Composite

Introduction

The ParaView ArcticViewer is able to load several type of datasets, but this guide will focus on the basic composite one and will explain the requirements so you can create your own dataset.

Dataset structure

ParaView ArcticViewer expects a dataset descriptor that will formalize that convention in a way it can be understood by the application. The application expects a file named index.json at the root of the tree structure (if any) with a content similar to the following one.

{
    "type": [ "tonic-query-data-model", "composite-pipeline" ],
    "arguments_order": ["time"],
    "arguments": {
        "time": {
            "values": [ "0", "1", "2", "3", "4", "5" ]
        }
    },
    "data": [
        {
            "name": "sprite",
            "type": "blob",
            "mimeType": "image/jpg",
            "pattern": "{time}/rgb.jpg"
        },{
            "name": "composite",
            "type": "json",
            "pattern": "{time}/composite.json"
        },
    ],
    "CompositePipeline": {
        "layers": ["A", "B", "C"],
        "dimensions": [ 500, 500 ],
        "fields": {
            "A": "salinity",
            "B": "temperature",
            "C": "bottomDepth"
        },
        "layer_fields": {
            "A": [ "C" ],
            "B": [ "B", "A" ],
            "C": [ "B", "A" ]
        },
        "offset": {
            "AC": 1,
            "BA": 3,
            "BB": 2,
            "CA": 5,
            "CB": 4
        },
        "pipeline": [
            {
                "ids": [ "A" ],
                "name": "Earth core",
            },{
                "ids": [ "B", "C" ],
                "name": "Contour by temperature",
                "children": [
                    {
                        "ids": [ "B" ],
                        "name": "t=25.0",
                        "type": "layer"
                    },{
                        "ids": [ "C" ],
                        "name": "t=20.0",
                        "type": "layer"
                    }
                ],
            }
        ]
    }
}

In that meta description, we find the pipeline tree described but also we have the mapping of the image sprite with the layer and its “ColorBy”. The first character is the layer and the second one is the actual field used for the ColorBy.

Then for each query (view position, time, configuration, …), the dataset is composed of an image sprite and a composite.json file.

The image sprite represents each layer independently with each of its ColorBy settings.

The composite.json file, on the other hand, provides the pixel ordering for each of the possible layers with a content similar to the following one:

{
    "pixel-order": "@2258+A+CBA+CA+CB+..."
}

The pixel-order string describes the layer order for each pixel, starting from the top left corner of the image.

The encoding can be described as follows:

• @456 : Skip 456 pixels as no layers are contributing to those pixels.
• CBA: For the given pixel, the layer _C_ is on top of _B_ which is on top of _A_. This implies if _C_ is hidden, for example, then the pixel of layer _B_ should be used.
• + : Character delimiter between pixel or pixel group.

Image Sprite