From 8921deb6ffb0e36350760605798e2362bc6852cc Mon Sep 17 00:00:00 2001
From: alxspiker <alxtheprogrammer@gmail.com>
Date: Fri, 12 May 2023 23:44:02 -0600
Subject: [PATCH 1/8] GUI SUPPORT

First working edition!
Ill continue adding to this PR to make it more complete.

I want to see if I can get an ingest page to ingest using the GUI too.

Run `streamlit run .\gui.py` to use it.
---
 gui.py           | 63 ++++++++++++++++++++++++++++++++++++++++++++++++
 requirements.txt |  3 +++
 startLLM.py      |  7 ++++--
 3 files changed, 71 insertions(+), 2 deletions(-)
 create mode 100644 gui.py

diff --git a/gui.py b/gui.py
new file mode 100644
index 0000000..f6253a1
--- /dev/null
+++ b/gui.py
@@ -0,0 +1,63 @@
+import streamlit as st
+from streamlit_chat import message
+from streamlit_extras.colored_header import colored_header
+from streamlit_extras.add_vertical_space import add_vertical_space
+
+st.set_page_config(page_title="CASALIOY")
+
+# Sidebar contents
+with st.sidebar:
+    st.title('CASALIOY')
+    st.markdown('''
+    ## About
+    This app is an LLM-powered chatbot built using:
+    - [Streamlit](https://streamlit.io/)
+    - [su77ungr/CASALIOY](https://github.com/alxspiker/CASALIOY) LLM Toolkit
+    
+    💡 Note: No API key required!
+
+    GUI does not support live response yet, so you have to wait for the tokens to process.
+    ''')
+    add_vertical_space(5)
+    st.write('Made with ❤️ by [su77ungr/CASALIOY](https://github.com/alxspiker/CASALIOY)')
+
+# Generate empty lists for generated and past.
+## generated stores AI generated responses
+if 'generated' not in st.session_state:
+    st.session_state['generated'] = ["I can help you answer questions about the documents you have ingested into the vector store."]
+## past stores User's questions
+if 'past' not in st.session_state:
+    st.session_state['past'] = ['Hi, what can you help me with!']
+
+# Layout of input/response containers
+input_container = st.container()
+colored_header(label='', description='', color_name='blue-30')
+response_container = st.container()
+
+# User input
+## Function for taking user provided prompt as input
+def get_text():
+    input_text = st.text_input("You: ", "", key="input")
+    return input_text
+## Applying the user input box
+with input_container:
+    user_input = get_text()
+
+# Response output
+## Function for taking user prompt as input followed by producing AI generated responses
+def generate_response(prompt):
+    import startLLM
+    response = startLLM.main(prompt, True)
+    return response
+
+## Conditional display of AI generated responses as a function of user provided prompts
+with response_container:
+    if user_input:
+        response = generate_response(user_input)
+        st.session_state.past.append(user_input)
+        st.session_state.generated.append(response)
+        
+    if st.session_state['generated']:
+        for i in range(len(st.session_state['generated'])):
+            message(st.session_state['past'][i], is_user=True, key=str(i) + '_user')
+            message(st.session_state["generated"][i], key=str(i))
\ No newline at end of file
diff --git a/requirements.txt b/requirements.txt
index 344b43a..96f6850 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -4,3 +4,6 @@ qdrant-client==1.1.7
 llama-cpp-python==0.1.49
 pdfminer.six==20221105
 python-dotenv==1.0.0
+steamlit
+streamlit-chat
+streamlit-extras
\ No newline at end of file
diff --git a/startLLM.py b/startLLM.py
index c71dee5..dfb6001 100644
--- a/startLLM.py
+++ b/startLLM.py
@@ -15,7 +15,7 @@
 model_temp = os.environ.get('MODEL_TEMP')
 model_stop = os.environ.get('MODEL_STOP').split(",")
 
-def main():
+def main(prompt="", gui=False):
     # Load stored vectorstore
     llama = LlamaCppEmbeddings(model_path=llama_embeddings_model, n_ctx=model_n_ctx)
     # Load ggml-formatted model 
@@ -44,7 +44,7 @@ def main():
 
