Create app.py

app.py

@@ -0,0 +1,93 @@
+import gradio as gr
+import fitz
+import re
+import faiss
+import torch
+import numpy as np
+from sentence_transformers import SentenceTransformer
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+# -------- Load Models --------
+embedding_model = SentenceTransformer("all-MiniLM-L6-v2")
+
+llm_name = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
+tokenizer = AutoTokenizer.from_pretrained(llm_name)
+llm = AutoModelForCausalLM.from_pretrained(
+    llm_name,
+    torch_dtype=torch.float32
+)
+
+# -------- Helper Functions --------
+def extract_text(pdf_file):
+    doc = fitz.open(pdf_file)
+    text = ""
+    for page in doc:
+        text += page.get_text()
+    return text
+
+def clean_text(text):
+    return re.sub(r"\s+", " ", text)
+
+def chunk_text(text, chunk_size=500, overlap=50):
+    chunks = []
+    start = 0
+    while start < len(text):
+        end = start + chunk_size
+        chunks.append(text[start:end])
+        start = end - overlap
+    return chunks
+
+def build_vector_db(chunks):
+    embeddings = embedding_model.encode(chunks)
+    embeddings = np.array(embeddings).astype("float32")
+    index = faiss.IndexFlatL2(embeddings.shape[1])
+    index.add(embeddings)
+    return index, chunks
+
+def retrieve_context(query, index, chunks, k=3):
+    q_emb = embedding_model.encode([query]).astype("float32")
+    _, indices = index.search(q_emb, k)
+    return [chunks[i] for i in indices[0]]
+
+def generate_answer(question, context_chunks):
+    context = "\n\n".join(context_chunks)
+    prompt = f"""
+Answer the question using ONLY the context below.
+If not found, say "Information not found in the document."
+
+Context:
+{context}
+
+Question:
+{question}
+
+Answer:
+"""
+    inputs = tokenizer(prompt, return_tensors="pt", truncation=True)
+    with torch.no_grad():
+        output = llm.generate(**inputs, max_new_tokens=200)
+    response = tokenizer.decode(output[0], skip_special_tokens=True)
+    return response.split("Answer:")[-1].strip()
+
+# -------- Main Pipeline --------
+def pdf_chat(pdf, question):
+    text = extract_text(pdf.name)
+    text = clean_text(text)
+    chunks = chunk_text(text)
+    index, chunks = build_vector_db(chunks)
+    context = retrieve_context(question, index, chunks)
+    return generate_answer(question, context)
+
+# -------- Gradio UI --------
+interface = gr.Interface(
+    fn=pdf_chat,
+    inputs=[
+        gr.File(label="Upload PDF"),
+        gr.Textbox(label="Ask a question")
+    ],
+    outputs=gr.Textbox(label="Answer"),
+    title="📄 PDF RAG Chatbot (Open-Source AI)",
+    description="Upload a PDF and ask questions. Runs on free CPU using Hugging Face open-source models."
+)
+
+interface.launch()