Update app.py

app.py

@@ -1,31 +1,16 @@
-import openai, subprocess
+import openai
 import gradio as gr
 from gradio.components import Audio, Textbox
-
 import os
 import re
-import tiktoken
 from transformers import GPT2Tokenizer
-tokenizer = GPT2Tokenizer.from_pretrained('gpt2')
 import whisper
 import pandas as pd
-import os
 from datetime import datetime, timezone, timedelta
-# import dropbox
-# from notion_client import Client
 import notion_df
 
-API_KEY = os.environ["API_KEY"]
-# # Define your API key
-
-# my_API_KEY = os.environ["NOTION"]
-# notion = Client(auth=my_API_KEY)
-# # find the page you want to upload the file to
-
-# ACCESS_TOKEN = os.environ["ACCESS_TOKEN"]
-# dbx = dropbox.Dropbox(ACCESS_TOKEN)
-
 openai.api_key = os.environ["OPENAI_API_KEY"]
+tokenizer = GPT2Tokenizer.from_pretrained('gpt2')
 
 initial_message = {"role": "system", "content": 'You are a USMLE Tutor. Respond with ALWAYS layered "bullet points" (listing rather than sentences) to all input with a fun mneumonics to memorize that list. But you can answer up to 1200 words if the user requests longer response.'}
 messages = [initial_message]
@@ -35,8 +20,6 @@ answer_count = 0
 # set up whisper model
 model = whisper.load_model("base")
 
-
-
 def num_tokens_from_messages(messages, model="gpt-3.5-turbo-0301"):
     """Returns the number of tokens used by a list of messages."""
     try:
@@ -58,175 +41,129 @@ def num_tokens_from_messages(messages, model="gpt-3.5-turbo-0301"):
 See https://github.com/openai/openai-python/blob/main/chatml.md for information on how messages are converted to tokens.""")
 
 def transcribe(audio, text):
-
     global messages
     global answer_count
-    transcript = None
-    
+
     if audio is not None:
         audio_file = open(audio, "rb")
         transcript = openai.Audio.transcribe("whisper-1", audio_file, language="en")
-        # transcript = model.transcribe(audio_file, language="english")
         messages.append({"role": "user", "content": transcript["text"]})
-        
-    if transcript is None:
+
+    if text is not None:
         # Split the input text into sentences
         sentences = re.split("(?<=[.!?]) +", text)
-    
-        # Initialize a list to store the tokens
-        input_tokens = []
-    
-        # Add each sentence to the input_tokens list
-        for sentence in sentences:
-            # Tokenize the sentence using the GPT-2 tokenizer
-            sentence_tokens = tokenizer.encode(sentence)
-            # Check if adding the sentence would exceed the token limit
-            if len(input_tokens) + len(sentence_tokens) < 1440:
-                # Add the sentence tokens to the input_tokens list
-                input_tokens.extend(sentence_tokens)
-            else:
-                # If adding the sentence would exceed the token limit, truncate it
-                sentence_tokens = sentence_tokens[:1440-len(input_tokens)]
-                input_tokens.extend(sentence_tokens)
-                break
-        # Decode the input tokens into text
-        input_text = tokenizer.decode(input_tokens)
-    
-        # Add the input text to the messages list
-        messages.append({"role": "user", "content": input_text})
-
-    # Get the current date and time in the local timezone
-    now_local = datetime.now()
-    # Create a timezone object for Eastern Time (ET)
-    et_tz = timezone(timedelta(hours=-5))
-    # Adjust the date and time to Eastern Time (ET)
-    now_et = now_local.astimezone(et_tz)
-        
-    # Check if the accumulated tokens have exceeded 2096
+
+        # Tokenize the sentences using the GPT-2 tokenizer
+        sentence_tokens = [tokenizer.encode(sentence) for sentence in sentences]
+
+        # Flatten the list of tokens
+        input_tokens = [token for sentence in sentence_tokens for token in sentence]
+
+        # Check if adding the input tokens would exceed the token limit
+        num_tokens = num_tokens_from_messages(messages)
+        if num_tokens + len(input_tokens) > 2200:
+            # Reset the messages list and answer counter
+            messages = [initial_message]
+            answer_count = 0
+            input_text = 'Can you click the Submit button one more time? (say Yes)'
+            messages.append({"role": "user", "content": input_text})
+        else:
+            # Add the input tokens to the messages list
+            input_text = tokenizer.decode(input_tokens)
+            messages.append({"role": "user", "content": input_text})
+
+    # Check if the accumulated tokens have exceeded the limit
     num_tokens = num_tokens_from_messages(messages)
     if num_tokens > 2096:
         # Concatenate the chat history
         chat_transcript = ""
         for message in messages:
             if message['role'] != 'system':
-                chat_transcript += f"[ANSWER {answer_count}]" + message['role'] + ": " + message['content'] + "\n\n"
+                chat_transcript += f"[ANSWER {answer_count}]{message['role']}: {message['content']}\n\n"
         # Append the number of tokens used to the end of the chat transcript
-            
-        chat_transcript_copy = chat_transcript
-        chat_transcript_copy += f"Number of tokens used: {num_tokens}\n\n"
+        chat_transcript += f"Number of tokens used: {num_tokens}\n\n"
         
-        # Get the current UTC time
-        utc_time = datetime.now(timezone.utc)
-        # Convert to Eastern Time Zone
-        eastern_time = utc_time + timedelta(hours=-5)
-        # Format as string (YY-MM-DD HH:MM)
-        published_date = eastern_time.strftime('%m-%d-%y %H:%M')
-        # string dataframe?
+        # Get the current time in Eastern Time (ET)
+        now_et = datetime.now(timezone(timedelta(hours=-5)))
+        # Format the time as string (YY-MM-DD HH:MM)
+        published_date = now_et.strftime('%m-%d-%y %H:%M')
+        
+        # Upload the chat transcript to Notion
         df = pd.DataFrame([chat_transcript])
         notion_df.upload(df, 'https://www.notion.so/personal-5e3978680ca848bda844452129955138?pvs=4', title=str(published_date), api_key=API_KEY)
         