     # Interactive questions and answers
     while True:
-        query = input("\nEnter a query: ")
+        query = prompt if prompt.strip() != "" else input("\nEnter a query: ")
         if query == "exit":
             break
         
@@ -62,6 +62,9 @@ def main():
         for document in docs:
             print("\n> " + document.metadata["source"] + ":")
             print(document.page_content)
+        
+        if gui:
+            return answer
 
 if __name__ == "__main__":
     main()

From 8f40781eb9879c007e75e2988f62be9d05077e7c Mon Sep 17 00:00:00 2001
From: alxspiker <alxtheprogrammer@gmail.com>
Date: Sat, 13 May 2023 00:47:26 -0600
Subject: [PATCH 2/8] Preload models only once instead of every query.

Text input doesnt disable like I tried to do after a query.
---
 gui.py      | 19 +++++++++++++++---
 startLLM.py | 56 ++++++++++++++++++++++++++++++++---------------------
 2 files changed, 50 insertions(+), 25 deletions(-)

diff --git a/gui.py b/gui.py
index f6253a1..0e45ca5 100644
--- a/gui.py
+++ b/gui.py
@@ -2,8 +2,10 @@
 from streamlit_chat import message
 from streamlit_extras.colored_header import colored_header
 from streamlit_extras.add_vertical_space import add_vertical_space
+from startLLM import main as startLLM
 
 st.set_page_config(page_title="CASALIOY")
+input_text = None
 
 # Sidebar contents
 with st.sidebar:
@@ -34,10 +36,20 @@
 colored_header(label='', description='', color_name='blue-30')
 response_container = st.container()
 
+if "disabled" not in st.session_state:
+    st.session_state["disabled"] = False
+
+def enable():
+    st.session_state["disabled"] = False
+
+def disable():
+    st.session_state["disabled"] = True
+
 # User input
 ## Function for taking user provided prompt as input
 def get_text():
-    input_text = st.text_input("You: ", "", key="input")
+    global input_text
+    input_text = st.text_input("You: ", "", key="input", disabled=st.session_state["disabled"])
     return input_text
 ## Applying the user input box
 with input_container:
@@ -46,8 +58,9 @@ def get_text():
 # Response output
 ## Function for taking user prompt as input followed by producing AI generated responses
 def generate_response(prompt):
-    import startLLM
-    response = startLLM.main(prompt, True)
+    #input_text.disabled = True
+    disable()
+    response = startLLM(prompt, True)
     return response
 
 ## Conditional display of AI generated responses as a function of user provided prompts
diff --git a/startLLM.py b/startLLM.py
index dfb6001..4a36e7b 100644
--- a/startLLM.py
+++ b/startLLM.py
@@ -15,7 +15,11 @@
 model_temp = os.environ.get('MODEL_TEMP')
 model_stop = os.environ.get('MODEL_STOP').split(",")
 
-def main(prompt="", gui=False):
+qa_system=None
+llm=None
+qdrant=None
+
+def initialize_qa_system():
     # Load stored vectorstore
     llama = LlamaCppEmbeddings(model_path=llama_embeddings_model, n_ctx=model_n_ctx)
     # Load ggml-formatted model 
@@ -24,6 +28,7 @@ def main(prompt="", gui=False):
     client = qdrant_client.QdrantClient(
     path=persist_directory, prefer_grpc=True
     )
+    global qdrant
     qdrant = Qdrant(
         client=client, collection_name="test", 
         embeddings=llama
@@ -31,6 +36,7 @@ def main(prompt="", gui=False):
 
     # Prepare the LLM chain 
     callbacks = [StreamingStdOutCallbackHandler()]
+    global llm
     match model_type:
         case "LlamaCpp":
             from langchain.llms import LlamaCpp
@@ -41,30 +47,36 @@ def main(prompt="", gui=False):
         case _default:
             print("Only LlamaCpp or GPT4All supported right now. Make sure you set up your .env correctly.")
     qa = RetrievalQA.from_chain_type(llm=llm, chain_type="stuff", retriever=qdrant.as_retriever(search_type="mmr"), return_source_documents=True)
+    return qa
 
