@aglyn/plugins-data

The Data plugin for Aglyn: datasets, their records, and CSV or JSON import and export, managed in the Aglyn console. Install it if you are running or building on the Aglyn platform and want datasets.

Published to npm from the Aglyn monorepo. Versions follow semantic versioning.

Install

npm install @aglyn/plugins-data

Peer dependencies:

• @mui/material
• firebase
• react

None is optional.

What's in it

This package is a first-party Aglyn plugin. It is loaded through Aglyn's plugin manager: the apps read its entry in the monorepo's plugins.config.json, import the module each surface names, and call the registrar declared for that surface. It is not a standalone library; installing it on its own loads nothing.

The plugin is console-only and declares one registrar, registerDataConsole (the console surface). Datasets are organization-shared collections that repeatable components read at render time through bindings; they are not a canvas component, so there is no site registrar and no server entry point in this package.

registerDataConsole() registers:

• A Data nav item and page at /data, behind the dataStore feature flag. The page is code-split.
• The organization datasets card, HostDatasetsCard, as a widget in the orgData slot.
• Datasets as a repeat source (DATASET_REPEAT_SOURCE), which is what offers "Repeat over dataset" on an element in Besigner and draws a repeat's copies from real rows on the canvas.

It also keeps the datasetSchema installable artifact type (artifactTypes in plugins.config.json, registered from registerDataConsoleServerDeclarations): a dataset's schema published to a marketplace, installed as a new, empty dataset, and updated in place. The installer asks this plugin through @aglyn/aglyn/plugin-manager/plugin-artifact-types and never reads the datasets collection; the schema's model (sanitizeDatasetSchema, resolveInstalledDatasetSchema, summarizeSchemaChange) is @aglyn/plugins-data/artifact/dataset-schema.

On the server, registerDataServerDeclarations() also registers the two automation steps that write a dataset record, datasetAppend and updateDataset (server/dataset-steps.server.ts). The Workflows engine runs them through the platform's server-step seam (registerServerStepExecutor), and plugins.config.json declares them under serverSteps. A record a step writes refreshes the live pages that repeat over its dataset (server/dataset-live-pages.ts).

Entry points

Usage

The registrar is normally called by Aglyn's generated console loader. Called directly:

import { registerDataConsole } from '@aglyn/plugins-data'
registerDataConsole()

Import and export run on the platform's transfer framework (AGL-3530). The plugin declares data.dataset under transferResources with instances, so each dataset is its own resource, data.dataset:<datasetId>; it registers the server half from registerDataConsoleServerDeclarations (transfer/dataset-transfer.server.ts: the field catalog, match keys, the export's pages and count, the dry run held to the dataset's model, the writes and their undo) and the client half from registerDataConsole. The Data card opens the wizard and the export dialog through the console's launcher:

import { useTransferLauncher } from '@aglyn/aglyn'
import { datasetTransferResourceKey } from '@aglyn/plugins-data'

useTransferLauncher()?.openImport({ resource: datasetTransferResourceKey(datasetId), scope: 'org' })

How it fits

A plugin may import the tenant runtime, the renderer, the Besigner logic, the core and the shared packages. It never imports another plugin, and the core never imports a plugin. Data depends on @aglyn/aglyn (which owns the dataset model and the CSV helpers), @aglyn/tenant-runtime, @aglyn/tenant-feature-instance and a few @aglyn/shared-* packages. Other plugins that need CSV parsing take it from the core, not from this package.

License

Apache-2.0. Source: https://github.com/aglyn/aglyn/tree/main/libs/plugins/data