install.rst

Installation

This section includes information for installing beat.editor a couple different ways.

Method 1: Virtualbox Image (easiest)

This method is the easiest - one command will generate a new Virtualbox image with beat.editor installed.

Requirements

1. Working internet connection
2. Virtualbox installed and working (might need configuration)
3. Packer installed
4. About 3G of disk space (at the time of writing, 2.7G needed)
5. Git

Installation Steps

1. Clone the repository holding the scripts:

   $ git clone https://gitlab.idiap.ch/beat/beat.tutorial.git

2. Change directory to the folder holding the scripts:

   $ cd beat.tutorial/packer_scripts

3. Generate the Virtualbox image:

   $ packer-io build ubuntu-18.04-beat.json

This will generate the OVF file in packer_scripts/output-virtualbox-iso.

To use this OVF file in the Virtualbox GUI, select "Import Appliance" under the "File" menu in Virtualbox and choose to import the OVF file.

The username and password for the image are both beat.

The BEAT commands, beat and beat editor serve, should be available in the shell in the VM. The Firefox browser is available, as is the Vim, Emacs, Nano, and Gedit editors.

Method 2: Conda (easyish)

This method integrates with one of your Conda environments' packages on your local machine, but has more strict requirements.

Requirements

1. Linux or MacOS (Docker features not available on MacOS for now)
2. Working internet connection
3. Useable Conda setup (see Bob's documentation on it)
4. Firefox or Chrome browser
5. Docker installed and available to the current user (not strictly necessary)

Installation Steps

1. Activate the Conda environment you will be adding beat.editor to

2. Add two new Conda channels:

   $ conda config --env --add channels https://www.idiap.ch/software/beat/conda/label/beta
   $ conda config --env --add channels https://www.idiap.ch/software/beat/conda

3. Install the beat.editor package:

   $ conda install beat.editor

4. The BEAT commands, beat and beat editor serve, should be available.

Method 3: For developers

This method is for developers looking to work on beat.editor.

Requirements

1. Linux or MacOS
2. Working internet connection
3. Useable Conda setup (see Bob's documentation on it)
4. Up-to-date Firefox and Chrome browsers (both need headless support, a relatively recent feature)
5. Docker installed and available to the current user
6. A relatively recent version of NodeJS & NPM.

Installation Steps

1. Clone the beat.editor repository:

   $ git clone git@gitlab.idiap.ch:beat/beat.editor.git

2. Clone the bob.admin repository, necessary for building a development Conda environment:

   $ git clone git@gitlab.idiap.ch:bob/bob.admin.git

3. Use bob.admin to create a development Conda environment named beatedit for Python 3.6:

   $ conda activate base
   $ conda install pyyaml # install a dependency that isn't fetched by default
   $ cd beat.editor
   $ ../bob.admin/conda/conda-bootstrap.py --overwrite --python=3.6 beatedit

4. Activate the created environment and use the buildout command from bob.buildout in the new environment to generate the executables for beat.editor:

   $ conda activate beatedit
   $ buildout

5. Note for development mode: Add the line dependent-scripts = true then use the buildout command to ensure to generate all the executables in the bin/ folder even beat for beat.editor:

   $ buildout

6. In bin/ you should several new executables for the project. The Python setup part is done.

7. Go into the root folder for the Javascript and install the dependencies via NPM:

   $ cd conda/js
   $ npm install

8. Test the JS to make sure everything is working properly:

   $ npm test

9. Assuming the tests pass, you can start the development webpack server:

   $ npm start

10. In a separate terminal tab/window, go back to the root of the beat.editor project and run the web server in development mode:

    $ ./bin/beat editor serve --dev

11. You should now be able to go localhost:9101 in your browser to see the beat.editor web app, now served by the webpack dev server.