| import gradio as gr | |
| from langchain.document_loaders import OnlinePDFLoader | |
| from langchain.text_splitter import CharacterTextSplitter | |
| text_splitter = CharacterTextSplitter(chunk_size=350, chunk_overlap=0) | |
| from langchain.llms import HuggingFaceHub | |
| flan_ul2 = HuggingFaceHub(repo_id="google/flan-ul2", model_kwargs={"temperature":0.1, "max_new_tokens":300}) | |
| from langchain.embeddings import HuggingFaceHubEmbeddings | |
| embeddings = HuggingFaceHubEmbeddings() | |
| from langchain.vectorstores import Chroma | |
| from langchain.chains import RetrievalQA | |
| def pdf_changes(pdf_doc): | |
| loader = OnlinePDFLoader(pdf_doc.name) | |
| documents = loader.load() | |
| texts = text_splitter.split_documents(documents) | |
| db = Chroma.from_documents(texts, embeddings) | |
| retriever = db.as_retriever() | |
| global qa | |
| qa = RetrievalQA.from_chain_type(llm=flan_ul2, chain_type="stuff", retriever=retriever, return_source_documents=True) | |
| return "Ready" | |
| def add_text(history, text): | |
| history = history + [(text, None)] | |
| return history, "" | |
| def bot(history): | |
| response = infer(history[-1][0]) | |
| history[-1][1] = response | |
| return history | |
| def infer(question): | |
| query = question | |
| result = qa({"query": query}) | |
| return result | |
| with gr.Blocks() as demo: | |
| with gr.Column(): | |
| pdf_doc = gr.File(label="Load a pdf", file_types=['.pdf'], type="file") | |
| langchain_status = gr.Textbox() | |
| load_pdf = gr.Button("Load pdf to langchain") | |
| chatbot = gr.Chatbot([], elem_id="chatbot").style(height=350) | |
| question = gr.Textbox(label="Question") | |
| load_pdf.click(pdf_changes, pdf_doc, langchain_status, queue=False) | |
| question.submit(add_text, [chatbot, question], [chatbot, question]).then( | |
| bot, chatbot, chatbot | |
| ) | |
| demo.launch() |