Update app.py

app.py

@@ -179,7 +179,7 @@ langchain_document_loader()
 
 text_splitter = RecursiveCharacterTextSplitter(
     separators = ["\n\n", "\n", " ", ""],    
-    chunk_size = 1500,
+    chunk_size = 1500,    # You could also use recursive, semantic, or document specific chunking techniques -- see https://medium.com/the-ai-forum/semantic-chunking-for-rag-f4733025d5f5
     chunk_overlap= 200
 )
 chunks = text_splitter.split_documents(documents=documents)
@@ -470,7 +470,7 @@ def instantiate_LLM(LLM_provider,api_key,temperature=0.8,top_p=0.95,model_name=N
 
 
 
-#  This creates history (memory) of prior questions. I am using Gemini for this but I left the code if I decide to go to GPT later on.
+#  This creates history (memory) of prior questions. The Website UI does this for you, but with API you have to do this on your own. I am using Gemini for this but I left the code if I decide to go to GPT later on.
 
 def create_memory(model_name='gemini-pro',memory_max_token=None):
 #def create_memory(model_name='gpt-3.5-turbo',memory_max_token=None):
@@ -497,7 +497,7 @@ def create_memory(model_name='gemini-pro',memory_max_token=None):
         )  
     return memory
 
-# Set a small memory_max_token, just to show how older messages are summarized if max_token_limit is exceeded.
+# You can set a small memory_max_token, just to show how older messages are summarized if max_token_limit is exceeded.
 
 memory = create_memory(model_name='gemini-pro',memory_max_token=None)
 #memory = create_memory(model_name='gpt-3.5-turbo',memory_max_token=20)