Lab Assistant

An Alexa skill for managing your open-source projects hosted on GitLab.com using your voice.

Installation

Enable the Lab Assistant skill on your Amazon site of choice:

• Amazon.com
• Amazon.ca
• Amazon.co.uk
• Amazon.com.au
• Amazon.in

Don't see your local Amazon store? You can help out by translating Lab Assistant into your language! See the Contributing section below for more details.

Usage

• "Alexa, open Lab Assistant"
• "Alexa, ask Lab Assistant to [any of the commands below]"

Todos

• "Todos"
• "Do I have any todos?"
• "How many todos do I have?"
• "What does my day look like?"
• "Read me my todos"

Issues

• "Issues"
• "Do I have any issues assigned to me?"
• "How many issues do I have?"
• "Read me my issues"

Merge Requests

• "Merge requests"
• "Do I have any merge requests?"
• "How many merge requests do I have"
• "Read me my merge requests"

Pipelines

• "Run a new pipeline"
• "Execute a deployment"
• "Create a test pipeline"
• "Initiate a test build"
• "Begin a new analysis job on my project"

This command in particular is quite flexible in how it can be invoked. To get a better understanding of all the ways you can run a pipeline on your project, see the RunPipelineIntent entry in this skill's grammar file.

Pipeline variables

Lab Assistant will always trigger your pipeline with the environment variable LAB_ASSISTANT=true.

You can trigger your pipeline with an additional environment variable in the format LAB_ASSISTANT_<VARIABLE>=true by including this variable in your invocation:

• "Run a new test pipeline"
• "Execute a test"
• "Launch a new test job"
• "Perform the test process"
• "Start a test on our code"

All of the invocations above will result in your pipeline being triggered with the following variables:

• LAB_ASSISTANT=true
• LAB_ASSISTANT_TEST=true

Pro tip: To avoid misunderstandings, use simple, one-word variables that are easy for Alexa to understand. Some good examples are "deployment", "test", "analysis", or "lint".

Using these variables in your .gitlab-ci.yml

Here's an example of how you might use these variables in your .gitlab-ci.yml:

a normal job:
  script:
    - echo 'This job will always run'

a normal job:
  script:
    - echo 'This job will not run when trigged by Lab Assistant'
  except:
    variables:
      - $LAB_ASSISTANT

a lab assistant job:
  script:
    - echo 'This job will only run when triggered by Lab Assistant'
  only:
    variables:
      - $LAB_ASSISTANT

a lab assistant deploy job:
  script:
    - echo 'This job will only run when triggered by Lab Assistant AND when the $LAB_ASSISTANT_DEPLOY variable is set'
  only:
    variables:
      - $LAB_ASSISTANT && $LAB_ASSISTANT_DEPLOY

a lab assistant test job:
  script:
    - echo 'This job will only run when triggered by Lab Assistant AND when the $LAB_ASSISTANT_TEST variable is set'
  only:
    variables:
      - $LAB_ASSISTANT && $LAB_ASSISTANT_TEST

If only a few of your jobs should run when triggered by Lab Assistant, and you'd like to avoid the boilerplate of adding except: variables: [$LAB_ASSISTANT] to all of the jobs the shouldn't be executed by Lab Assistant, you can make use of a default job definition like this:

.except-default: &except-default
  except:
    variables:
      - $LAB_ASSISTANT

build:
  <<: *except-default
  script:
    - echo 'This job will run as long as the pipeline was not triggered by Lab Assistant'

test:
  <<: *except-default
  script:
    - echo 'Same as above'

deploy:
  script:
    - echo 'This job will always run, regardless of how the pipeline is triggered'

report:
  script:
    - echo 'This job will ONLY run when the pipeline is triggered by Lab Assistant'
  only:
    variables:
      - $LAB_ASSISTANT

For a real-life example, see this project's .gitlab-ci.yml.

For more information on how to use only and except in a .gitlab-ci.yml file, see this page.

Pipeline ref

All pipelines are executed against the master branch of the project you specify. Interested in the ability to execute pipelines against branches other than master? Open an issue and let me know.