-    if num_tokens > 2200:
         # Reset the messages list and answer counter
         messages = [initial_message]
         answer_count = 0
-        input_text = 'Can you click the Submit button one more time? (say Yes)'
-        # Add the input text to the messages list
-        messages.append({"role": "user", "content": input_text})
 
     # Increment the answer counter
     answer_count += 1
-    # Add the answer counter to the system message
+
+    # Generate the system message using the OpenAI API
     system_message = openai.ChatCompletion.create(
         model="gpt-3.5-turbo",
         messages=messages,
         max_tokens=2000
     )["choices"][0]["message"]
+
+
+
     # Add the system message to the messages list
-    messages.append(system_message)
+    messages.append({"role": "system", "content": system_message})
 
     # Concatenate the chat history
     chat_transcript = ""
     for message in messages:
         if message['role'] != 'system':
-            chat_transcript += f"[ANSWER {answer_count}]" + message['role'] + ": " + message['content'] + "\n\n"
-    # Append the number of tokens used to the end of the chat transcript
+            chat_transcript += f"[ANSWER {answer_count}]{message['role']}: {message['content']}\n\n"
     
-    with open("conversation_history.txt", "a") as f:
-        f.write(chat_transcript)
-        
-    chat_transcript_copy = chat_transcript
-    chat_transcript_copy += f"Number of tokens used: {num_tokens}\n\n"
-    filename = datetime.now().strftime("%m%d%y_%H:%M_conversation_history.txt")
+    # Append the number of tokens used to the end of the chat transcript
+    num_tokens = num_tokens_from_messages(messages)
+    chat_transcript += f"Number of tokens used: {num_tokens}\n\n"
     
-    # dbx.files_upload(chat_transcript_copy.encode('utf-8'), f'/{filename}', mode=dropbox.files.WriteMode.overwrite, autorename=False, client_modified=None, mute=False)
-    # dbx.files_upload(chat_transcript_copy.encode('utf-8'), '/conversation_history.txt', mode=dropbox.files.WriteMode.overwrite, autorename=False, client_modified=None, mute=False)
+    # Get the current time in Eastern Time (ET)
+    now_et = datetime.now(timezone(timedelta(hours=-5)))
+    # Format the time as string (YY-MM-DD HH:MM)
+    published_date = now_et.strftime('%m-%d-%y %H:%M')
     
-    # Get the current UTC time
-    utc_time = datetime.now(timezone.utc)
-    # Convert to Eastern Time Zone
-    eastern_time = utc_time + timedelta(hours=-5)
-    # Format as string (YY-MM-DD HH:MM)
-    published_date = eastern_time.strftime('%m-%d-%y %H:%M')
-
-    # Get the current UTC time
-    utc_time = datetime.now(timezone.utc)
-    # Convert to Eastern Time Zone
-    eastern_time = utc_time + timedelta(hours=-5)
-    # Format as string (YY-MM-DD HH:MM)
-    published_date = eastern_time.strftime('%m-%d-%y %H:%M')
-    # string dataframe
-    df = pd.DataFrame([chat_transcript_copy])
+    # Upload the chat transcript to Notion
+    df = pd.DataFrame([chat_transcript])
     notion_df.upload(df, 'https://www.notion.so/personal-5e3978680ca848bda844452129955138?pvs=4', title=str(published_date), api_key=API_KEY)
-
-    return chat_transcript
+    
+    # Reset the messages list and answer counter if the token limit is exceeded
+    if num_tokens > 2096:
+        messages = [initial_message]
+        answer_count = 0
+    else:
+        # Increment the answer counter
+        answer_count += 1
+    
+    # Generate the system message using the OpenAI API
+    system_message = openai.Completion.create(
+        engine="text-davinci-002",
+        prompt=[{"text": f"{message['role']}: {message['content']}\n\n"} for message in messages],
+        temperature=0.7,
+        max_tokens=2000,
+        n=1,
+        stop=None,
+    )[0]["text"]
+    
+    # Add the system message to the messages list
+    messages.append({"role": "system", "content": system_message})
 
 
 audio_input = Audio(source="microphone", type="filepath", label="Record your message")
 text_input = Textbox(label="Type your message", max_length=4096)
 
 output_text = gr.outputs.Textbox(label="Response")
-output_audio = Audio()
 
 iface = gr.Interface(
     fn=transcribe,
     inputs=[audio_input, text_input],
-    # outputs=(["audio", "text"]),
     outputs="text",
-    title="Your Excellence Never Abates (YENA)",
-    description="Talk to the AI Tutor YENA",
-    capture_session=True,
-    autoplay=True)
-
+    title="YENA",
+    description="Tutor YENA")
 
 # Launch Gradio interface
 iface.launch()
 
-
-
-
-# from transformers import pipeline, T5Tokenizer
-# import pyttsx3
-# import threading
-# import time
-
-
-
-
-# Set up speech engine
-# engine = pyttsx3.init()
-
-# def speak(text):
-#     # Get the current rate of the engine
-#     rate = engine.getProperty('rate')
-
-#     # Calculate the estimated time in seconds based on the length of the message and the current rate
-#     estimated_time = len(text) / (rate / 10)
-
-#     # Speak the text using the text-to-speech engine
-    
-#     engine.say(text)
-#     engine.runAndWait()
-#     if engine._inLoop:
-#         # Wait for the speech engine to finish speaking
-#         time.sleep(estimated_time*1.5)
-#         engine.endLoop()