import gradio as gr import os import time import shutil import base64 from pdfquery import PDFQuery pquery = PDFQuery() def openai_create(s): global pquery return pquery.ask(s) def chatgpt_clone(input, history, chatbot): if input == "": return chatbot, history, "" history = history or [] s = list(sum(history, ())) s.append(input) inp = ' '.join(s) output = openai_create(input) history.append((inp, output)) chatbot.append((input, output)) return chatbot, history, "" title_html = f"

Chat With Pdf

" gr_L1 = lambda: gr.Row().style() gr_L2 = lambda scale, elem_id: gr.Column(scale=scale, elem_id=elem_id) def pdf_to_markdown(file_obj): try: shutil.rmtree('./private_upload/') except: pass time_tag = time.strftime("%Y-%m-%d-%H-%M-%S", time.localtime()) os.makedirs(f'private_upload/{time_tag}', exist_ok=True) file_name = os.path.basename(file_obj.name) destination = f'private_upload/{time_tag}/{file_name}' shutil.copy(file_obj.name, destination) global pquery pquery.ingest(destination) with open(destination, "rb") as f: pdf = base64.b64encode(f.read()).decode('utf-8') pdf_display = f'' return [pdf_display, gr.update(visible=False),gr.update(visible=True),gr.update(visible=True),gr.update(visible=True), gr.update(visible=True),gr.update(visible=True)] # 清空 cle = lambda :"" with gr.Blocks(title="Chat With Pdf") as demo: gr.HTML(title_html) file = gr.File() with gr_L1(): with gr_L2(scale=1.5, elem_id="gpt-chat"): out = gr.Markdown() with gr_L2(scale=1, elem_id="gpt-chat"): title = gr.Markdown("""

文档问答

""", visible=False) chatbot = gr.Chatbot(scale=3, height=600, visible=False) with gr_L1(): message = gr.Textbox(placeholder="Input question here.", scale=10, visible=False) state = gr.State([]) submit = gr.Button("发送", scale=1, visible=False) file.upload(pdf_to_markdown, file, [out, file, out, title, chatbot, message, submit]) submit.click(chatgpt_clone, inputs=[message, state, chatbot], outputs=[chatbot, state, message]) demo.launch(share=True)