BlueBrainAtlas WebExporter

The web exporter is a suite of tools to facilitate the conversion and export of the Blue Brain Atlas pipeline to web-friendly datasets to be consumed by the Atlas web app.

Once installed, this package exposes two CLIs: - cellexport - parcellationexport

Installation

pip install "blue-brain-atlas-web-exporter"

This will automatically install the CLIs listed below and make them available in your PATH.

Examples

cellexport

The CLI cellexport converts a cell-position SONATA file into a set of RandomAccessBuffer files (RAB) as well as a JSON metadata file containing a few computed metrics.
Here is the list of arguments: - --help displays the help menu - --version displays the version number - --hierarchy <FILE PATH> input path to the brain region hierarchy file (commonly called 1.json) — mandatory - --sonata <FILE PATH> input path to the SONATA file containing cell positions/orientation/brain-regions/celltypes — mandatory - --downsampling-factor <NUMBER> factor to downsample the data — optional (default: 1, no downsampling. Example: a factor of 100 will keep only one value every 100) - --out-postion <FILE PATH> output path to the RAB file that contains the cell positions — mandatory - --out-cell-type <FILE PATH> output path to the RAB file that contains the cell types — mandatory - --out-orientation <FILE PATH> output path to the RAB file that contains the cell orientations — mandatory - --out-region <FILE PATH> output path to the RAB file that contains the cell region IDs — mandatory - --out-metadata <FILE PATH> output path the JSON file that contains the metadata — mandatory

parcellationexport

The CLI parcellationexport computes all the meshes (.OBJ files) as well as the binary masks (.nrrd files) for all the regions available in both the hierarchy file (input) and the annotation volume (input). Many intermediate regions from the hierarchy file do not have any voxel representation and instead are composed of voxels from their child region, thus, this CLI computes all the aggregations necessary to recreate binary masks for those regions. This CLI also output a JSON metadata file with metrics regarding the volume and volume ration for all the regions.

For each region, this CLI also generates a list of neighbour regions (adjacentTo) and the list of neighbour region that are on the same layer, when applicable (continuousWith). Note that continuousWith is a subset of adjacentTo and while adjacentTo entirely relies on volumetric data, continuousWith leverages the region hierarchy and in particular the regions' acronyms. Those metadata are part of the main JSON payload, alongside with the region volumes.

Here is the list of arguments:
- --help displays the help menu - --version displays the version number - --hierarchy <FILE PATH> input path to the brain region hierarchy file (commonly called 1.json) — mandatory - --parcellation-volume <FILE PATH> input path to the parcellation file format (.nrrd) — mandatory - --out-mesh-dir <DIR PATH> output path for the directory where the region/parcellation mesh files (.obj) will be written. Each file will have the name corresponding to the ID of the region (example: 997.obj for the whole brain) — mandatory - --out-mask-dir <DIR PATH> output path for the directory where the region/parcellation binary mask volumes (.nrrd). Each file will have the name corresponding to the ID of the region (example: 997.nrrd for the whole brain) — mandatory - --out-metadata <FILE PATH> output path to the JSON file that contains the metadata - mandatory

Funding & Acknowledgment

The development of this software was supported by funding to the Blue Brain Project, a research center of the École polytechnique fédérale de Lausanne (EPFL), from the Swiss government's ETH Board of the Swiss Federal Institutes of Technology.

Copyright © 2015-2022 Blue Brain Project/EPFL