vsvi2precomputed

Package for converting VSVI (used in VAST) image datasets to precomputed volumes. Supports conversion of local and AWS S3 datasets.

Requirements:

• Python
• AWS CLI (if using S3)

Usage

Convert a cloud dataset and store in new cloud path:

pip install -r requirements.txt
python vsvi2precomputed.py -i s3://path/to/config.vsvi -o s3://path/to/output/dir/

Don't forget the trailing slash on the output dir.

Convert a local dataset and upload to the cloud:

python vsvi2precomputed.py --i path/to/config.vsvi --o s3://path/to/output/dir/

Convert a cloud dataset and upload to the cloud:

python vsvi2precomputed.py --i s3://path/to/config.vsvi --o path/to/output/dir/

Convert a dataset locally:

python vsvi2precomputed.py --i path/to/config.vsvi --o path/to/output/dir/

Optional Arguments

Argument	Description	Default
--profile	AWS CLI profile name	default

Tests

pip install pytest
pytest

To use an non-default AWS CLI profile:

pytest --profile <profile-name>

About VSVI and precomputed formats

VSVI format is native to the VAST ecosystem. Precomputed format is native to the Neuroglancer/CloudVolume ecosystem.

To view converted data in Neuroglancer:

• Navigate to neuroglancer.bossdb.io.
• Add a new layer using the Data Source URL input box on the top right.
  • S3: The Data Source URL will be the S3 URI of the directory containing the info file, prepended with precomputed://. Example: precomputed://s3://mambo-datalake/connects49a/vsvi2precomputed/local_aligned/.
  • Local: You will need to serve the data first. Navigate to the directory containing the info file, then open a terminal and run the following code. The Data Source URL will then follow the format precomputed://localhost:<port>/.

  from cloudvolume import CloudVolume
  cv = CloudVolume("file://.")
  cv.viewer()
  

  • Click the yellow "Create as image layer" button at the bottom right.

Acknowledgements

We thank the Visual Computing Group at Harvard for building the VAST software. https://www.frontiersin.org/journals/neural-circuits/articles/10.3389/fncir.2018.00088/full

---

Copyright (c) 2024 The Johns Hopkins University Applied Physics Laboratory LLC.