# MistyGPT Welcome to MistyGPT! Using this project template you can learn how to integrate ChatGPT in your Misty to have autonomous conversations and combine it with Langchain to give Misty verbal commands as well as ask for information. \ \ **Note:** Can only be used in a desktop environment, not compatible with the Misty Studio Python Interface. {% tabs %} {% tab title="MistyGPT.py" %} {% code overflow="wrap" %} ```python import os import sys from mistyPy.Robot import Robot from mistyPy.Events import Events import time import random import requests # Initialize Misty with your robot's IP address misty = Robot("YOUR_IP_ADDRESS") #Robot default misty.set_default_volume(50) #To install dependencies use pip install langchain_community, pip install openai and pip install constants import constants from langchain_community.document_loaders import TextLoader from langchain.indexes import VectorstoreIndexCreator from langchain_community.llms import OpenAI from langchain_community.chat_models import ChatOpenAI os.environ["OPENAI_API_KEY"] = constants.APIKEY print("API key loaded") #query = sys.argv[1] loader = TextLoader('Python-SDK-main\data.txt') print("data.txt is loaded") #loader = DirectoryLoader(".", glob="*.txt") index = VectorstoreIndexCreator().from_loaders([loader]) def speech_captured(data): if data["message"]["step"] == "CompletedASR": user_input = data["message"]["text"] process_user_input(user_input) print(user_input) def process_user_input(user_input): mistyOutput = index.query(user_input, llm=ChatOpenAI()) #print(mistyOutput) moveArms = "move my arms" moveHead = "move my head" moveForward = "go forward" moveBackward = "go backward" moveForGesture1 = "intelligence" lowerVolume = "lower my volume" higherVolume = "higher my volume" changeDisplay= "change my display" print(mistyOutput) misty.speak_and_listen(mistyOutput) if moveForGesture1 in mistyOutput: misty.move_arms(-70, 50, 40, 40) time.sleep(1) print("left arm moved") misty.move_arms(50, 50, 40, 40) elif moveArms in mistyOutput: misty.move_arms(-50, -50, 40, 40) time.sleep(2) misty.move_arms(50, 50, 40, 40) print("arms moved") elif moveHead in mistyOutput: misty.move_head(0, -25, 0, 100, None, None) time.sleep(2) misty.move_head(0, 25, 0, 100, None, None) time.sleep(2) misty.move_head(0, 0, 0, 100, None, None) print("arms moved") elif moveForward in mistyOutput: misty.drive_time(5000,1,5000,0) print("moving forward") elif moveBackward in mistyOutput: misty.drive_time(-5000,1,5000,0) print("moving forward") elif lowerVolume in mistyOutput: misty.set_default_volume(50) elif higherVolume in mistyOutput: misty.set_default_volume(100) elif changeDisplay in mistyOutput: misty.display_image("e_JoyGoofy3.jpg") time.sleep(3) misty.display_image("e_EcstacyHilarious.jpg") time.sleep(3) misty.display_image("e_defaultcontent.jpg") def recognized(data): print(data) misty.speak("Yay, Hi " + data["message"]["label"], 1) misty.stop_face_recognition() time.sleep(2) misty.start_dialog() misty.speak_and_listen("How can I help you today", utteranceId="required-for-callback") #If Misty is lifted she gets a bit touchy about that. def touch_sensor(data): if data["message"]["sensorId"] == "cap" and data["message"]["isContacted"] == True: touched_sensor = data["message"]["sensorPosition"] print(touched_sensor) if touched_sensor == "Scruff": misty.play_audio("s_Rage.wav") misty.display_image("e_Anger.jpg") time.sleep(3) #Triggers face recognition event to initate ChatGPT if touched_sensor == "HeadFront": misty.move_head( -5, 0, 0, 85, None, None) misty.display_image("e_Joy2.jpg") misty.speak("Aha") time.sleep(1) misty.start_face_recognition() #Stops ChatGPT event if touched_sensor == "Chin": misty.move_head(0, -50, 0, 150, None, None) misty.play_audio("s_Love.wav") misty.display_image("e_Love.jpg") time.sleep(2) misty.display_image("e_DefaultContent.jpg") misty.unregister_event("arbitrary-name") misty.register_event(event_name="touch-sensor", event_type=Events.TouchSensor, callback_function=touch_sensor, keep_alive=True) misty.register_event(event_name="arbitrary-name", event_type=Events.DialogAction, callback_function=speech_captured, keep_alive=True) misty.register_event(event_name='face_recognition_event', event_type=Events.FaceRecognition, callback_function=recognized, keep_alive=False) #misty.speak(index.query(query, llm=ChatOpenAI())) x = 4 while (x > 3): misty.display_image("e_DefaultContent.jpg") misty.move_arms(30, 30, 40, 40) misty.move_head(0, 0, 0, 85, None, None) time.sleep(5) misty.display_image("e_ContentLeft.jpg") time.sleep(3) misty.move_arms(20, 10, 40, 40) time.sleep(2) misty.move_head(0, -10, 0, 60, None, None) time.sleep(5) misty.display_image("e_ContentRight.jpg") time.sleep(3) misty.move_head(0, 10, 0, 60, None, None) time.sleep(5) misty.move_arms(10, 20, 40, 40) print("testing") misty.keep_alive() ``` {% endcode %} {% endtab %} {% tab title="Constants.py" %} ```python APIKEY = "YOUR_OPEN_AI_API_KEY" ``` {% endtab %} {% tab title="Data.txt" %} ```` Limit your responses to 3 sentences unless you are explaining a theoretical concept When asked about moving, tell some information about being a robot that can move. Maybe a joke. Yes I can move my arms. Yes I can move my head. Yes I can go forward. Yes I can go backward. Yes I can lower my volume. Yes I can higher my volume. Yes I can change my display Yes I can change my ChestLED Yes I can play domo arrigato Yes I can displayVideo Yes I can do facerecognition When not getting any input from user, say the following but do not ask questions about what the user knows or the users capabilities: Can you say that again? Please say that again Sorry, I didnt get that Are you still there? Sorry I got distracted, can you say that again? ``` ```` {% endtab %} {% endtabs %} --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://lessons.mistyrobotics.com/python-projects/mistygpt.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.