DMP2 (formerly discodomeuprimo)

Source code for https://discodomeuprimo.lol, a comprehensive and heavily biased catalogue of artists and bands worth knowing about.

Tech stack

• Nuxt 3
• Typescript
• Pinia
• Musicbrainz API
• Firebase / Firestore
• Netlify

Commands

Run project

yarn dev

Runs the project locally (defaults to port 3000).

Build

bash build.sh

Reads local files based on the path provided in .env and builds a JSON file with the directory tree (by using tree).

The JSON dump is stored in dumps/[timestamp].json.

Sync

yarn sync

1. Reads the JSON file previously created with build.sh
2. Calls Musicbrainz's API so that we can get the unique ID of the artist from their database (in case of multiple findings, assumes that the one with highest score is the one we want)
3. Builds each Artist with all the information
4. Synchronizes the Firestore database with the new artist (overwrites it if the artist already exists)

Scan slicing

# Start scanning from the letter n
yarn sync --from=n

# End scanning in the letter n
yarn sync --to=n

# Example
yarn sync --from=sa --to=sat

# Output - stops BEFORE Satyr:
# * Processed Sabaton
# * Processed Salvador Sobral
# * Processed Sam Amidon
# * Processed Sam Cooke
# * Processed Samuel Úria
# * Processed Santana

Dry run

Doesn't insert anything in the database.

yarn --dry

How does it work?

The local catalogue

The files stored in the BASE_PATH you have set in .env are read with bash build.sh, stored in a JSON file, and then, by running yarn sync, each of the artists is stored in the database (Firestore), along with its catalogue.

In order for this to work, a strict structure must be met:

/[your BASE_PATH]
/[your BASE_PATH]/A Perfect Circle
/[your BASE_PATH]/A Perfect Circle/Albums
/[your BASE_PATH]/A Perfect Circle/Albums/2000 - Mer de Noms
/[your BASE_PATH]/A Perfect Circle/Boxset & Compilations/
/[your BASE_PATH]/A Perfect Circle/Boxset & Compilations/2004 - aMOTION (rEMIXED)

The official catalogue

We read the official catalogue in real-time from Musicbrainz - it isn't stored in our database, since that would prevent us to have it consistently up-to-date.

We are able to do this because yarn sync attempts to find each artist (folder name) in Musicbrainz API and we store its unique Musicbrainz ID in our database - thus allowing us to have a link between our local catalogue and the "remote" one.

To do

API

• Refactor the direct approach to also use musicbrainz-api (better type support, throttling)

• Save artist description and external links in our database, since we are making a request when running yarn sync anyway and this information will unlikely change too often

Catalogue management

Local catalogue

• Normalize the whole catalogue so that EVERY album follows the [year] - [name] folder convention

UI

List page

• Lowercase artist names should appear in alphabetical order as if they were uppercase (not at the end of the list)

Catalogue page

• Include artist description, external links

• Allow togging local / official panels

• Download album covers from coverartarchive.org

Misc

• Decent loading states

Infrastructure

• Move the conversion from Listing to object map to a centralized location (code duplication in nuxt.config.ts and index.vue)

• Organize types and interfaces better