: center; justify-content: center; color: #666; font-size: 12px;">
+                        Poster {i+1}
+                    </div>
+                    <div style="position: absolute; bottom: 0; left: 0; right: 0; background: linear-gradient(transparent, rgba(0,0,0,0.7)); color: white; padding: 8px; font-size: 12px; text-align: center;">
+                        Movie Title {i+1}
+                    </div>
+                </div>
+                """)
+            
+            html_parts.append("</div>")
+            return "".join(html_parts)
+            
+        except Exception as e:
+            return f"<div style='color: red; padding: 20px;'>❌ Error: {str(e)}</div>"
+        
     with gr.Blocks(
-        theme=gr.themes.Soft(), title="TV-Series and Movie Recommend"
+        theme=gr.themes.Soft(), 
+        title="TV-Series and Movie Recommend",
+        css="""
+        .gradio-container {
+            max-width: 1200px !important;
+        }
+        """
     ) as demo:
         gr.Markdown("# 🎬 TV-Series and Movie Recommend")
 
@@ -28,19 +86,31 @@ def create_interface(engine):
 
                 search_btn = gr.Button("🔍 Search", variant="primary")
 
-            with gr.Column(scale=2):
+            with gr.Column(scale=1):
                 results_text = gr.Textbox(
-                    label="Recommended Movies and TV-Series",
+                    label="Detailed Results",
                     lines=20,
                     max_lines=25,
                     show_copy_button=True,
                     interactive=False,
                 )
+                
+            with gr.Column(scale=1):
+                thumbnails_display = gr.HTML(
+                    label="Movie Posters",
+                    value="<div style='text-align: center; padding: 40px; color: #666;'>Movie thumbnails will appear here</div>"
+                )
 
         search_btn.click(
             fn=get_recommendations_text,
             inputs=[query_input],
             outputs=[results_text],
         )
+        
+        search_btn.click(
+            fn=get_thumbnails_from_results,
+            inputs=[query_input],
+            outputs=[thumbnails_display],
+        )
 
-    return demo
+    return demo

components/similarity.py

@@ -22,7 +22,7 @@ class SimilarityCalculator:
             }
 
         start_time = time.time()
-
+        print(f"🔍 Calculating similarity for query: {query}")
         query_embedding = self.model.encode([query])
         query_embedding = torch.tensor(query_embedding, dtype=torch.float32)
 
@@ -35,7 +35,7 @@ class SimilarityCalculator:
         similarities = similarities[0]
 
         hybrid_scores = self._calculate_hybrid_score(
-            similarities, filtered_data, similarity_weight=0.6, rating_weight=0.4
+            similarities, filtered_data, similarity_weight=0.8, rating_weight=0.2
         )
 
         top_indices = (
@@ -50,6 +50,7 @@ class SimilarityCalculator:
             row = filtered_data.iloc[idx]
 
             result = {
+                "tconst": row["tconst"],
                 "title": row["primaryTitle"],
                 "type": row["titleType"],
                 "year": row["startYear"],
@@ -82,8 +83,8 @@ class SimilarityCalculator:
         self,
         similarities: torch.Tensor,
         data: pd.DataFrame,
-        similarity_weight: float = 0.6,
-        rating_weight: float = 0.4,
+        similarity_weight: float = 0.8,
+        rating_weight: float = 0.2,
     ) -> torch.Tensor:
 
