final changes

README.md

@@ -164,6 +164,9 @@ The system follows a multi-stage pipeline that processes movie data and user cur
 
 ### Pipeline Stages
 
+The complete preprocessing pipeline code is available in our GitHub repository:
+🔗 **[Preprocessing Code](https://github.com/RecoFM/HF-hackathon2025)**
+
 #### 1. Preprocessing Stage
 - **Movie Title Processing**: Raw movie titles are converted into semantic embeddings using Mistral AI
 - **Graph Convolution Enhancement**: Movie embeddings are enriched through Graph Convolution Layers (GCL) that capture user interaction patterns and movie relationships

app.py

@@ -13,7 +13,6 @@ from ranking_agent import rank_with_ai
 from scipy.sparse import load_npz
 from rapidfuzz import process, fuzz
 import re
-from sklearn.metrics.pairwise import cosine_similarity
 
 load_dotenv()
 
@@ -269,7 +268,7 @@ def create_interface():
         Get personalized movie recommendations based on your taste and preferences utilizing zero-shot predicitons from Foundation Recommender!
         
         **How to use:**
-        1. Search and select movies you've enjoyed 
+        1. Search and select movies you've enjoyed
         2. Describe the type of film you are looking for. Consider factors such as genre, length, whether it is animated, etc.
         3. Adjust the preference weight (α) to balance between your description and movie history
         4. Get personalized recommendations