Create a virtual environment and install dependencies

python3 -m venv venv

source venv/bin/activate (Linux/UNIX)

venv\Scripts\activate (Windows)

pip install -r requirements.txt

Train model

python3 -m rasa train

Run rasa in terminal

python3 -m rasa shell

Train rasa with examples while running it on terminal

python3 -m rasa interactive

Run Rasa on raspberry pi

python3 -m rasa run --enable-api --cors="*" In a new terminal tab, run the following command in the directory - ngrok http 5005

Connect to pi through ssh and run the following command in the directory - cd priya code . Copy the https url and paste it in the frontend code in the src/App.js file in the axios.post() function. (there are 2 axios commands, search ngrok and you'll find them. Replace those ngrok links with the new one.)

On the pi, run the following command in the terminal - cd priya npm install npm start

This should open a browser window with chatbot screen. If it doesn't, open a browser and go to localhost:3000 and you should see the chatbot screen.

Test the chatbot all you want!

Run RASA on web (needs to be updated, use shell/interactive in the meantime)

python3 -m rasa run --enable-api --cors="*" Clone bizdamn/priya and run the following command in the directory - npm install npm start

How to add data to dataset

Add intent examples to data/nlu.yml

- intent: new_intent_name
  examples: |
    - some intent examples
    - make sure the examples help train the AI
    - and has all necessary keywords as well as variations of the keywords
    - minimum 5 examples are recommended
    - these examples all correspond to the same intent and will be mapped to a particular response so kindly do not club various examples together.

Add intent name and response to domain.yml

intents:
  - existing_intent_name
  - new_intent_name
  .
  .
  .
responses:
  utter_new_intent_name:
  - text: "The response to new_intent_name"
  - image: "https://image.com/image.png" # optional

Add rules to data/rules.yml to map the intent and response

- rule: Do action when asked about intent
  steps:
  - intent: new_intent_name
    action: utter_new_intent_name

TODO

• Add frontend
• Update Frontend for Pi
• Add voice support (speech synthesis and speech recognition)
• Increase dataset (and add buttons and slots/entities)
• Add a fallback intent.
• Add university map.
• Look into better STT and TTS.
• Connect Mic and Speakers
• Give the Pi a body
• Add the frontend code to this repository

Optional/Future Work:

• Wake word
• Interface with mouth movements and face (refer something like UwU)
• Design 3d printed body

Known bugs/Issues

• Does not work on Safari (M1 Mac)

Connecting to raspberry pi through ssh

follow this link - https://www.onlogic.com/company/io-hub/how-to-ssh-into-raspberry-pi/