StackGenVis: Alignment of Data, Algorithms, and Models for Stacking Ensemble Learning Using Performance Metrics https://doi.org/10.1109/TVCG.2020.3030352
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
StackGenVis/frontend/node_modules/parcoord-es/README.md

161 lines
4.3 KiB

4 years ago
# parcoords-es
ES6 module of Syntagmatic's [Parallel Coordinates](https://github.com/syntagmatic/parallel-coordinates) (aka. parcoords). This library is completely based on D3 V5 API.
## Features
Please refer to [Parallel Coordinates](https://github.com/syntagmatic/parallel-coordinates)'s project page for concepts and API usage
All examples of the original project has been verified. You can play with them via running:
```
npm install
npm run dev
```
## API
This section only lists api that are <b>deviated</b> from the original parallel coordinates.
<a name="parcoords_brush_extents" href="#parcoords_brush_extents">#</a> parcoords.<b>brushExtents</b>(extents) supports 1D multi brushes:
1D brush [<>](https://github.com/BigFatDog/parcoords-es/blob/master/demo/setterForBrushes.html "Source")
```javascript
.brushMode("1D-axes")
.brushExtents({"2": [3,4]});
```
1D multi brush [<>](https://github.com/BigFatDog/parcoords-es/blob/master/demo/setterForMultiBrushes.html "Source")
```javascript
.brushMode("1D-axes-multi")
.brushExtents({"2": [[3,4], [6,8], [10, 14]]});
```
when parameter `extents` is not provided, this function returns
```
const extents = parcoords.brushExtents();
// format is:
{
extents,
selection: {
raw, //raw coordinate
scaled //y-scale transformed
}
}
```
<a name="parcoords_marking" href="#parcoords_marking">#</a> parcoords.<b>mark</b>([values])
allows user to permanently highlight a data element in chart
```
const data = [...];
const pc = ParCoords().data(data)...;
parcoords.mark(data.filter(d => d.volume > 10));
```
Please refer to [marking demo](https://github.com/BigFatDog/parcoords-es/blob/develop/demo/marking.html "Source") for details
<a name="parcoords_unmark" href="#parcoords_unmark">#</a> parcoords.<b>unmark</b>()
clears all permanently highlighted data that is added by <a href="#parcoords_marking">mark([values])</a>
<a name="parcoords_brush_arg" href="parcoords_brush_arg">#</a> parccords.<b>on</b>(function(brushed, args){})
adds brush arguments to `brushstart`, `brush`, and `brushend` events.
```
parcoords.on('brushstart', function(brushed, args){
const {
selection: {
raw, //raw coordinate
scaled //y-scale transformed
},
node, // svg node
axis // dimension name
} = args;
})
```
Please refer to [brushing with arguments demo](https://github.com/BigFatDog/parcoords-es/blob/develop/demo/brush-with-arguments.html "Source") for details
## Usage
### ES6
1. Install library in your project
```
npm install parcoord-es --save
```
2. import module
```
import 'parcoord-es/parcoords.css';
import ParCoords from 'parcoord-es';
const chart = ParCoords()....
```
### Standalone
parcoords.standalone.js contains all dependencies and can be used directly in your html page. Please note that only essential D3 V5 modules are bundled, your global namespace won't be polluted.
```
<link rel="stylesheet" type="text/css" href="./parcoords.css">
<script src="./parcoords.standalone.js"></script>
var parcoords = ParCoords()("#example")
```
You are free to use either D3 V3 or D3 V5 in your html. demo/superformula.html demonstrates how to use parcoords-es with d3 V3.
## Development
Follow this instruction to setup development environment for parcoords-es
### Prerequisites
npm
### Installing
```
npm install
```
### Building
```
npm run build
```
### Development
Internal server will be launched, hosting all demos at localhost:3004
```
npm run dev
```
### Testing (Coverage)
run all unit tests and generate test coverage report.
```
npm run test:cover
```
## Built With
* [D3 V5](https://d3js.org/) - D3 modules are used
* [Rollup](https://github.com/rollup/rollup) - Module bundler
## License
This project is licensed under the MIT License - see the [LICENSE.md](LICENSE.md) file for details
## Contributors
This project gets supports from open-source community. Many thanks to our [contributors](https://github.com/BigFatDog/parcoords-es/graphs/contributors)
## Contribution Notes
Run `npm run pretty` before committing.
## Acknowledgments
This project is based on [Parallel Coordinates](https://github.com/syntagmatic/parallel-coordinates) v0.7.0. Many thanks to [parcoords contributors](https://github.com/syntagmatic/parallel-coordinates/graphs/contributors) for such a complicated and useful D3 visualization.