update chainlit

Dockerfile

@@ -8,7 +8,8 @@ USER user
 
 # Set the home directory and path
 ENV HOME=/home/user \
-    PATH=/home/user/.local/bin:$PATH        
+    PATH=/home/user/.local/bin:$PATH \
+    CHAINLIT_CONFIG_DIR=/home/user/app
 
 ENV UVICORN_WS_PROTOCOL=websockets
 
@@ -18,6 +19,9 @@ WORKDIR $HOME/app
 # Copy the app to the container
 COPY --chown=user . $HOME/app
 
+# Ensure proper permissions
+RUN chmod -R 755 $HOME/app
+
 # Install the dependencies
 # RUN uv sync --frozen
 RUN uv sync

app.py

@@ -43,7 +43,7 @@ HF_TOKEN = os.environ["HF_TOKEN"]
 document_loader = TextLoader("./data/paul_graham_essays.txt")
 documents = document_loader.load()
 
-text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=30)
+text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50)
 split_documents = text_splitter.split_documents(documents)
 
 hf_embeddings = HuggingFaceEndpointEmbeddings(
@@ -82,7 +82,7 @@ async def main():
             pbars.append(pbar)
             
             if i == 0:
-                vectorstore = await process_batch(None, batch, True, pbar)
+                vectorstore = await FAISS.afrom_documents(batch, hf_embeddings)
             else:
                 tasks.append(process_batch(vectorstore, batch, False, pbar))
         
@@ -94,7 +94,15 @@ async def main():
     
     await process_all_batches()
     
-    hf_retriever = vectorstore.as_retriever()
+    # Configure retriever with search parameters
+    hf_retriever = vectorstore.as_retriever(
+        search_kwargs={
+            "k": 3,  # Number of documents to retrieve
+            "fetch_k": 5,  # Number of documents to fetch before filtering
+            "maximal_marginal_relevance": True,  # Use MMR to ensure diversity
+            "filter": None  # No filtering
+        }
+    )
     print("\nIndexing complete. Vectorstore is ready for use.")
     return hf_retriever
 
@@ -111,7 +119,7 @@ hf_retriever = asyncio.run(run())
 """
 RAG_PROMPT_TEMPLATE = """\
 <|start_header_id|>system<|end_header_id|>
-You are a helpful assistant. You answer user questions based on provided context. If you can't answer the question with the provided context, say you don't know.<|eot_id|>
+You are a helpful assistant. You answer user questions based on provided context. If you can't answer the question with the provided context, say you don't know. Keep your responses concise and focused.<|eot_id|>
 
 <|start_header_id|>user<|end_header_id|>
 User Query:
@@ -131,7 +139,7 @@ rag_prompt = PromptTemplate.from_template(RAG_PROMPT_TEMPLATE)
 """
 hf_llm = HuggingFaceEndpoint(
     endpoint_url=HF_LLM_ENDPOINT,
-    max_new_tokens=512,
+    max_new_tokens=256,
     top_k=10,
     top_p=0.95,
     temperature=0.3,