Why set up a virtual environment? Virtual environments let you scope Python packages to a project directory, instead of installing the package on the system (global) level. This means you can use different versions of a package for different projects and helps to prevent conflicts.

We’re using venv in this demo to create the virtual environment, because it’s built into the Python standard library. This process is recommended for Mac and Linux users.

Note for Windows Users: If you’re using Windows, we recommend using Anaconda to set up your virtual environment instead of venv. You can find detailed instructions here and in this video tutorial to help. The important parts to note are installing ujson, tensorflow, and the Visual Studio C++ build tools. Also, it is important to activate the option “Add Anaconda to my PATH environment variable" during the installation of Anaconda.

To create an environment in conda type conda create --name pydata_env python==3.7.6

Be sure to clone this project using git clone https://github.com/kunan-sa/BotDelClima-PyConAr2020.git and install the dependencies (instructions below).

• Inside the root directory of this project
• Create the virtual environment
  • python3 -m venv ./venv
• Activate the virtual environment
  • source ./venv/bin/activate

NB: In order to leave the virtual environment simply use the command deactivate; delete/ remove the environment with rm -rf venv.

To install Rasa Open Source

• Make sure you are in the root directory of this project
• Upgrade pip
  • pip install --upgrade pip
• Install dependencies
  • pip install rasa

Installing Rasa for Windows Users: This video provides a guide to installing Rasa on a Windows operating system.

You need to register to obtain the api key that you will use. https://openweathermap.org/appid then, put this key in the constant API_KEY on api_key.py file.

Before we can start talking to the assistant, we need to train the model and start the servers.

• Train the model

  rasa train

• Once the model has finished training, run the action server in one terminal

  rasa run actions

• In another terminal start the Rasa Open Source server. This will also start a session to chat with the assistant on the command line.

  rasa shell

Talk to the bot! To stop the Rasa Open Source server and the chat session, type /stop