|
|
--- |
|
|
title: Bottttt |
|
|
emoji: 📉 |
|
|
colorFrom: gray |
|
|
colorTo: indigo |
|
|
sdk: gradio |
|
|
sdk_version: 5.36.2 |
|
|
app_file: app.py |
|
|
pinned: false |
|
|
license: apache-2.0 |
|
|
--- |
|
|
|
|
|
Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference |
|
|
Great — let’s prepare your **RAG app** for **deployment on Hugging Face Spaces** with: |
|
|
|
|
|
* ✅ Gradio as UI |
|
|
* ✅ LLaMA3-Instruct via Groq API |
|
|
* ✅ Sentence Transformers |
|
|
* ✅ ChromaDB with persistence |
|
|
* ✅ PDF upload + student Q\&A |
|
|
|
|
|
--- |
|
|
|
|
|
## ✅ STEP 1: Project Structure |
|
|
|
|
|
Create this directory structure for your Hugging Face Space: |
|
|
|
|
|
``` |
|
|
rag-student-assistant/ |
|
|
├── app.py |
|
|
├── requirements.txt |
|
|
└── .env (optional, but don’t upload publicly) |
|
|
``` |
|
|
|
|
|
--- |
|
|
|
|
|
## ✅ STEP 2: `app.py` (Full Code) |
|
|
|
|
|
```python |
|
|
import os |
|
|
import gradio as gr |
|
|
import fitz # PyMuPDF |
|
|
from sentence_transformers import SentenceTransformer |
|
|
import chromadb |
|
|
from chromadb.utils import embedding_functions |
|
|
import openai |
|
|
|
|
|
# Load GROQ API Key |
|
|
openai.api_key = os.getenv("GROQ_API_KEY") |
|
|
openai.api_base = "https://api.groq.com/openai/v1" |
|
|
|
|
|
# Load embedding model |
|
|
embedder = SentenceTransformer("all-MiniLM-L6-v2") |
|
|
|
|
|
# Set up ChromaDB with persistence |
|
|
persist_path = "./chroma_db" |
|
|
db = chromadb.Client(chromadb.config.Settings(persist_directory=persist_path)) |
|
|
collection = db.get_or_create_collection("papers") |
|
|
|
|
|
# Extract text from uploaded PDF |
|
|
def extract_text_from_pdf(file): |
|
|
text = "" |
|
|
doc = fitz.open(stream=file.read(), filetype="pdf") |
|
|
for page in doc: |
|
|
text += page.get_text() |
|
|
return text |
|
|
|
|
|
# Chunk and store in vector DB |
|
|
def chunk_and_store(text): |
|
|
chunks = [text[i:i+500] for i in range(0, len(text), 500)] |
|
|
embeddings = embedder.encode(chunks).tolist() |
|
|
|
|
|
for i, chunk in enumerate(chunks): |
|
|
collection.add(documents=[chunk], ids=[f"id_{len(collection.get()['ids']) + i}"], embeddings=[embeddings[i]]) |
|
|
db.persist() |
|
|
|
|
|
# Retrieve relevant chunks and send to LLaMA3 via Groq |
|
|
def retrieve_and_ask(query): |
|
|
if len(collection.get()["documents"]) == 0: |
|
|
return "Please upload a paper first." |
|
|
|
|
|
query_embedding = embedder.encode([query]).tolist()[0] |
|
|
results = collection.query(query_embeddings=[query_embedding], n_results=3) |
|
|
context = "\n".join(results["documents"][0]) |
|
|
|
|
|
system_prompt = "You are an academic assistant helping students understand research papers." |
|
|
user_prompt = f"Based on the following context:\n{context}\n\nAnswer the question:\n{query}" |
|
|
|
|
|
try: |
|
|
response = openai.ChatCompletion.create( |
|
|
model="llama3-70b-8192", |
|
|
messages=[ |
|
|
{"role": "system", "content": system_prompt}, |
|
|
{"role": "user", "content": user_prompt} |
|
|
] |
|
|
) |
|
|
return response['choices'][0]['message']['content'] |
|
|
except Exception as e: |
|
|
return f"Error: {str(e)}" |
|
|
|
|
|
# Gradio UI |
|
|
def handle_upload(file): |
|
|
if file is None: |
|
|
return "Upload a valid PDF file." |
|
|
text = extract_text_from_pdf(file) |
|
|
chunk_and_store(text) |
|
|
return "✅ Paper uploaded and processed." |
|
|
|
|
|
def handle_query(query): |
|
|
return retrieve_and_ask(query) |
|
|
|
|
|
with gr.Blocks() as demo: |
|
|
gr.Markdown("### 📘 RAG Academic Assistant\nUpload a paper and ask questions.") |
|
|
|
|
|
with gr.Row(): |
|
|
file = gr.File(label="Upload PDF", type="binary") |
|
|
upload_btn = gr.Button("Process") |
|
|
upload_output = gr.Textbox() |
|
|
|
|
|
with gr.Row(): |
|
|
query = gr.Textbox(label="Ask a question") |
|
|
response = gr.Textbox(label="Answer") |
|
|
ask_btn = gr.Button("Ask") |
|
|
|
|
|
upload_btn.click(handle_upload, inputs=[file], outputs=[upload_output]) |
|
|
ask_btn.click(handle_query, inputs=[query], outputs=[response]) |
|
|
|
|
|
demo.launch() |
|
|
``` |
|
|
|
|
|
--- |
|
|
|
|
|
## ✅ STEP 3: `requirements.txt` |
|
|
|
|
|
```txt |
|
|
gradio |
|
|
chromadb |
|
|
sentence-transformers |
|
|
PyMuPDF |
|
|
openai |
|
|
``` |
|
|
|
|
|
> Hugging Face Spaces will auto-install these on build. |
|
|
|
|
|
--- |
|
|
|
|
|
## ✅ STEP 4: GROQ API Key |
|
|
|
|
|
### 🔐 Option 1: Use Hugging Face "Secrets" |
|
|
|
|
|
* Go to your Space → **Settings > Secrets** |
|
|
* Add a new secret: |
|
|
|
|
|
* **Name:** `GROQ_API_KEY` |
|
|
* **Value:** `your-api-key-here` |
|
|
|
|
|
No need to change code. It will use `os.getenv("GROQ_API_KEY")`. |
|
|
|
|
|
--- |
|
|
|
|
|
## ✅ STEP 5: Deploy on Hugging Face |
|
|
|
|
|
1. Go to [Hugging Face Spaces](https://huggingface.co/spaces) |
|
|
2. Click **Create New Space** |
|
|
3. Choose: |
|
|
|
|
|
* **Gradio** |
|
|
* **Public or Private** |
|
|
4. Upload: |
|
|
|
|
|
* `app.py` |
|
|
* `requirements.txt` |
|
|
5. Add GROQ API key under **Settings > Secrets** |
|
|
|
|
|
--- |
|
|
|
|
|
## ✅ You’re Done! |
|
|
|
|
|
After deployment: |
|
|
|
|
|
* Students can upload PDF papers |
|
|
* Ask natural language questions |
|
|
* Get Groq/LLaMA3-generated answers from your vector database |
|
|
|
|
|
--- |
|
|
|
|
|
Would you like me to: |
|
|
|
|
|
* 🎁 Zip the files for direct upload? |
|
|
* 🧪 Add test examples? |
|
|
* 🎓 Add UI branding for universities or students? |
|
|
|
|
|
Let me know what extras you want! |
|
|
|
