Spaces:
Runtime error
Runtime error
Commit
·
203e2cd
1
Parent(s):
e866a92
initial upload
Browse files
app.py
ADDED
|
@@ -0,0 +1,61 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import os
|
| 3 |
+
from langchain.chains import RetrievalQA
|
| 4 |
+
from langchain.llms import OpenAI
|
| 5 |
+
from langchain.document_loaders import PyPDFLoader
|
| 6 |
+
from langchain.text_splitter import CharacterTextSplitter
|
| 7 |
+
from langchain.embeddings import OpenAIEmbeddings
|
| 8 |
+
from langchain.vectorstores import Chroma
|
| 9 |
+
|
| 10 |
+
def qa_system(pdf_file, openai_key, prompt, chain_type, k):
|
| 11 |
+
os.environ["OPENAI_API_KEY"] = openai_key
|
| 12 |
+
|
| 13 |
+
# load document
|
| 14 |
+
loader = PyPDFLoader(pdf_file.name)
|
| 15 |
+
documents = loader.load()
|
| 16 |
+
|
| 17 |
+
# split the documents into chunks
|
| 18 |
+
text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
|
| 19 |
+
texts = text_splitter.split_documents(documents)
|
| 20 |
+
|
| 21 |
+
# select which embeddings we want to use
|
| 22 |
+
embeddings = OpenAIEmbeddings()
|
| 23 |
+
|
| 24 |
+
# create the vectorestore to use as the index
|
| 25 |
+
db = Chroma.from_documents(texts, embeddings)
|
| 26 |
+
|
| 27 |
+
# expose this index in a retriever interface
|
| 28 |
+
retriever = db.as_retriever(search_type="similarity", search_kwargs={"k": k})
|
| 29 |
+
|
| 30 |
+
# create a chain to answer questions
|
| 31 |
+
qa = RetrievalQA.from_chain_type(
|
| 32 |
+
llm=OpenAI(), chain_type=chain_type, retriever=retriever, return_source_documents=True)
|
| 33 |
+
|
| 34 |
+
# get the result
|
| 35 |
+
result = qa({"query": prompt})
|
| 36 |
+
return result['result'], [doc.page_content for doc in result["source_documents"]]
|
| 37 |
+
|
| 38 |
+
# define the Gradio interface
|
| 39 |
+
input_file = gr.inputs.File(label="PDF File")
|
| 40 |
+
openai_key = gr.inputs.Textbox(label="OpenAI API Key")
|
| 41 |
+
prompt = gr.inputs.Textbox(label="Question Prompt")
|
| 42 |
+
chain_type = gr.inputs.Radio(['stuff', 'map_reduce', "refine", "map_rerank"], label="Chain Type")
|
| 43 |
+
k = gr.inputs.Slider(minimum=1, maximum=5, default=1, label="Number of Relevant Chunks")
|
| 44 |
+
|
| 45 |
+
output_text = gr.outputs.Textbox(label="Answer")
|
| 46 |
+
output_docs = gr.outputs.Textbox(label="Relevant Source Text")
|
| 47 |
+
|
| 48 |
+
gr.Interface(qa_system, inputs=[input_file, openai_key, prompt, chain_type, k], outputs=[output_text, output_docs],
|
| 49 |
+
title="Question Answering with PDF File and OpenAI",
|
| 50 |
+
description="Upload a PDF file, enter your OpenAI API key, type a question prompt, select a chain type, and choose the number of relevant chunks to use for the answer.").launch(debug = True)
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
|
| 56 |
+
|
| 57 |
+
|
| 58 |
+
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
|