Create app.py

app.py

@@ -0,0 +1,146 @@
+# =========================
+# IMPORTS
+# =========================
+from langchain_community.document_loaders import PyPDFLoader
+from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain.embeddings import HuggingFaceEmbeddings
+from langchain.vectorstores import Chroma
+from groq import Groq
+from duckduckgo_search import DDGS
+import gradio as gr
+
+
+# =========================
+# CONFIG
+# =========================
+GROQ_API_KEY = "Ai_tutor"   # 🔥 set in HF secrets instead
+
+client = Groq(api_key=GROQ_API_KEY)
+
+vectorstore = None
+retriever = None
+
+
+# =========================
+# PROMPT
+# =========================
+def build_prompt(context, question):
+    return f"""
+You are an expert AI assistant.
+
+Use ONLY the context below.
+If answer is not present, say "Not found in document".
+
+CONTEXT:
+{context}
+
+QUESTION:
+{question}
+
+ANSWER:
+"""
+
+
+# =========================
+# WEB SEARCH
+# =========================
+def web_search(query):
+    results = []
+    with DDGS() as ddgs:
+        for r in ddgs.text(query, max_results=3):
+            results.append(r["body"])
+    return "\n\n".join(results)
+
+
+# =========================
+# PROCESS PDF
+# =========================
+def process_pdf(file):
+
+    global vectorstore, retriever
+
+    loader = PyPDFLoader(file.name)
+    documents = loader.load()
+
+    text_splitter = RecursiveCharacterTextSplitter(
+        chunk_size=600,
+        chunk_overlap=100
+    )
+
+    chunks = text_splitter.split_documents(documents)
+
+    embedding_model = HuggingFaceEmbeddings(
+        model_name="sentence-transformers/all-MiniLM-L6-v2"
+    )
+
+    vectorstore = Chroma.from_documents(
+        documents=chunks,
+        embedding=embedding_model
+    )
+
+    retriever = vectorstore.as_retriever(search_kwargs={"k": 4})
+
+    return "✅ PDF processed. Ask questions."
+
+
+# =========================
+# RAG FUNCTION (HYBRID)
+# =========================
+def ask_rag(query):
+
+    global retriever
+
+    if retriever is None:
+        return "⚠️ Upload a PDF first."
+
+    docs = retriever.invoke(query)
+    pdf_context = "\n\n".join([d.page_content for d in docs])
+
+    # Hybrid fallback
+    if len(pdf_context.strip()) < 50:
+        web_context = web_search(query)
+        context = pdf_context + "\n\nWEB:\n" + web_context
+    else:
+        context = pdf_context
+
+    prompt = build_prompt(context, query)
+
+    response = client.chat.completions.create(
+        messages=[{"role": "user", "content": prompt}],
+        model="llama3-8b-8192"
+    )
+
+    return response.choices[0].message.content
+
+
+# =========================
+# CHAT
+# =========================
+def chat(user_message, history):
+
+    response = ask_rag(user_message)
+
+    history.append({"role": "user", "content": user_message})
+    history.append({"role": "assistant", "content": response})
+
+    return history, history
+
+
+# =========================
+# UI
+# =========================
+with gr.Blocks() as app:
+
+    gr.Markdown("# 🧠 Hybrid RAG Chatbot (PDF + Web)")
+
+    file = gr.File(label="Upload PDF")
+    status = gr.Textbox(label="Status")
+
+    chatbot = gr.Chatbot(type="messages", allow_tags=False)
+    msg = gr.Textbox(placeholder="Ask your question...")
+    state = gr.State([])
+
+    file.change(process_pdf, file, status)
+    msg.submit(chat, [msg, state], [chatbot, state])
+
+app.launch()