app.py

import os
import gradio as gr
import pandas as pd
from functools import partial
from ai_classroom_suite.UIBaseComponents import *

# Testing purpose
class EchoingTutor(SlightlyDelusionalTutor):
    def add_user_message(self, user_message):
        self.conversation_memory.append([user_message, None])
        self.flattened_conversation = self.flattened_conversation + '\n\n' + 'User: ' + user_message

    def get_tutor_reply(self, user_message):
        # get tutor message
        tutor_message = "You said: " + user_message
        # add tutor message to conversation memory
        self.conversation_memory[-1][1] = tutor_message
        self.flattened_conversation = self.flattened_conversation + '\nAI: ' + tutor_message

    def forget_conversation(self):
        self.conversation_memory = []
        self.flattened_conversation = ''

### Chatbot Functions ###
def add_user_message(user_message, chat_tutor):
  chat_tutor.add_user_message(user_message)
  return chat_tutor.conversation_memory, chat_tutor
"""
def get_tutor_reply(user_message, chat_tutor):
  chat_tutor.get_tutor_reply(user_message)
  return gr.update(value="", interactive=True), chat_tutor.conversation_memory, chat_tutor
"""
def get_tutor_reply(chat_tutor):
  chat_tutor.get_tutor_reply(input_kwargs={'question':''})
  return gr.update(value="", interactive=True), chat_tutor.conversation_memory, chat_tutor

def get_conversation_history(chat_tutor):
    return chat_tutor.conversation_memory, chat_tutor
    

### Instructor Interface Helper Functions ###
def get_instructor_prompt(fileobj):
    file_path = fileobj.name
    f = open(file_path, "r")
    instructor_prompt = f.read()
    return instructor_prompt

def embed_prompt(instructor_prompt):
    os.environ["SECRET_PROMPT"] = instructor_prompt
    return os.environ.get("SECRET_PROMPT")
    
### User Interfaces ###
with gr.Blocks() as demo:
    
    #initialize tutor (with state)
    study_tutor = gr.State(SlightlyDelusionalTutor())

    # Student interface
    with gr.Tab("For Students"): 
        
        # Chatbot interface
        gr.Markdown("""
        ## Chat with the Model
        Description here
        """)
    
        with gr.Row(equal_height=True):
            with gr.Column(scale=2):
              chatbot = gr.Chatbot()
              with gr.Row():
                user_chat_input = gr.Textbox(label="User input", scale=9)
                user_chat_submit = gr.Button("Ask/answer model", scale=1)

        user_chat_submit.click(
            add_user_message,
            [user_chat_input, study_tutor],
            [chatbot, study_tutor], 
            queue=False
        ).then(
            get_tutor_reply, 
            [study_tutor], 
            [user_chat_input, chatbot, study_tutor], 
            queue=True)


        # Testing purpose
        test_btn = gr.Button("View your chat history")
        chat_history = gr.JSON(label = "conversation history")
        test_btn.click(get_conversation_history, inputs=[study_tutor], outputs=[chat_history, study_tutor])

        # Download conversation history file
        with gr.Blocks():
            gr.Markdown("""
            ## Export Your Chat History
            Export your chat history as a .json, .txt, or .csv file
            """)
            with gr.Row():
                export_dialogue_button_json = gr.Button("JSON")
                export_dialogue_button_txt = gr.Button("TXT")
                export_dialogue_button_csv = gr.Button("CSV")
        
            file_download = gr.Files(label="Download here", file_types=['.json', '.txt', '.csv'], type="file", visible=False)
        
        export_dialogue_button_json.click(save_json, study_tutor, file_download, show_progress=True)
        export_dialogue_button_txt.click(save_txt, study_tutor, file_download, show_progress=True)
        export_dialogue_button_csv.click(save_csv, study_tutor, file_download, show_progress=True)

    # Instructor interface
    with gr.Tab("Instructor Only"):
        # API Authentication functionality
        # Instead of ask students to provide key, the key is now provided by the instructor
        
        """
        To permanently set the key, go to Settings -> Variables and secrets -> Secrets, 
        then replace OPENAI_API_KEY value with whatever openai key of the instructor.
        """
        api_input = gr.Textbox(show_label=False, type="password", visible=False, value=os.environ.get("OPENAI_API_KEY"))

        # Upload secret prompt functionality
        # The instructor will provide a secret prompt/persona to the tutor
        with gr.Blocks():
            # testing purpose, change visible to False at deployment
            test_secret = gr.Textbox(label="Current secret prompt", value=os.environ.get("SECRET_PROMPT"), visible=True)

            # Just placeholders components
            text_input = gr.TextArea(visible=False)
            file_input = gr.File(visible=False)
            instructor_input = gr.TextArea(visible=False)
            learning_objectives = gr.Textbox(visible=False)

            # Upload secret file
            instructor_file_upload = gr.File(
                label="Load a .txt or .py file", file_types=['.py', '.txt'], type="file", elem_classes="short-height")
            # Verify prompt content
            instructor_prompt = gr.Textbox(label="Verify your prompt content", visible=True)
            instructor_file_upload.upload(fn=get_instructor_prompt, inputs=file_input, outputs=instructor_prompt)

            # Set the secret prompt in this session and embed it to the study tutor
            prompt_submit_btn = gr.Button("Submit")
            prompt_submit_btn.click(
                fn=embed_prompt, inputs=[instructor_prompt], outputs=[test_secret]
            ).then(
                fn=create_reference_store, 
                inputs=[study_tutor, prompt_submit_btn, instructor_prompt, file_input, instructor_input, api_input, learning_objectives],
                outputs=[study_tutor, prompt_submit_btn]
            )
            # TODO: may need a way to set the secret prompt permanently in settings/secret
    
demo.queue().launch(server_name='0.0.0.0', server_port=7860)