+def main(prompt="", gui=False):
+    global qa_system
+    if qa_system is None:
+        qa_system = initialize_qa_system()
     # Interactive questions and answers
-    while True:
-        query = prompt if prompt.strip() != "" else input("\nEnter a query: ")
-        if query == "exit":
-            break
-        
-        # Get the answer from the chain
-        res = qa(query)    
-        answer, docs = res['result'], res['source_documents']
+    if prompt.strip() != "":
+        while True:
+            query = prompt if prompt.strip() != "" else input("\nEnter a query: ")
+            if query == "exit":
+                break
+            
+            # Get the answer from the chain
+            res = qa_system(query)    
+            answer, docs = res['result'], res['source_documents']
 
-        # Print the result
-        print("\n\n> Question:")
-        print(query)
-        print("\n> Answer:")
-        print(answer)
-        
-        # Print the relevant sources used for the answer
-        for document in docs:
-            print("\n> " + document.metadata["source"] + ":")
-            print(document.page_content)
-        
-        if gui:
-            return answer
+            # Print the result
+            print("\n\n> Question:")
+            print(query)
+            print("\n> Answer:")
+            print(answer)
+            
+            # Print the relevant sources used for the answer
+            for document in docs:
+                print("\n> " + document.metadata["source"] + ":")
+                print(document.page_content)
+            
+            if gui:
+                return answer
 
 if __name__ == "__main__":
     main()

From d5499e6d30b501e600786e6daea25859b0d573ea Mon Sep 17 00:00:00 2001
From: alxspiker <alxtheprogrammer@gmail.com>
Date: Sat, 13 May 2023 13:21:37 -0600
Subject: [PATCH 3/8] Preload models + better interface

Still needs some work.
---
 gui.py | 59 +++++++++++++++++++++++++++++-----------------------------
 1 file changed, 29 insertions(+), 30 deletions(-)

diff --git a/gui.py b/gui.py
index 0e45ca5..250170b 100644
--- a/gui.py
+++ b/gui.py
@@ -4,8 +4,12 @@
 from streamlit_extras.add_vertical_space import add_vertical_space
 from startLLM import main as startLLM
 
+# Initialization
+if "visibility" not in st.session_state:
+    st.session_state.visibility = "visible"
+    st.session_state.running = False
+
 st.set_page_config(page_title="CASALIOY")
-input_text = None
 
 # Sidebar contents
 with st.sidebar:
@@ -36,41 +40,36 @@
 colored_header(label='', description='', color_name='blue-30')
 response_container = st.container()
 
-if "disabled" not in st.session_state:
-    st.session_state["disabled"] = False
-
-def enable():
-    st.session_state["disabled"] = False
-
-def disable():
-    st.session_state["disabled"] = True
-
 # User input
 ## Function for taking user provided prompt as input
-def get_text():
-    global input_text
-    input_text = st.text_input("You: ", "", key="input", disabled=st.session_state["disabled"])
-    return input_text
+#def get_text():
+    #return input_text
+
 ## Applying the user input box
-with input_container:
-    user_input = get_text()
+#with input_container:
+    #user_input = get_text()
 
 # Response output
 ## Function for taking user prompt as input followed by producing AI generated responses
 def generate_response(prompt):
-    #input_text.disabled = True
-    disable()
-    response = startLLM(prompt, True)
-    return response
+    if prompt.strip() != "":
+        st.session_state.running = True
+        response = startLLM(prompt, True)
+        st.session_state.running = False
+        st.session_state.past.append(prompt)
+        st.session_state.generated.append(response)
+        return response
 
