Update app.py

app.py

@@ -102,30 +102,25 @@ def cleanup_expired_files():
 
 ## Context Taking, PDF Upload, and Mode Selection
 with st.sidebar:
-    st.title("Select Mode:")
-    option = st.selectbox(
-        'Choose your interaction mode',
-        ('Chat PDF', 'Chat LLM')
-    )
+    st.title("Upload PDF:")
     
-    if option == "Chat PDF":
-        st.title("Upload PDF:")
-        research_field = st.text_input("Research Field: ", key="research_field", placeholder="Enter research fields with commas")
-        if not research_field:
-            st.info("Please enter a research field to proceed.")
-            uploaded_file = st.file_uploader("", type=["pdf"], disabled=True)
-        else:
-            uploaded_file = st.file_uploader("", type=["pdf"], disabled=False)
+    research_field = st.text_input("Research Field: ", key="research_field", placeholder="Enter research fields with commas")
+    option = ''
+
+    if not research_field:
+        st.info("Please enter a research field to proceed.")
+        option = st.selectbox('Select Mode', ('Chat', 'Code'), disabled=True)
+        uploaded_file = st.file_uploader("", type=["pdf"], disabled=True)
     else:
-        research_field = None
-        uploaded_file = None
+        option = st.selectbox('Select Mode', ('Chat', 'Code'))
+        uploaded_file = st.file_uploader("", type=["pdf"], disabled=False)
 
     temperature = st.slider("Select Temperature", min_value=0.0, max_value=1.0, value=0.05, step=0.01)
     selected_llm_model = st.selectbox("Select LLM Model", options=list(llm_model.keys()), index=3)
     top_k = st.slider("Select Top K Matches", min_value=1, max_value=20, value=5)
 
-## Initialize unique ID, db_client, db_path, and timestamp if needed
-if 'db_client' not in st.session_state and option == "Chat PDF":
+## Initialize unique ID, db_client, db_path, and timestamp if not already in session state
+if 'db_client' not in st.session_state:
     unique_id = str(uuid.uuid4())
     st.session_state['unique_id'] = unique_id
     db_path = os.path.join(VECTOR_DB_DIR, unique_id)
@@ -137,19 +132,18 @@ if 'db_client' not in st.session_state and option == "Chat PDF":
     log_upload_time(unique_id)
 
 # Access session-stored variables
-if option == "Chat PDF":
-    db_client = st.session_state['db_client']
-    unique_id = st.session_state['unique_id']
-    db_path = st.session_state['db_path']
+db_client = st.session_state['db_client']
+unique_id = st.session_state['unique_id']
+db_path = st.session_state['db_path']
 
-    if 'document_text' not in st.session_state:
-        st.session_state['document_text'] = None
+if 'document_text' not in st.session_state:
+    st.session_state['document_text'] = None
 
-    if 'text_embeddings' not in st.session_state:
-        st.session_state['text_embeddings'] = None
+if 'text_embeddings' not in st.session_state:
+    st.session_state['text_embeddings'] = None
 
 ## Handle PDF Upload and Processing
-if option == "Chat PDF" and uploaded_file is not None and st.session_state['document_text'] is None:
+if uploaded_file is not None and st.session_state['document_text'] is None:
     os.makedirs(UPLOAD_DIR, exist_ok=True)
     file_path = os.path.join(UPLOAD_DIR, f"{unique_id}_paper.pdf")
     with open(file_path, "wb") as file:
@@ -162,18 +156,16 @@ if option == "Chat PDF" and uploaded_file is not None and st.session_state['docu
     text_contents_embeddings = contextEmbeddingChroma(embeddModel, text_content_chunks, db_client, db_path=db_path)
     st.session_state['text_embeddings'] = text_contents_embeddings
 
-if option == "Chat PDF" and st.session_state['document_text'] and st.session_state['text_embeddings']:
+if st.session_state['document_text'] and st.session_state['text_embeddings']:
     document_text = st.session_state['document_text']
     text_contents_embeddings = st.session_state['text_embeddings']
 else:
-    if option == "Chat PDF":
-        st.stop()
+    st.stop()
 
-## Chat Input for Both Modes
 q_input = st.chat_input(key="input", placeholder="Ask your question")
 
 if q_input:
-    if option == "Chat PDF":
+    if option == "Chat":
         query_embedding = ragQuery(embeddModel, q_input)
         top_k_matches = similarityChroma(query_embedding, db_client, top_k)
 
@@ -182,25 +174,18 @@ if q_input:
         prompt_template = q_input
         user_content = top_k_matches
         max_tokens = max_tokens[selected_llm_model]
+        print(max_tokens)
+        top_p = 1
+        stream = True
+        stop = None
 
-        groq_completion = GroqCompletion(groq_client, LLMmodel, domain, prompt_template, user_content, temperature, max_tokens, top_p=1, stream=True, stop=None)
+        groq_completion = GroqCompletion(groq_client, LLMmodel, domain, prompt_template, user_content, temperature, max_tokens, top_p, stream, stop)
         result = groq_completion.create_completion()
         
         with st.spinner("Processing..."):
             chat_response(q_input, result)
 
-    elif option == "Chat LLM":
-        LLMmodel = llm_model[selected_llm_model]
-        domain = "General"
-        prompt_template = q_input
-        user_content = ""
-        max_tokens = max_tokens[selected_llm_model]
+## Call the cleanup function periodically
+cleanup_expired_files()
 
-        groq_completion = GroqCompletion(groq_client, LLMmodel, domain, prompt_template, user_content, temperature, max_tokens, top_p=1, stream=True, stop=None)
-        result = groq_completion.create_completion()
-        
-        with st.spinner("Processing..."):
-            chat_response(q_input, result)
 
-## Periodic Cleanup
-cleanup_expired_files()