Spaces:
Sleeping
Sleeping
import os | |
import gradio as gr | |
from langchain.document_loaders import PDFMinerLoader,CSVLoader ,UnstructuredWordDocumentLoader,TextLoader,OnlinePDFLoader | |
from langchain.text_splitter import CharacterTextSplitter | |
from langchain.embeddings import SentenceTransformerEmbeddings | |
from langchain.vectorstores import FAISS | |
from langchain import HuggingFaceHub | |
from langchain.chains import RetrievalQA | |
from langchain.prompts import PromptTemplate | |
DEVICE = 'cpu' | |
FILE_EXT = ['pdf','text','csv','word','wav'] | |
def loading_file(): | |
return "Loading..." | |
def get_openai_chat_model(API_key): | |
try: | |
from langchain.llms import OpenAI | |
except ImportError as err: | |
raise "{}, unable to load openAI. Please install openai and add OPENAIAPI_KEY" | |
os.environ["OPENAI_API_KEY"] = API_key | |
llm = OpenAI() | |
return llm | |
def process_documents(documents,data_chunk=1000,chunk_overlap=50): | |
text_splitter = CharacterTextSplitter(chunk_size=data_chunk, chunk_overlap=chunk_overlap,separator='\n') | |
texts = text_splitter.split_documents(documents) | |
return texts | |
def get_hugging_face_model(model_id,API_key,temperature=0.1): | |
chat_llm = HuggingFaceHub(huggingfacehub_api_token=API_key, | |
repo_id=model_id, | |
model_kwargs={"temperature": temperature, "max_new_tokens": 2048}) | |
return chat_llm | |
def chat_application(llm_service,key): | |
if llm_service == 'HuggingFace': | |
llm = get_hugging_face_model(model_id='tiiuae/falcon-7b-instruct',API_key=key) | |
else: | |
llm = get_openai_chat_model(API_key=key) | |
return llm | |
def summarize_contents(): | |
question = "Generate a summary of the contents. Do not return the response in json format" | |
return qa.run(question) | |
def document_loader(file_path,api_key,doc_type='pdf',llm='Huggingface'): | |
embedding_model = SentenceTransformerEmbeddings(model_name='all-mpnet-base-v2',model_kwargs={"device": DEVICE}) | |
document = None | |
if doc_type == 'pdf': | |
document = process_pdf_document(document_file=file_path) | |
elif doc_type == 'text': | |
document = process_text_document(document_file=file_path) | |
elif doc_type == 'csv': | |
document = process_csv_document(document_file=file_path) | |
elif doc_type == 'word': | |
document = process_word_document(document_file=file_path) | |
if document: | |
texts = process_documents(documents=document) | |
vector_db = FAISS.from_documents(documents=texts, embedding= embedding_model) | |
global qa | |
qa = RetrievalQA.from_chain_type(llm=chat_application(llm_service=llm,key=api_key), | |
chain_type='stuff', | |
retriever=vector_db.as_retriever(), | |
# chain_type_kwargs=chain_type_kwargs, | |
return_source_documents=True | |
) | |
else: | |
return "Error in loading Documents " | |
return "Ready..." | |
def process_text_document(document_file): | |
loader = TextLoader(document_file.name) | |
document = loader.load() | |
return document | |
def process_csv_document(document_file): | |
loader = CSVLoader(file_path=document_file.name) | |
document = loader.load() | |
return document | |
def process_word_document(document_file): | |
loader = UnstructuredWordDocumentLoader(file_path=document_file.name) | |
document = loader.load() | |
return document | |
def process_pdf_document(document_file): | |
print("Document File Name :",document_file.name) | |
loader = PDFMinerLoader(document_file.name) | |
document = loader.load() | |
return document | |
css=""" | |
#col-container {max-width: 700px; margin-left: auto; margin-right: auto;} | |
""" | |
title = """ | |
<div style="text-align: center;max-width: 700px;"> | |
<h1>Chat with Data • OpenAI/HuggingFace</h1> | |
<p style="text-align: center;">Upload a file from your computer, click the "Load data to LangChain" button, <br /> | |
when everything is ready, you can start asking questions about the data you uploaded ;) <br /> | |
This version is just for QA retrival so it will not use chat history, and uses Hugging face as LLM, | |
so you don't need any key</p> | |
</div> | |
""" | |
with gr.Blocks(css=css) as demo: | |
with gr.Column(elem_id="col-container"): | |
gr.HTML(title) | |
with gr.Column(): | |
with gr.Box(): | |
gr.Row() | |
LLM_option = gr.Dropdown(['HuggingFace','OpenAI'],label='Large Language Model Selection',info='LLM Service') | |
file_extension = gr.Dropdown(FILE_EXT, label="File Extensions", info="Select your files extensions!") | |
API_key = gr.Textbox(label="Add API key", type="password") | |
with gr.Column(): | |
with gr.Box(): | |
pdf_doc = gr.File(label="Upload File to start QA", file_types=FILE_EXT, type="file") | |
with gr.Row(): | |
langchain_status = gr.Textbox(label="Status", placeholder="", interactive=True) | |
load_pdf = gr.Button("Upload File & Generate Embeddings",).style(full_width=False) | |
with gr.Row(): | |
summary = gr.Textbox(label="Summary") | |
summarize_pdf = gr.Button("Summarize the Contents").style(full_width=False) | |
chatbot = gr.Chatbot() | |
question = gr.Textbox(label="Question", placeholder="Type your question and hit Enter") | |
submit_button = gr.Button("Send Message") | |
load_pdf.click(loading_file, None, langchain_status, queue=False) | |
load_pdf.click(document_loader, inputs=[pdf_doc,API_key,file_extension,LLM_option], outputs=[langchain_status], queue=False) | |
summarize_pdf.click(summarize_contents,outputs=summary) | |
# question.submit(add_text, [chatbot, question], [chatbot, question]).then( | |
# bot, chatbot, chatbot | |
# ) | |
demo.launch() |