         sim_normalized = (similarities - similarities.min()) / (

components/tmdb_api.py

@@ -0,0 +1,54 @@
+import requests
+from config import Config
+
+class TMDBApi:
+    def __init__(self):
+        self.config = Config()
+        self.base_url = self.config.TMDB_BASE_URL
+        self.api_key = self.config.TMDB_API_KEY
+        self.image_base_url = self.config.TMDB_IMAGE_BASE_URL
+    
+    def get_poster_by_imdb_id(self, imdb_id: str):
+        try:
+            if not imdb_id.startswith('tt'):
+                imdb_id = f"tt{imdb_id}"
+            
+            endpoint = f"{self.base_url}/find/{imdb_id}"
+            params = {
+                "api_key": self.api_key,
+                "external_source": "tconst"
+            }
+            
+            response = requests.get(endpoint, params=params)
+            response.raise_for_status()
+            
+            data = response.json()
+            
+            poster_path = None
+            if data.get("movie_results"):
+                poster_path = data["movie_results"][0].get("poster_path")
+            elif data.get("tv_results"):
+                poster_path = data["tv_results"][0].get("poster_path")
+            
+            if poster_path:
+                return f"{self.image_base_url}{poster_path}"
+            return None
+            
+        except Exception as e:
+            print(f"❌ TMDB API Error for IMDB ID {imdb_id}: {str(e)}")
+            return None
+    
+    def get_multiple_posters_by_imdb(self, items: list):
+        results = []
+        for item in items:
+            imdb_id = item.get('tconst')
+            
+            if imdb_id:
+                poster_url = self.get_poster_by_imdb_id(imdb_id)
+                item['poster_url'] = poster_url
+            else:
+                item['poster_url'] = None
+                
+            results.append(item)
+            
+        return results

config.py

@@ -39,7 +39,9 @@ GENRE_LIST = Literal[
 
 class Config:
     OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
-
+    TMDB_API_KEY = os.getenv("TMDB_API_KEY")
+    TMDB_BASE_URL = "https://api.themoviedb.org/3"
+    TMDB_IMAGE_BASE_URL = "https://image.tmdb.org/t/p/w500"
     EMBEDDING_MODEL = "Qwen/Qwen3-Embedding-0.6B"
     DATA_FILE = "data/demo_data.parquet"
 

models/recommendation_engine.py

@@ -6,7 +6,7 @@ from models.pydantic_schemas import Features
 from components.similarity import SimilarityCalculator
 from components.filters import MovieFilter
 from sentence_transformers import SentenceTransformer
-
+from components.tmdb_api import TMDBApi
 
 class RecommendationEngine:
     def __init__(self):
@@ -19,6 +19,7 @@ class RecommendationEngine:
 
         self.similarity_calc = SimilarityCalculator(self.model)
         self.filter = MovieFilter()
+        self.tmdb_api = TMDBApi()
 
         print(f"✅ Recommendation engine initialized with {len(self.data)} items.")
 
@@ -33,8 +34,17 @@ class RecommendationEngine:
             filtered_data = self.filter.apply_filters(self.data, features)
 
             search_results = self.similarity_calc.calculate_similarity(
-                user_query, filtered_data, top_k
+                features.themes, filtered_data, top_k
             )
+            if search_results["results"]:
+                    print(f"🔍 First result keys: {search_results['results'][0].keys()}")
+                    
+                    for i, result in enumerate(search_results["results"]):
+                        print(f"🔍 Result {i}: tconst = {result.get('tconst', 'NOT FOUND')}")
+                        
+                    search_results["results"] = self.tmdb_api.get_multiple_posters_by_imdb(
+                        search_results["results"]
+                    )
 
             formatted_results = self._format_results(search_results)
 
@@ -44,14 +54,13 @@ class RecommendationEngine:
             return f"❌ Error: {str(e)}", None
 
     def _parse_user_query(self, query: str) -> Features:
-        """GPT ile kullanıcı sorgusu parse et"""
         try:
             response = self.client.beta.chat.completions.parse(
                 model="gpt-4o-mini",
                 messages=[
                     {
                         "role": "system",
-                        "content": "You are an AI that converts user requests into structured movie/TV-series features. Be smart about interpreting user preferences.",
+                        "content": "You are an AI that converts user requests into structured movie/TV-series features. ONLY extract genres that are explicitly mentioned by the user. Do not infer or add additional genres unless clearly stated.",
                     },
                     {"role": "user", "content": query},
                 ],
@@ -127,5 +136,5 @@ class RecommendationEngine:
                     "Overview": result["overview"],
                 }
             )
-
+        print(df_data)
         return pd.DataFrame(df_data)