Upload 3 files

app.py

@@ -0,0 +1,52 @@
+import gradio as gr
+from rag_pipeline import build_rag_chain
+from dotenv import load_dotenv
+
+load_dotenv()
+
+chain = None
+
+def upload_file(files):
+    global chain
+    if files is None:
+        return "No file uploaded."
+    if isinstance(files, str):
+        paths = [files]
+    else:
+        paths = [f if isinstance(f, str) else f.name for f in files] if isinstance(files, list) else [files.name]
+    
+    full_chain, _ = build_rag_chain(paths)
+    chain = full_chain
+    return f"✅ {len(paths)} document(s) loaded. Ask away!"
+
+def ask_question(message, history):
+    if chain is None:
+        return "Upload a document first."
+    
+    history_text = ""
+    for h in history:
+        if isinstance(h, (list, tuple)) and len(h) == 2:
+            history_text += f"User: {h[0]}\nAssistant: {h[1]}\n"
+    
+    full_query = f"{history_text}Question: {message}" if history_text else message
+    
+    result = chain.invoke(full_query)
+    answer = result["answer"]
+    pages = sorted(set(
+        d.metadata.get("page", 0) + 1
+        for d in result["source_documents"]
+    ))
+    return f"{answer}\n\n📄 Source pages: {pages}"
+
+with gr.Blocks() as full_app:
+    gr.Markdown("# 📄 RAG Document Q&A")
+    file_input = gr.File(label="Upload PDF(s)", file_types=[".pdf"], file_count="multiple")
+    status = gr.Textbox(label="Status", interactive=False)
+    file_input.upload(upload_file, inputs=file_input, outputs=status)
+    gr.Markdown("---")
+    question = gr.Textbox(label="Ask a question")
+    answer = gr.Textbox(label="Answer", interactive=False, lines=10)
+    btn = gr.Button("Ask")
+    btn.click(fn=ask_question, inputs=[question, answer], outputs=answer)
+
+full_app.launch()

rag_pipeline.py

@@ -0,0 +1,48 @@
+from langchain_community.document_loaders import PyPDFLoader
+from langchain_text_splitters import RecursiveCharacterTextSplitter
+from langchain_community.embeddings import HuggingFaceEmbeddings
+from langchain_community.vectorstores import FAISS
+from langchain_groq import ChatGroq
+from langchain_core.prompts import PromptTemplate
+from langchain_core.runnables import RunnablePassthrough, RunnableParallel
+from langchain_core.output_parsers import StrOutputParser
+import os
+
+def build_rag_chain(file_paths):
+    if isinstance(file_paths, str):
+        file_paths = [file_paths]
+
+    all_chunks = []
+    splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50)
+
+    for path in file_paths:
+        loader = PyPDFLoader(path)
+        docs = loader.load()
+        all_chunks.extend(splitter.split_documents(docs))
+
+    embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
+    vectorstore = FAISS.from_documents(all_chunks, embeddings)
+    retriever = vectorstore.as_retriever(search_kwargs={"k": 4})
+
+    llm = ChatGroq(
+        groq_api_key=os.getenv("GROQ_API_KEY"),
+        model_name="llama-3.3-70b-versatile"
+    )
+
+    prompt = PromptTemplate.from_template(
+        "Use the context below to answer the question.\n\nContext: {context}\n\nQuestion: {question}"
+    )
+
+    answer_chain = (
+        {"context": retriever, "question": RunnablePassthrough()}
+        | prompt
+        | llm
+        | StrOutputParser()
+    )
+
+    full_chain = RunnableParallel(
+        answer=answer_chain,
+        source_documents=retriever
+    )
+
+    return full_chain, retriever

requirements.txt

@@ -0,0 +1,10 @@
+langchain
+langchain-community
+langchain-groq
+langchain-core
+langchain-text-splitters
+faiss-cpu
+sentence-transformers
+gradio
+python-dotenv
+pypdf