Spaces:
Build error
Build error
Commit
•
c6709ba
1
Parent(s):
ed64697
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,81 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from langchain.callbacks.manager import CallbackManager
|
3 |
+
from langchain.callbacks.streaming_stdout import StreamingStdOutCallbackHandler
|
4 |
+
from langchain.prompts import PromptTemplate
|
5 |
+
from langchain_community.llms import LlamaCpp
|
6 |
+
from langchain_core.runnables import RunnablePassthrough
|
7 |
+
from langchain_core.output_parsers import StrOutputParser
|
8 |
+
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
9 |
+
from langchain_community.document_loaders import WebBaseLoader
|
10 |
+
from langchain_huggingface.embeddings import HuggingFaceEmbeddings
|
11 |
+
from langchain_community.vectorstores import Chroma
|
12 |
+
from langchain import hub
|
13 |
+
|
14 |
+
# Set up callback manager and model parameters
|
15 |
+
callback_manager = CallbackManager([StreamingStdOutCallbackHandler()])
|
16 |
+
n_gpu_layers = 0
|
17 |
+
n_batch = 512
|
18 |
+
|
19 |
+
llm = LlamaCpp(
|
20 |
+
model_path="./models/phi-2.Q2_K.gguf",
|
21 |
+
n_gpu_layers=n_gpu_layers, n_batch=n_batch,
|
22 |
+
n_ctx = 4096,
|
23 |
+
temperature=0.7,
|
24 |
+
max_tokens=4096,
|
25 |
+
top_p=1,
|
26 |
+
callback_manager=callback_manager,
|
27 |
+
verbose=False,
|
28 |
+
)
|
29 |
+
|
30 |
+
# Load the prompt
|
31 |
+
prompt = hub.pull("rlm/rag-prompt")
|
32 |
+
|
33 |
+
# Function to format documents
|
34 |
+
def format_docs(docs):
|
35 |
+
return "\n\n".join(doc.page_content for doc in docs)
|
36 |
+
|
37 |
+
# Main function to process the question and URL
|
38 |
+
def get_answer(question, url):
|
39 |
+
# Load data from the provided URL
|
40 |
+
loader = WebBaseLoader(url)
|
41 |
+
data = loader.load()
|
42 |
+
|
43 |
+
# Split the data into small chunks
|
44 |
+
text_splitter = RecursiveCharacterTextSplitter(chunk_size=2000, chunk_overlap=0)
|
45 |
+
all_splits = text_splitter.split_documents(data)
|
46 |
+
|
47 |
+
# Store the data in Vector Store
|
48 |
+
vectorstore = Chroma.from_documents(documents=all_splits, embedding=HuggingFaceEmbeddings())
|
49 |
+
retriever = vectorstore.as_retriever(search_type="similarity", search_kwargs={"k": 3})
|
50 |
+
|
51 |
+
retrieved_docs = retriever.invoke(question)
|
52 |
+
|
53 |
+
rag_chain = (
|
54 |
+
{"context": retriever | format_docs, "question": RunnablePassthrough()}
|
55 |
+
| prompt
|
56 |
+
| llm
|
57 |
+
| StrOutputParser()
|
58 |
+
)
|
59 |
+
|
60 |
+
answer = ""
|
61 |
+
for chunk in rag_chain.stream(question):
|
62 |
+
answer += chunk
|
63 |
+
yield answer
|
64 |
+
|
65 |
+
yield answer
|
66 |
+
|
67 |
+
# Create the Gradio interface
|
68 |
+
iface = gr.Interface(
|
69 |
+
fn=get_answer,
|
70 |
+
inputs=[gr.Textbox(lines=1, placeholder="Enter your question here..."),
|
71 |
+
gr.Textbox(lines=1, placeholder="Enter the website URL here...")],
|
72 |
+
outputs="text",
|
73 |
+
title="Web-based Question Answering System",
|
74 |
+
description="Ask a question about the content of a webpage and get an answer.",
|
75 |
+
examples=[
|
76 |
+
["Which are the top 5 companies in the world with their revenue in table format?", "https://www.investopedia.com/biggest-companies-in-the-world-by-market-cap-5212784"]
|
77 |
+
]
|
78 |
+
)
|
79 |
+
|
80 |
+
# Launch the app
|
81 |
+
iface.launch(share=True)
|