Jongleur

jongleur.app

Your learning. As a musician. Simplified.

About

Jongleur helps musicians practice and track their hours by ingesting recordings and returning useful information and visuals. Currently we support Scarlatti solo sonatas.

Development

Initial setup

Most of the code is written in Typescript, so the setup is simply

$ npm i

A good deal of the workflow is supported by vscode, so install the recommended vscode extensions. The settings should already configure eslint to run on save without any further modification.

If working on any Python files locally (a few files), you'll need ffmpeg and libsndfile. On MacOS, for example:

$ brew update
$ brew install ffmpeg libsndfile

(You may need some other libraries depending on your exact setup; if so, install them and feel free to add them to this README!). You'll also want pipenv (pip3 install pipenv) so you can use the virtual environment:

$ pipenv install
$ pipenv shell

Running the development server

Use the command

$ npm run dev

This will run the Remix web app and watch the graphql schema to regenerate Typescript types upon modification. It will also print out the Jongleur "wordmark," just for fun. Try it!

Right now, we don't have emulation for the backend when developing locally. You just run the web app locally but connect to the remote server. This is generally fine, but it requires having a working server running (see deployment below). Right now you can connect to the version running at jongleur.app but once the app is out of initial development this will change.

Deployment

While it's not necessary for front-end development, you might want to deploy a stack of your very own.

The command

$ npm run deploy

deploys the entire application to AWS (the front end, the backend, the GraphQL API, and more, all together). This command handles

1. Boxing up the web app into a docker container and pushing it to ECR
2. Starting the ECR image through ECS (Fargate)
3. Deploying the GraphQL API and its data resolvers
4. Creating the data stores themselves (databases, buckets, etc.)
5. Creating the Cognito user pool
6. Launching all the misc lambdas and rest APIs
7. Etc

But it deploys it by default to my AWS account and to my domain (jongleur.app). If you want to deploy the app yourself, you should be able to do it just by changing these values in infrastructure/bin/jongleur-infrastructure.ts.

Then you simply add the validation records to your DNS during deployment for the HTTPS certificate (see the ACM panel in your AWS console) and the app should be good to go. If you're asked to bootstrap your CDK environment when attempting to deploy, do it.

Initial data

One thing you'll notice after deployment is that there is no initial data in the application's data stores. While you probably want the app to be deployed without any user accounts or user data, you might want to tell the app to support some pieces so you can play.

No problem :) We've got a script and data file that will upload about 500 Scarlatti sonatas. You just point the script at the data file and give it the name of the Piece table that was deployed. (We use comma of course!)

For example, I can run the command like this:

$ python infrastructure/populate_piece_table.py asset/scarlatti_piano_solo_sonatas.csv "JongleurInfrastructureStack-JongleurDataConstructPieceTableECC218AA-KYBQW0X5HSTU"

(If you get a syntax error, did you forget to run pipenv shell?).

Because this command is being run locally, just make sure that your aws config is pointed at an IAM user with permissions on your account. Of course, it isn't pretty that you have to find the name of the Piece table yourself, but then again, the point of this script is just to bootstrap your app deployment, it's definitely not meant to be used regularly. A remote function that adds pieces will be added to our GraphQL resolver

Staging

Following what I curently consider to be best practice, if you want to have {prod, stage, qa, etc.} environments, you should deploy the identical stack to separate accounts. It should be simple enough to set up a couple different deployment commands to choose the appropriate account, but that hasn't been done since this app has no active user-base yet! But soon!

Documentation

Docs are kept in documentation/. To build the report from latex, I've set up my Mac as follows:

1. Install MacTex
2. Run sudo tlmgr update --self and then sudo tlmgr install latexmk

This will make the Late Workshop VSCode extension work.