b3b9ff269a
## Description 🔖
This adds the ability to "customize & track" using the [new tracking ui](https://github.com/hasura/graphql-engine-mono/pull/5391).
A new button was added to implement this:
<img width="870" alt="Screen Shot 2022-09-16 at 12 37 14 PM" src="https://user-images.githubusercontent.com/49927862/190701948-1ad86717-f6be-4f67-8e0c-17b618790795.png">
## Solution and Design 🎨
This feature mostly makes use of components and hooks already created.
I was able to refactor some code to reduce code duplication and type duplication.
A few highlights on the refactor:
- `useTrackTable`, `useUntrackTable` and `useTrackSelectedTables` were all refactored into a single hook: `useTrackTable`. This hook has one main function but returns 4 wrapper functions: `trackTable`, `untrackTable`, `trackTables`, and `untrackTables`. This should make maintaining easier in the future.
- Synced up types between `MetadataTableConfig` and the customization form. Previously, the customization form had duplicated this same type, and there was some slight discrepancies between them.
- Modified `TableTrackingCustomizationModal` `onSubmit` return with a 2nd argument that's in the exact shape of `MetadataTableConfig` for convenience.
- Did some refactoring of the `DropDownMenu` component that should not interfere with anything. Exposed a few of the inner components for export and used these modular pieces in the current implementation. This ended up not being used in the feature, but left it in as it's a slight improvement.
I also added a function called `delayAsyncAction` [here](
|
||
|---|---|---|
| .. | ||
| src | ||
| .gitignore | ||
| .nvmrc | ||
| Dockerfile | ||
| package-lock.json | ||
| package.json | ||
| README.md | ||
| tsconfig.json | ||
Data Connector Agent Reference Implementation
This directory contains a barebones implementation of the Data Connector agent specification which fetches its data from static JSON files. It can be used as a reference implementation for testing, and as a reference for developers working on backend services.
Requirements
- NodeJS 16
Build & Run
> npm install
> npm start
Docker Build & Run
> docker build . -t dc-reference-agent:latest
> docker run -it --rm -p 8100:8100 dc-reference-agent:latest
Dataset
The dataset exposed by the reference agent is sourced from https://github.com/lerocha/chinook-database/
More specifically, the ChinookData.xml.gz file is a GZipped version of https://raw.githubusercontent.com/lerocha/chinook-database/ce27c48d9f375f81b7b68bacdfddf3c4458acc49/ChinookDatabase/DataSources/_Xml/ChinookData.xml
The schema-tables.json is manually derived from the schema of the data as can be seen from the CREATE TABLE etc DML statements in the various per-database-vendor SQL scripts that can be found in /ChinookDatabase/DataSources in that repo.
Configuration
The reference agent supports some configuration properties that can be set via the value property of configuration on a source in Hasura metadata. The configuration is passed to the agent on each request via the X-Hasura-DataConnector-Config header.
The configuration that the reference agent can take supports two properties:
tables: This is a list of table names that should be exposed by the agent. If omitted all Chinook dataset tables are exposed. If specified, it filters all available table names by the specified list.schema: If specified, this places all the tables within a schema of the specified name. For example, ifschemais set tomy_schema, all table names will be namespaced undermy_schema, for example["my_schema","Album"]. If not specified, then tables are not namespaced, for example["Album"].
Here's an example configuration that only exposes the Artist and Album tables, and namespaces them under my_schema:
{
"tables": ["Artist", "Album"],
"schema": "my_schema"
}
Here's an example configuration that exposes all tables, un-namespaced:
{}