refactoring for azure and langsmith

app.py

@@ -21,9 +21,11 @@ GPT_MODEL = "gpt-4o-mini"
 # Used for Langsmith
 unique_id = uuid4().hex[0:8]
 os.environ["LANGCHAIN_TRACING_V2"] = "true"
-os.environ["LANGCHAIN_PROJECT"] = f"LangSmith LCEL RAG - {unique_id}"
+if os.environ.get("LANGCHAIN_PROJECT") is None:
+    os.environ["LANGCHAIN_PROJECT"] = f"LangSmith LCEL RAG - {unique_id}"
 
 is_azure = False if os.environ.get("AZURE_DEPLOYMENT") is None else True
+is_azure_qdrant_inmem = True if os.environ.get("AZURE_QDRANT_INMEM") else False
 
 # Utility functions
 def save_file(file: AskFileResponse,file_ext:str,is_azure:bool) -> str:
@@ -33,7 +35,7 @@ def save_file(file: AskFileResponse,file_ext:str,is_azure:bool) -> str:
         file_ext = ".txt"
     else:
         raise ValueError(f"Unknown file type: {file_ext}")
-    dir = "/tmp" if is_azure else None
+    dir = "/tmp" if is_azure_qdrant_inmem else None
     with tempfile.NamedTemporaryFile(
         mode="wb", delete=False, suffix=file_ext,dir=dir
     ) as temp_file:
@@ -42,6 +44,28 @@ def save_file(file: AskFileResponse,file_ext:str,is_azure:bool) -> str:
     return temp_file_path
 
 
+def setup_vectorstore(documents: List[str], embedding_model: OpenAIEmbeddings,is_azure:bool) -> Qdrant:
+    if is_azure:
+        if is_azure_qdrant_inmem: 
+            qdrant_vectorstore = Qdrant.from_documents(
+                documents=documents,
+                embedding=embedding_model,
+                location=":memory:"
+            )
+        else:
+            qdrant_vectorstore = Qdrant.from_documents(
+                documents=documents,
+                embedding=embedding_model,
+                url="http://qdrant:6333", # Docker compose setup
+            )
+    else:
+        qdrant_vectorstore = Qdrant.from_documents(
+            documents=documents,
+            embedding=embedding_model,
+            location=":memory:"
+        )
+    return qdrant_vectorstore
+
 # Prepare the components that will form the chain
 
 ## Step 1: Create a prompt template
@@ -83,20 +107,20 @@ async def on_chat_start():
         ).send()
     
     ## Load file and split into chunks
-    msg = cl.Message(content=f"Processing `{files[0].name}`...")
-    await msg.send()
+    await cl.Message(content=f"Processing `{files[0].name}`...").send()
 
     current_file_path = save_file(files[0], files[0].type,is_azure)
     loader_splitter = TextLoaderAndSplitterWrapper(ChunkingStrategy.RECURSIVE_CHARACTER_CHAR_SPLITTER, current_file_path)
     documents = loader_splitter.load_documents() 
 
+    await cl.Message(content="    Data Chunked...").send()
+
     ## Vectorising the documents
-    qdrant_vectorstore = Qdrant.from_documents(
-        documents=documents,
-        embedding=embedding_model,
-        location=":memory:"
-    )
+    
+    qdrant_vectorstore = setup_vectorstore(documents, embedding_model,is_azure)
+    
     qdrant_retriever = qdrant_vectorstore.as_retriever()
+    await cl.Message(content="    Created Vector store").send()
 
     # create the chain on new chart session
     retrieval_augmented_qa_chain = (

docker-compose.yml

@@ -0,0 +1,38 @@
+version: '3.8'
+
+services:
+  # Your Python Application Service
+  app:
+    build:
+      context: .
+      dockerfile: Dockerfile  # Assuming your Dockerfile is named Dockerfile
+    container_name: app
+    user: "user"  # Matching the user created in your Dockerfile
+    ports:
+      - "7860:7860"  # Expose your application's port
+    environment:
+      - HOME=/home/user
+      - PATH=/home/user/.local/bin:$PATH
+      - AZURE_DEPLOYMENT=true
+      - AZURE_QDRANT_INMEM=true  # False means use Qdrant service from the network
+    depends_on:
+      - qdrant  # Ensure Qdrant starts before this service
+    volumes:
+      - .:/home/user/app  # Mount current directory to container
+
+  # Qdrant Service
+  qdrant:
+    image: qdrant/qdrant:latest
+    restart: always
+    container_name: qdrant
+    ports:
+      - "6333:6333"
+      - "6334:6334"
+    expose:
+      - "6333"
+      - "6334"
+      - "6335"
+    # volumes:
+    #   - ./qdrant_data:/qdrant/storage  # Persist Qdrant data
+    #   - ./qdrant_config/production.yaml:/qdrant/config/production.yaml  # Mount configuration file
+