+with st.form("my_form", clear_on_submit=True):
 ## Conditional display of AI generated responses as a function of user provided prompts
-with response_container:
-    if user_input:
-        response = generate_response(user_input)
-        st.session_state.past.append(user_input)
-        st.session_state.generated.append(response)
-        
-    if st.session_state['generated']:
-        for i in range(len(st.session_state['generated'])):
-            message(st.session_state['past'][i], is_user=True, key=str(i) + '_user')
-            message(st.session_state["generated"][i], key=str(i))
\ No newline at end of file
+    user_input = st.text_input("You: ", "", key="input", disabled=st.session_state.running)
+    st.form_submit_button('SUBMIT', on_click=generate_response(user_input), disabled=st.session_state.running)
+#if user_input:
+    #response = generate_response(user_input)
+    #st.session_state.past.append(user_input)
+    #st.session_state.generated.append(response)
+    
+if st.session_state['generated']:
+    for i in range(len(st.session_state['generated'])):
+        message(st.session_state['past'][i], is_user=True, key=str(i) + '_user')
+        message(st.session_state["generated"][i], key=str(i))
\ No newline at end of file

From 004cc797f544cb3383f225b2092d8a05410a98b7 Mon Sep 17 00:00:00 2001
From: su77ungr <69374354+su77ungr@users.noreply.github.com>
Date: Sat, 13 May 2023 22:03:42 +0200
Subject: [PATCH 4/8] specified version of streamlit else None

---
 requirements.txt | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/requirements.txt b/requirements.txt
index 50254b2..3734f00 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -4,8 +4,8 @@ qdrant-client==1.1.7
 llama-cpp-python==0.1.49
 pdfminer.six==20221105
 python-dotenv==1.0.0
-steamlit
-streamlit-chat
-streamlit-extras
+streamlit==1.22.0 
+streamlit-chat==0.0.2.2 
+streamlit-extras==0.2.7 
 pandoc==2.3
 unstructured==0.6.6

From c8bbe2b40f22c0b61fb98d1946a5850c58624c88 Mon Sep 17 00:00:00 2001
From: alxspiker <alxtheprogrammer@gmail.com>
Date: Sat, 13 May 2023 14:50:02 -0600
Subject: [PATCH 5/8] Working version: Added loading message and improved UI

Still need to do cleanup, but should perform properly.
---
 gui.py | 37 +++++++++++++++++++++----------------
 1 file changed, 21 insertions(+), 16 deletions(-)

diff --git a/gui.py b/gui.py
index 250170b..e6b451e 100644
--- a/gui.py
+++ b/gui.py
@@ -5,8 +5,8 @@
 from startLLM import main as startLLM
 
 # Initialization
-if "visibility" not in st.session_state:
-    st.session_state.visibility = "visible"
+if "input" not in st.session_state:
+    st.session_state.input = ""
     st.session_state.running = False
 
 st.set_page_config(page_title="CASALIOY")
@@ -52,24 +52,29 @@
 # Response output
 ## Function for taking user prompt as input followed by producing AI generated responses
 def generate_response(prompt):
-    if prompt.strip() != "":
+    if st.session_state['generated']:
+        for i in range(len(st.session_state['generated'])):
+            message(st.session_state["past"][i], is_user=True, key=str(i) + '_user')
+            message(st.session_state["generated"][i], key=str(i))
+    #if prompt.strip() != "":
+    
+    if st.session_state.running==False and prompt.strip() != "":
         st.session_state.running = True
-        response = startLLM(prompt, True)
-        st.session_state.running = False
         st.session_state.past.append(prompt)
+        message(prompt, is_user=True)
+        message("Loading response. Please wait...", key="rmessage")
+        response = startLLM(prompt, True)
         st.session_state.generated.append(response)
-        return response
-
-with st.form("my_form", clear_on_submit=True):
-## Conditional display of AI generated responses as a function of user provided prompts
-    user_input = st.text_input("You: ", "", key="input", disabled=st.session_state.running)
-    st.form_submit_button('SUBMIT', on_click=generate_response(user_input), disabled=st.session_state.running)
+        message(response)
+        st.session_state.running = False
+    #st.session_state.rmessage = response
+    #return response
+    st.text_input("You: ", "", key="input", disabled=st.session_state.running)
 #if user_input:
     #response = generate_response(user_input)
     #st.session_state.past.append(user_input)
     #st.session_state.generated.append(response)
-    
-if st.session_state['generated']:
-    for i in range(len(st.session_state['generated'])):
-        message(st.session_state['past'][i], is_user=True, key=str(i) + '_user')
-        message(st.session_state["generated"][i], key=str(i))
\ No newline at end of file
+with st.container():
+    with st.form("my_form", clear_on_submit=True):
+    ## Conditional display of AI generated responses as a function of user provided prompts
+        st.form_submit_button('SUBMIT', on_click=generate_response(st.session_state.input), disabled=st.session_state.running)
\ No newline at end of file

