Update app.py

app.py

@@ -376,21 +376,16 @@ class MistralRAGChatbot:
 
         if not context or not self.is_context_relevant(context, user_query):
             prompt = f"""You are an intelligent assistant.
-
     User Question:
     {user_query}
-
     Instruction:
     The document database does not contain relevant information to answer the question. Please inform the user that no relevant documents were found and refrain from generating an imaginative or unrelated response."""
         else:
             prompt = f"""You are an intelligent assistant.
-
     Context:
     {context}
-
     User Question:
     {user_query}
-
     Instruction:
     {style_instruction}"""
 
@@ -413,7 +408,7 @@ def create_vector_db_and_annoy_index(pdf_path, vector_db_path, annoy_index_path)
 # Cell 9: Run the store embeddings function (example)
 # Replace 'example.pdf' with your PDF file path.
 # It will create 'vector_db.pkl' and 'vector_index.ann'
-create_vector_db_and_annoy_index('med.pdf', 'vector_db.pkl', 'vector_index.ann')
+# create_vector_db_and_annoy_index('med.pdf', 'vector_db.pkl', 'vector_index.ann')
 
 # # Cell 10: Query the chatbot with user input
 # async def query_chatbot():
@@ -502,7 +497,7 @@ create_vector_db_and_annoy_index('med.pdf', 'vector_db.pkl', 'vector_index.ann')
 
 import gradio as gr
 
-def chatbot_interface(user_query, response_style, selected_retrieval_methods, selected_reranking_methods, chunk_size, overlap):
+def chatbot_interface(user_query, response_style, selected_retrieval_methods, selected_reranking_methods, file, chunk_size, overlap):
     vector_db_path = "vector_db.pkl"
     annoy_index_path = "vector_index.ann"
 
@@ -511,7 +506,7 @@ def chatbot_interface(user_query, response_style, selected_retrieval_methods, se
     
 
     #Load the documents and create embeddings with the provided chunk_size and overlap
-    store_embeddings_in_vector_db('med.pdf', 'vector_db.pkl', 'vector_index.ann', chunk_size, overlap)
+    store_embeddings_in_vector_db(file.name, 'vector_db.pkl', 'vector_index.ann', chunk_size, overlap)
 
     chatbot = MistralRAGChatbot(vector_db_path, annoy_index_path)
 
@@ -546,6 +541,7 @@ iface = gr.Interface(
         gr.Dropdown(["Detailed", "Concise", "Creative", "Technical"], label="Response Style"),
         gr.Dropdown(["annoy", "tfidf", "bm25", "euclidean", "jaccard"], label="Retrieval Methods", multiselect=True), # This line is changed
         gr.Dropdown(["advanced_fusion", "reciprocal_rank_fusion", "weighted_score_fusion", "semantic_similarity"], label="Reranking Methods"),
+        gr.File(label="Upload a PDF"),
         gr.Slider(minimum=1024, maximum=2048, step=128, value=2048, label="Chunk Size"),
         gr.Slider(minimum=100, maximum=300, step=100, value=200, label="Overlap")
     ],