Update app.py

app.py

@@ -1,32 +1,35 @@
-import streamlit as st 
-import os
 import base64
-import time
-from transformers import AutoTokenizer, AutoModelForSeq2SeqLM 
-from transformers import pipeline
-import torch 
-import textwrap 
-from langchain.document_loaders import PyPDFLoader, DirectoryLoader, PDFMinerLoader 
-from langchain.text_splitter import RecursiveCharacterTextSplitter 
-from langchain.embeddings import SentenceTransformerEmbeddings 
-from langchain.vectorstores import FAISS
+import os
+
+import streamlit as st
+from langchain.chains import RetrievalQA
+from langchain.document_loaders import PDFMinerLoader
+from langchain.embeddings import SentenceTransformerEmbeddings
 from langchain.llms import HuggingFacePipeline
-from langchain.chains import RetrievalQA 
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.vectorstores import FAISS
 from streamlit_chat import message
+from transformers import AutoModelForSeq2SeqLM, AutoTokenizer, pipeline
+import torch
 
-st.set_page_config(layout="wide")
+# Constants
+CHECKPOINT = "MBZUAI/LaMini-T5-738M"
+TOKENIZER = AutoTokenizer.from_pretrained(CHECKPOINT)
+BASE_MODEL = AutoModelForSeq2SeqLM.from_pretrained(CHECKPOINT, device_map=torch.device('cpu'), torch_dtype=torch.float32)
 
-device = torch.device('cpu')
 
-checkpoint = "MBZUAI/LaMini-T5-738M"
-tokenizer = AutoTokenizer.from_pretrained(checkpoint)
-base_model = AutoModelForSeq2SeqLM.from_pretrained(
-    checkpoint,
-    device_map=device,
-    torch_dtype=torch.float32
-)
+def process_answer(instruction, qa_chain):
+    response = ''
+    generated_text = qa_chain.run(instruction)
+    return generated_text
+
+
+def get_file_size(file):
+    file.seek(0, os.SEEK_END)
+    file_size = file.tell()
+    file.seek(0)
+    return file_size
 
-persist_directory = "db"
 
 @st.cache_resource
 def data_ingestion():
@@ -35,36 +38,35 @@ def data_ingestion():
             if file.endswith(".pdf"):
                 print(file)
                 loader = PDFMinerLoader(os.path.join(root, file))
-    
+
     documents = loader.load()
     text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=500)
-    splits = text_splitter.split_documents(document)
-    
-    #create embeddings here
+    splits = text_splitter.split_documents(documents)
+
+    # create embeddings here
     embeddings = SentenceTransformerEmbeddings(model_name="all-MiniLM-L6-v2")
     vectordb = FAISS.from_documents(splits, embeddings)
     vectordb.save_local("faiss_index")
 
-    
+
 @st.cache_resource
-def qa_llm():
+def initialize_qa_chain():
     pipe = pipeline(
         'text2text-generation',
-        model = base_model,
-        tokenizer = tokenizer,
-        max_length = 256,
-        do_sample = True,
-        temperature = 0.3,
-        top_p= 0.95,
-        device=device
+        model=BASE_MODEL,
+        tokenizer=TOKENIZER,
+        max_length=256,
+        do_sample=True,
+        temperature=0.3,
+        top_p=0.95,
+        # device=torch.device('cpu')
     )
-    
+
     llm = HuggingFacePipeline(pipeline=pipe)
     embeddings = SentenceTransformerEmbeddings(model_name="all-MiniLM-L6-v2")
-    
+
     vectordb = FAISS.load_local("faiss_index", embeddings)
-    retriever = db.as_retriever()
-    
+
     # Build a QA chain
     qa_chain = RetrievalQA.from_chain_type(
         llm=llm,
@@ -73,44 +75,35 @@ def qa_llm():
     )
     return qa_chain
 
-def process_answer(instruction):
-    response = ''
-    instruction = instruction
-    qa_chain = qa_llm()
-    
-    generated_text = qa_chain.run(instruction)
-    return generated_text
-
-def get_file_size(file):
-    file.seek(0, os.SEEK_END)
-    file_size = file.tell()
-    file.seek(0)
-    return file_size
 
 @st.cache_data
-#function to display the PDF of a given file 
-def displayPDF(file):
-    # Opening file from file path
-    with open(file, "rb") as f:
-        base64_pdf = base64.b64encode(f.read()).decode('utf-8')
+# function to display the PDF of a given file
+def display_pdf(file):
+    try:
+        # Opening file from file path
+        with open(file, "rb") as f:
+            base64_pdf = base64.b64encode(f.read()).decode('utf-8')
 
