pf-ai-chatbot

AI chatbot for PatternFly. Based on Rasa open source framework.

How to run

Initial setup

• create a python virtual environment: python3 -m venv my_venv
• start a python virtual environment: source my_venv/bin/activate
• install rasa and other necessary libraries into this virtual environment

Start a chatbot

• go to chatbot directory: cd chatbot
• train a model: rasa train
• in a separate terminal, start actions: rasa run actions
• start the model rasa shell

Content

• chatbot
  • rasa framework project
  • domain.yml includes list of user intents, chatbot responses, list of actions, entities, slots, forms, session_config, ...
  • data/nlu.yml defines user intents and examples (sentences) for each intent
  • data/rules.yml defines what action should the chatbot do as a reaction on an intent = strict
  • data/stories.yml defines what action should the chatbot do as a reaction on an intent. Can have more steps than one intent-action.
  • actions/actions.py defines custom actions (how to respond on intents in a more customizable way)
  • helpers/generate_component_intents.py creates intents based on the various components in actions/components.json file
• webscraper
  • main.py downloads data about PatternFly components and saves them to chatbot/actions/components.json
  • requires a patternfly-org website running at localhost:8003