add other parameters in dir

app.py

@@ -1,7 +1,5 @@
 import streamlit as st
 import pandas as pd
-# import os
-# import docx2txt
 from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM
 from langchain_community.vectorstores import FAISS
 from langchain_huggingface import HuggingFaceEmbeddings
@@ -9,8 +7,6 @@ from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain.schema import Document
 from sentence_transformers import SentenceTransformer
 from langchain_community.document_loaders import DirectoryLoader, TextLoader
-# from dotenv import load_dotenv
-# load_dotenv()
 
 
 model_name = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
@@ -25,12 +21,14 @@ extracted_text = None
 # ------------------------------
 # Load Model for pretraining
 # ------------------------------
+@st.cache_resource
 def load_model():
     # model_name = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
     tokenizer = AutoTokenizer.from_pretrained(model_name)
     model = AutoModelForCausalLM.from_pretrained(model_name)
     return pipeline("text-generation", model=model, tokenizer=tokenizer)
 
+@st.cache_resource
 def extract_text():
     uploaded_data_path = "./msci"
     loader = DirectoryLoader(
@@ -63,22 +61,27 @@ with st.spinner("🔄 Loading Knowldge Base..."):
 st.title("📚 RAG For MSCI Indexes")
 st.markdown("This app uses a local LLM model to answer questions about MSCI Indexes using RAG (Retrieval Augmented Generation).")   
 
-query = st.text_input("💬 Ask a question about MSCI Indexes", placeholder="MSCI World IMI Index")
-
-if query and db and extracted_text and len(docs) > 0:
-    retriever = db.as_retriever(search_kwargs={"k": 3})
-    retrieved_docs = retriever.get_relevant_documents(query)
-    context = "\n".join([doc.page_content for doc in retrieved_docs])
-
-    with st.spinner("🤔 Generating answer..."):
-        result = generator(
-            f"Context:\n{context}\n\nQuestion: {query}\nAnswer:",
-            max_new_tokens=150,
-            temperature=0.5,
-            top_p=0.9
-        )
-
-        generated = result[0]["generated_text"]
-        answer_only = generated.split("Answer:")[-1].strip()
-
-    st.write("📝 Answer:", answer_only)
+with st.form(key='my_form'):
+    query = st.text_input("💬 Ask a question about MSCI Indexes(Required)", placeholder="MSCI World IMI Index")
+    max_new_tokens_model = st.slider("Max New Tokens (Optional):", min_value=50, max_value=500, value=150, step=25)
+    temperature_model = st.slider("Temperature (Optional):", min_value=0.0, max_value=0.9, value=0.5, step=0.1)
+    submit_button = st.form_submit_button("Submit")
+
+if submit_button:
+    if query and db and extracted_text and len(docs) > 0:
+        retriever = db.as_retriever(search_kwargs={"k": 3})
+        retrieved_docs = retriever.get_relevant_documents(query)
+        context = "\n".join([doc.page_content for doc in retrieved_docs])
+
+        with st.spinner("🤔 Generating answer..."):
+            result = generator(
+                f"Context:\n{context}\n\nQuestion: {query}\nAnswer:",
+                max_new_tokens=max_new_tokens_model,
+                temperature=temperature_model,
+                top_p=0.9
+            )
+
+            generated = result[0]["generated_text"]
+            answer_only = generated.split("Answer:")[-1].strip()
+
+        st.write("📝 Answer:", answer_only)