From 4a146ce5b03041b53260e61b9e32edab246e05e3 Mon Sep 17 00:00:00 2001
From: alxspiker <alxtheprogrammer@gmail.com>
Date: Sat, 13 May 2023 14:53:58 -0600
Subject: [PATCH 6/8] Code cleanup for commit

---
 gui.py | 26 ++------------------------
 1 file changed, 2 insertions(+), 24 deletions(-)

diff --git a/gui.py b/gui.py
index e6b451e..9807cd1 100644
--- a/gui.py
+++ b/gui.py
@@ -27,37 +27,21 @@
     add_vertical_space(5)
     st.write('Made with ❤️ by [su77ungr/CASALIOY](https://github.com/alxspiker/CASALIOY)')
 
-# Generate empty lists for generated and past.
-## generated stores AI generated responses
 if 'generated' not in st.session_state:
     st.session_state['generated'] = ["I can help you answer questions about the documents you have ingested into the vector store."]
-## past stores User's questions
+
 if 'past' not in st.session_state:
     st.session_state['past'] = ['Hi, what can you help me with!']
 
-# Layout of input/response containers
 input_container = st.container()
 colored_header(label='', description='', color_name='blue-30')
 response_container = st.container()
 
-# User input
-## Function for taking user provided prompt as input
-#def get_text():
-    #return input_text
-
-## Applying the user input box
-#with input_container:
-    #user_input = get_text()
-
-# Response output
-## Function for taking user prompt as input followed by producing AI generated responses
 def generate_response(prompt):
     if st.session_state['generated']:
         for i in range(len(st.session_state['generated'])):
             message(st.session_state["past"][i], is_user=True, key=str(i) + '_user')
             message(st.session_state["generated"][i], key=str(i))
-    #if prompt.strip() != "":
-    
     if st.session_state.running==False and prompt.strip() != "":
         st.session_state.running = True
         st.session_state.past.append(prompt)
@@ -67,14 +51,8 @@ def generate_response(prompt):
         st.session_state.generated.append(response)
         message(response)
         st.session_state.running = False
-    #st.session_state.rmessage = response
-    #return response
     st.text_input("You: ", "", key="input", disabled=st.session_state.running)
-#if user_input:
-    #response = generate_response(user_input)
-    #st.session_state.past.append(user_input)
-    #st.session_state.generated.append(response)
+
 with st.container():
     with st.form("my_form", clear_on_submit=True):
-    ## Conditional display of AI generated responses as a function of user provided prompts
         st.form_submit_button('SUBMIT', on_click=generate_response(st.session_state.input), disabled=st.session_state.running)
\ No newline at end of file

From 98f38841cbee99403c4d3f2d673233a255fc8585 Mon Sep 17 00:00:00 2001
From: alxspiker <alxtheprogrammer@gmail.com>
Date: Sat, 13 May 2023 15:34:30 -0600
Subject: [PATCH 7/8] Faster responses + GUI | Run: streamlit run .\gui.py

---
 gui.py      | 7 +++++--
 startLLM.py | 8 ++------
 2 files changed, 7 insertions(+), 8 deletions(-)

diff --git a/gui.py b/gui.py
index 9807cd1..92a9452 100644
--- a/gui.py
+++ b/gui.py
@@ -2,7 +2,7 @@
 from streamlit_chat import message
 from streamlit_extras.colored_header import colored_header
 from streamlit_extras.add_vertical_space import add_vertical_space
-from startLLM import main as startLLM
+import startLLM
 
 # Initialization
 if "input" not in st.session_state:
@@ -21,6 +21,8 @@
     - [su77ungr/CASALIOY](https://github.com/alxspiker/CASALIOY) LLM Toolkit
     
     💡 Note: No API key required!
+    Refreshing the page will restart gui.py with a fresh chat history.
+    CASALIOY will not remember previous questions as of yet.
 