-    # Embedding PDF in HTML
-    pdf_display = F'<iframe src="data:application/pdf;base64,{base64_pdf}" width="100%" height="600" type="application/pdf"></iframe>'
+        # Embedding PDF in HTML
+        pdf_display = f'<iframe src="data:application/pdf;base64,{base64_pdf}" width="100%" height="600" type="application/pdf"></iframe>'
+
+        # Displaying File
+        st.markdown(pdf_display, unsafe_allow_html=True)
+    except Exception as e:
+        st.error(f"An error occurred while displaying the PDF: {e}")
 
-    # Displaying File
-    st.markdown(pdf_display, unsafe_allow_html=True)
 
 # Display conversation history using Streamlit messages
 def display_conversation(history):
     for i in range(len(history["generated"])):
-        message(history["past"][i], is_user=True, key=str(i) + "_user")
-        message(history["generated"][i],key=str(i))
+        message(history["past"][i], is_user=True, key=f"{i}_user")
+        message(history["generated"][i], key=str(i))
 
-def main():
-    st.markdown("<h1 style='text-align: center; color: blue;'>Chat with your PDF 🦜📄 </h1>", unsafe_allow_html=True)
-    st.markdown("<h3 style='text-align: center; color: grey;'>Built by <a href='https://github.com/AIAnytime'>AI Anytime with ❤️ </a></h3>", unsafe_allow_html=True)
 
-    st.markdown("<h2 style='text-align: center; color:red;'>Upload your PDF 👇</h2>", unsafe_allow_html=True)
+def main():
+    st.set_page_config(layout="wide")
+    st.markdown("<h1 style='text-align: center; color: blue;'>Custom PDF Chatbot 🦜📄 </h1>", unsafe_allow_html=True)
+    st.markdown("<h2 style='text-align: center; color:red;'>Upload your PDF, and Ask Questions 👇</h2>", unsafe_allow_html=True)
 
     uploaded_file = st.file_uploader("", type=["pdf"])
 
@@ -119,43 +112,46 @@ def main():
             "Filename": uploaded_file.name,
             "File size": get_file_size(uploaded_file)
         }
-        filepath = "docs/"+uploaded_file.name
-        with open(filepath, "wb") as temp_file:
+        filepath = os.path.join("docs", uploaded_file.name)
+        try:
+            with open(filepath, "wb") as temp_file:
                 temp_file.write(uploaded_file.read())
 
-        col1, col2= st.columns([1,2])
-        with col1:
-            st.markdown("<h4 style color:black;'>File details</h4>", unsafe_allow_html=True)
-            st.json(file_details)
-            st.markdown("<h4 style color:black;'>File preview</h4>", unsafe_allow_html=True)
-            pdf_view = displayPDF(filepath)
-
-        with col2:
-            with st.spinner('Embeddings are in process...'):
-                ingested_data = data_ingestion()
-            st.success('Embeddings are created successfully!')
-            st.markdown("<h4 style color:black;'>Chat Here</h4>", unsafe_allow_html=True)
-
-
-            user_input = st.text_input("", key="input")
-
-            # Initialize session state for generated responses and past messages
-            if "generated" not in st.session_state:
-                st.session_state["generated"] = ["I am ready to help you"]
-            if "past" not in st.session_state:
-                st.session_state["past"] = ["Hey there!"]
-                
-            # Search the database for a response based on user input and update session state
-            if user_input:
-                answer = process_answer({'query': user_input})
-                st.session_state["past"].append(user_input)
-                response = answer
-                st.session_state["generated"].append(response)
-
-            # Display conversation history using Streamlit messages
-            if st.session_state["generated"]:
-                display_conversation(st.session_state)
-        
+            col1, col2 = st.columns([1, 2])
+            with col1:
+                st.markdown("<h4 style color:black;'>File details</h4>", unsafe_allow_html=True)
+                st.json(file_details)
+                st.markdown("<h4 style color:black;'>File preview</h4>", unsafe_allow_html=True)
+                pdf_view = display_pdf(filepath)
+
+            with col2:
+                with st.spinner('Embeddings are in process...'):
+                    ingested_data = data_ingestion()
+                st.success('Embeddings are created successfully!')
+                st.markdown("<h4 style color:black;'>Chat Here</h4>", unsafe_allow_html=True)
+
+                user_input = st.text_input("", key="input")
+
+                # Initialize session state for generated responses and past messages
+                if "generated" not in st.session_state:
+                    st.session_state["generated"] = ["I am ready to help you"]
+                if "past" not in st.session_state:
+                    st.session_state["past"] = ["Hey there!"]
+
+                # Search the database for a response based on user input and update session state
+                if user_input:
+                    answer = process_answer({'query': user_input}, initialize_qa_chain())
+                    st.session_state["past"].append(user_input)
+                    response = answer
+                    st.session_state["generated"].append(response)
+
+                # Display conversation history using Streamlit messages
+                if st.session_state["generated"]:
+                    display_conversation(st.session_state)
+
+        except Exception as e:
+            st.error(f"An error occurred: {e}")
+
 
 if __name__ == "__main__":
     main()