Spaces:
Sleeping
Sleeping
Create rag_gradio_app.py
Browse files- rag_gradio_app.py +56 -0
rag_gradio_app.py
ADDED
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import torch
|
3 |
+
from langchain.embeddings import HuggingFaceEmbeddings
|
4 |
+
from langchain_community.vectorstores import Chroma
|
5 |
+
from sentence_transformers import SentenceTransformer
|
6 |
+
import openai
|
7 |
+
|
8 |
+
# Load pre-trained embedding model
|
9 |
+
model_name = 'intfloat/e5-small'
|
10 |
+
embedding_model = HuggingFaceEmbeddings(model_name=model_name)
|
11 |
+
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
12 |
+
|
13 |
+
# Load ChromaDB
|
14 |
+
persist_directory = './docs/chroma/'
|
15 |
+
vectordb = Chroma(persist_directory=persist_directory, embedding_function=embedding_model)
|
16 |
+
|
17 |
+
# OpenAI API Key
|
18 |
+
openai.api_key = 'your-api-key'
|
19 |
+
|
20 |
+
def retrieve_documents(question, k=5):
|
21 |
+
"""Retrieve top K relevant documents from ChromaDB"""
|
22 |
+
docs = vectordb.similarity_search(question, k=k)
|
23 |
+
return [doc.page_content for doc in docs]
|
24 |
+
|
25 |
+
def generate_response(question, context):
|
26 |
+
"""Generate response using OpenAI GPT-4"""
|
27 |
+
full_prompt = f"Context: {context}\n\nQuestion: {question}"
|
28 |
+
response = openai.ChatCompletion.create(
|
29 |
+
model="gpt-4",
|
30 |
+
messages=[{"role": "user", "content": full_prompt}],
|
31 |
+
max_tokens=300,
|
32 |
+
temperature=0.7
|
33 |
+
)
|
34 |
+
return response['choices'][0]['message']['content'].strip()
|
35 |
+
|
36 |
+
def rag_pipeline(question):
|
37 |
+
"""Full RAG Pipeline - Retrieve Docs & Generate Response"""
|
38 |
+
retrieved_docs = retrieve_documents(question, k=5)
|
39 |
+
context = " ".join(retrieved_docs)
|
40 |
+
response = generate_response(question, context)
|
41 |
+
return response, retrieved_docs
|
42 |
+
|
43 |
+
def gradio_interface(question):
|
44 |
+
response, retrieved_docs = rag_pipeline(question)
|
45 |
+
return response, "\n\n".join(retrieved_docs)
|
46 |
+
|
47 |
+
# Create Gradio App
|
48 |
+
iface = gr.Interface(
|
49 |
+
fn=gradio_interface,
|
50 |
+
inputs=gr.Textbox(label="Enter your question"),
|
51 |
+
outputs=[gr.Textbox(label="Generated Response"), gr.Textbox(label="Retrieved Documents")],
|
52 |
+
title="RAG-Based Question Answering System",
|
53 |
+
description="Enter a question and retrieve relevant documents along with the AI-generated response."
|
54 |
+
)
|
55 |
+
|
56 |
+
iface.launch()
|