Selecting a project

The first time you run a command that runs in the context of a project, Lab Assistant will ask you for the project's ID. You can find your project's ID on the project's page under the project name:

Lab Assistant will remember this ID for future invocations and won't ask you for it again.

If you'd like to switch projects, you can say:

• "Switch projects"
• "Select my default project"
• "Choose a different project"
• "Select project 1234"
• "Change my default project to project number 1234"

Connecting to your GitLab.com account

Any command that requires you to connect your GitLab.com account will automatically prompt you to login, so you shouldn't have to worry about manually connecting your account. If for some reason you do want to explicitly connect your account, you can say:

• "Connect my account"
• "Log in to my GitLab.com account"
• "Authenticate to GitLab.com"

Disconnecting your GitLab.com account

To disconnect your account, disable and re-enable the Lab Assistant skill.

FAQ

Can I use this skill for self-managed GitLab instances?

No, this skill only integrates with GitLab.com. This is because Alexa Skills must specify an OAuth provider at build time.

If you do want to use this skill with a self-managed GitLab instance, you can clone this repo and deploy this code as a separate skill that is pointed at your GitLab instance.

Why does the skill ask for the project ID instead of asking for the project's name?

Alexa's speech detection is really good at some things and really terrible at others. Alexa is really good at accurately understanding numbers; Alexa is really bad at accurately understanding open-ended phrases, like project names. To cut down on frustration, this skill asks for project IDs to ensure you can select the project you want.

Developing

Building

1. Clone this repo
2. cd lab-assistant/Lab_Assistant/lambda/custom
3. npm install
4. npm run build

Building the model

This skill's model is generated using the Alexa Skill Utterance and Schema Generator library. This library takes en-US.grammar and outputs en-US.json. To run this process, run npm run model. Note that you will need a JDK installed for this command to run; it executes alexa-generate.jar which is included in this repository.

Testing

The easiest way to develop on this project is using test-driven development through Jest. You can run the tests using npm run test or npm run test-watch. See the existing tests for some examples.

i18n

This project uses i18next for internationalization ("i18n"). i18next-scanner is used to extract the strings directly from the source into this project's i18n directory. You can run this extraction process by building the project (npm run build) and then running npm run translate. Alternatively, you can run npm run build-and-translate.

To make this skill available in all English-speaking Amazon stores, a utility script (./Lab_Assistant/lambda/custom/utility-scripts/copy-english-files.js) can be run using npm run copy-english-files to copy the appropriate en-US files/sections to en-CA, en-GB, en-AU, and en-IN.

Linting

This project uses Prettier and TSLint to help keep the codebase consistent. You can run all linting checks using npm run lint. Many of the more tedious errors can be fixed automatically; to do this, run npm run lint-fix.

Continuous Integration

This project includes a pipeline that runs a number of automated checks to ensure tests are passing, translation files are up-to-date, and that the code contains no linting errors. Currently, the pipeline is:

Deploying

To deploy, run npm run deploy. This command uses the Alexa Skills Kit (ASK) CLI to update the latest development version of this skill. Production deployments/submissions are a manual process.

A quick note

You may notice that this project's .gitlab-ci.yml file includes references to automated deployments (i.e., a job and stage named deploy). These jobs (and the corresponding deploy branch) are almost working, but ultimately were abandoned due to limitations with the ASK CLI tool. It's possible this automated deployment process will be reintroduced in the future if the ASK CLI makes this possible. For now, deployments are created by running the npm run deploy command locally as described above.

Contributing

Merge requests are welcome! 🙏 🙇

If you'd like to contribute code, see the Developing section above.

If you're interested in translating Lab Assistant into another language (thereby making it available in other Amazon stores), please open a merge request with translated versions of these files:

• ./Lab_Assistant/lambda/custom/i18n/en/translation.json
• ./Lab_Assistant/models/en-US.grammar and ./Lab_Assistant/models/en-US.base.json
  • This is used to generate ./Lab_Assistant/models/en-US.json
• ./Lab_Assistant/skill.json