     GUI does not support live response yet, so you have to wait for the tokens to process.
     ''')
@@ -47,7 +49,8 @@ def generate_response(prompt):
         st.session_state.past.append(prompt)
         message(prompt, is_user=True)
         message("Loading response. Please wait...", key="rmessage")
-        response = startLLM(prompt, True)
+        response = startLLM.main(prompt, True)
+        #startLLM.qdrant = None
         st.session_state.generated.append(response)
         message(response)
         st.session_state.running = False
diff --git a/startLLM.py b/startLLM.py
index 4a36e7b..aabfc81 100644
--- a/startLLM.py
+++ b/startLLM.py
@@ -16,8 +16,6 @@
 model_stop = os.environ.get('MODEL_STOP').split(",")
 
 qa_system=None
-llm=None
-qdrant=None
 
 def initialize_qa_system():
     # Load stored vectorstore
@@ -28,7 +26,6 @@ def initialize_qa_system():
     client = qdrant_client.QdrantClient(
     path=persist_directory, prefer_grpc=True
     )
-    global qdrant
     qdrant = Qdrant(
         client=client, collection_name="test", 
         embeddings=llama
@@ -36,7 +33,6 @@ def initialize_qa_system():
 
     # Prepare the LLM chain 
     callbacks = [StreamingStdOutCallbackHandler()]
-    global llm
     match model_type:
         case "LlamaCpp":
             from langchain.llms import LlamaCpp
@@ -54,9 +50,9 @@ def main(prompt="", gui=False):
     if qa_system is None:
         qa_system = initialize_qa_system()
     # Interactive questions and answers
-    if prompt.strip() != "":
+    if (prompt.strip() != "" and gui) or gui==False:
         while True:
-            query = prompt if prompt.strip() != "" else input("\nEnter a query: ")
+            query = prompt if gui else input("\nEnter a query: ")
             if query == "exit":
                 break
             

From 3fa8687efb148a090ea1bfe924c03643986e0d9b Mon Sep 17 00:00:00 2001
From: alxspiker <alxtheprogrammer@gmail.com>
Date: Sat, 13 May 2023 17:49:37 -0600
Subject: [PATCH 8/8] Edit .env directly in GUI

---
 gui.py      | 72 +++++++++++++++++++++++++++++++++++++----------------
 startLLM.py |  4 +--
 2 files changed, 53 insertions(+), 23 deletions(-)

diff --git a/gui.py b/gui.py
index 92a9452..fd3e2a0 100644
--- a/gui.py
+++ b/gui.py
@@ -1,8 +1,20 @@
+import dotenv
 import streamlit as st
 from streamlit_chat import message
 from streamlit_extras.colored_header import colored_header
 from streamlit_extras.add_vertical_space import add_vertical_space
 import startLLM
+import os
+
+dotenv_file = dotenv.find_dotenv(".env")
+dotenv.load_dotenv()
+llama_embeddings_model = os.environ.get("LLAMA_EMBEDDINGS_MODEL")
+persist_directory = os.environ.get('PERSIST_DIRECTORY')
+model_type = os.environ.get('MODEL_TYPE')
+model_path = os.environ.get('MODEL_PATH')
+model_n_ctx = int(os.environ.get('MODEL_N_CTX'))
+model_temp = float(os.environ.get('MODEL_TEMP'))
+model_stop = os.environ.get('MODEL_STOP')
 
 # Initialization
 if "input" not in st.session_state:
@@ -35,27 +47,45 @@
 if 'past' not in st.session_state:
     st.session_state['past'] = ['Hi, what can you help me with!']
 
