Upload 3 files

app.py

@@ -0,0 +1,89 @@
+import os
+import streamlit as st
+import yaml
+from dotenv import load_dotenv
+from src.document_loader import load_document
+from src.chunking_embedding import setup_chunking_and_embedding
+from src.vector_store import create_vectorstore
+from langchain_groq import ChatGroq
+from langchain.chains import ConversationalRetrievalChain
+from langchain.memory import ConversationBufferWindowMemory
+from src.utils import log_interaction
+
+# Load environment variables from .env file
+load_dotenv()
+
+
+# Load configuration
+config_path = os.path.join(os.getcwd(), "config", "config.yaml")  # Ensure correct path
+with open(config_path, "r") as f:
+    config = yaml.safe_load(f)
+
+# Get the GROQ API key from environment variables
+groq_api_key = os.getenv('groq_api_key')
+if not groq_api_key:
+    raise ValueError("GROQ_API_KEY not found. Please set it in the .env file.")
+
+# Streamlit UI
+st.set_page_config(page_title="Your Document AI Assistant", page_icon="📄", layout="centered")
+st.title("🪅 Document AI Assistant!")
+
+# Sidebar for document upload
+uploaded_file = st.sidebar.file_uploader(label="Upload your document (PDF/DOC/DOCX)", type=['pdf', 'doc', 'docx'])
+loader_type = st.sidebar.selectbox('Choose a loader type', config['loaders'])
+embedding_model = st.sidebar.selectbox('Choose an embedding model', config['embedding_models'])
+chunking_strategy = st.sidebar.selectbox('Choose a chunking strategy', config['chunking_strategies'])
+chunk_size = st.sidebar.number_input('Chunk Size', min_value=100, value=500, step=100)
+chunk_overlap = st.sidebar.number_input('Chunk Overlap', min_value=0, value=100, step=100)
+temperature = st.sidebar.slider('Temperature', min_value=0.0, max_value=1.0, value=0.0, step=0.1)
+top_p = st.sidebar.slider('Top-p', min_value=0.0, max_value=1.0, value=0.9, step=0.1)
+
+# Initialize session state for conversation
+if 'chat_history' not in st.session_state:
+    st.session_state.chat_history = []
+if 'conversation_chain' not in st.session_state:
+    st.session_state.conversation_chain = None
+
+# Process the uploaded file
+if uploaded_file is not None:
+    file_path = os.path.join("temp", uploaded_file.name)
+    os.makedirs("temp", exist_ok=True)
+    with open(file_path, "wb") as f:
+        f.write(uploaded_file.getbuffer())
+    
+    documents = load_document(file_path, loader_type)
+    doc_chunks, embeddings = setup_chunking_and_embedding(documents, chunking_strategy, chunk_size, chunk_overlap, embedding_model)
+    vectorstore = create_vectorstore(doc_chunks, embeddings)
+
+    # Create the conversational retrieval chain
+    llm = ChatGroq(groq_api_key=groq_api_key, model_name='llama-3.3-70b-versatile', temperature=temperature)
+    retriever = vectorstore.as_retriever()
+    memory = ConversationBufferWindowMemory(k=5, memory_key="chat_history", return_messages=True)
+    st.session_state.conversation_chain = ConversationalRetrievalChain.from_llm(llm=llm, retriever=retriever, memory=memory)
+
+# User input for asking a question
+user_input = st.text_input("Ask a question:")
+if st.button("Ask Question"):
+    if user_input:
+        with st.chat_message("user"):
+            st.markdown(user_input)
+        
+        with st.chat_message("assistant"):
+            response = st.session_state.conversation_chain({"question": user_input})
+            assistant_response = response['answer']
+            st.markdown(assistant_response)
+            st.session_state.chat_history.append({"role": "assistant", "content": assistant_response})
+
+            # Log interaction
+            params = {
+                "loader": loader_type,
+                "chunking_strategy": chunking_strategy,
+                "chunk_size": chunk_size,
+                "chunk_overlap": chunk_overlap,
+                "embedding_model": embedding_model,
+                "temperature": temperature,
+                "top_p": top_p
+            }
+            log_interaction(user_input, assistant_response, params)
+
+

interaction_log.json

@@ -0,0 +1,15 @@
+[
+    {
+        "question": "what is this document about?",
+        "response": "This document appears to be a sample PDF file containing a passage of Lorem Ipsum text, which is a placeholder text used to demonstrate the layout and formatting of a document. It doesn't seem to be about any specific topic or subject, but rather a filler text used for testing or demonstration purposes.",
+        "params": {
+            "loader": "PyPDF",
+            "chunking_strategy": "Recursive",
+            "chunk_size": 800,
+            "chunk_overlap": 100,
+            "embedding_model": "HuggingFace",
+            "temperature": 0.0,
+            "top_p": 0.9
+        }
+    }
+]

requirements.txt

@@ -0,0 +1,13 @@
+langchain
+pypdf
+python-dotenv
+langchain_groq
+langchain_community
+streamlit
+langchain_huggingface
+faiss-cpu
+pymupdf
+pdfplumber
+pdfminer
+python-docx
+python-doc