-input_container = st.container()
 colored_header(label='', description='', color_name='blue-30')
 response_container = st.container()
 
-def generate_response(prompt):
-    if st.session_state['generated']:
-        for i in range(len(st.session_state['generated'])):
-            message(st.session_state["past"][i], is_user=True, key=str(i) + '_user')
-            message(st.session_state["generated"][i], key=str(i))
-    if st.session_state.running==False and prompt.strip() != "":
-        st.session_state.running = True
-        st.session_state.past.append(prompt)
-        message(prompt, is_user=True)
-        message("Loading response. Please wait...", key="rmessage")
-        response = startLLM.main(prompt, True)
-        #startLLM.qdrant = None
-        st.session_state.generated.append(response)
-        message(response)
-        st.session_state.running = False
-    st.text_input("You: ", "", key="input", disabled=st.session_state.running)
-
-with st.container():
-    with st.form("my_form", clear_on_submit=True):
-        st.form_submit_button('SUBMIT', on_click=generate_response(st.session_state.input), disabled=st.session_state.running)
\ No newline at end of file
+
+            
+def generate_response(input=""):
+    with response_container:
+        col1, col2, col3 = st.columns(3)
+        with col1:
+            if st.number_input('Temperature', key="temp_input", value=float(model_temp), step=float(0.05), min_value=float(0), max_value=float(1)):
+                os.environ["MODEL_TEMP"] = str(st.session_state.temp_input)
+                dotenv.set_key(dotenv_file, "MODEL_TEMP", os.environ["MODEL_TEMP"])
+        with col2:
+            if st.number_input('Context', key="ctx_input", value=int(model_n_ctx), step=int(512), min_value=int(512), max_value=int(9000)):
+                os.environ["MODEL_N_CTX"] = str(st.session_state.ctx_input)
+                dotenv.set_key(dotenv_file, "MODEL_N_CTX", os.environ["MODEL_N_CTX"])
+        with col3:
+            if st.text_input('Stops', key="stops_input", value=str(model_stop)):
+                os.environ["MODEL_STOP"] = str(st.session_state.stops_input)
+                dotenv.set_key(dotenv_file, "MODEL_STOP", os.environ["MODEL_STOP"])
+        #with st.form("my_form", clear_on_submit=True):
+        if st.session_state['generated']:
+            for i in range(len(st.session_state['generated'])):
+                message(st.session_state["past"][i], is_user=True, key=str(i) + '_user')
+                message(st.session_state["generated"][i], key=str(i))
+        if input.strip() != "":
+            st.session_state.running=True
+            st.session_state.past.append(st.session_state.input)
+            if st.session_state.running:
+                message(st.session_state.input, is_user=True)
+                message("Loading response. Please wait for me to finish before refreshing the page...", key="rmessage")
+                #startLLM.qdrant = None #Not sure why this fixes db error
+                response = startLLM.main(st.session_state.input, True)
+                st.session_state.generated.append(response)
+                message(response)
+                st.session_state.running = False
+        st.text_input("You: ", "", key="input", disabled=st.session_state.running)
+
+
+with st.form("my_form", clear_on_submit=True):
+    st.form_submit_button('SUBMIT', on_click=generate_response(st.session_state.input), disabled=st.session_state.running)
+        
\ No newline at end of file
diff --git a/startLLM.py b/startLLM.py
index aabfc81..503a1b2 100644
--- a/startLLM.py
+++ b/startLLM.py
@@ -11,8 +11,8 @@
 persist_directory = os.environ.get('PERSIST_DIRECTORY')
 model_type = os.environ.get('MODEL_TYPE')
 model_path = os.environ.get('MODEL_PATH')
-model_n_ctx = os.environ.get('MODEL_N_CTX')
-model_temp = os.environ.get('MODEL_TEMP')
+model_n_ctx = int(os.environ.get('MODEL_N_CTX'))
+model_temp = float(os.environ.get('MODEL_TEMP'))
 model_stop = os.environ.get('MODEL_STOP').split(",")
 
 qa_system=None