NGSidekick

NGSidekick#

Tools for working with neuroglancer.

API Reference:

Feature Highlights#

  • Segment properties: read and write neuroglancer’s precomputed segment properties format from a pandas DataFrame.

  • Local annotations: construct local annotations directly in viewer state. (API subject to change.)

  • Precomputed annotations: export annotations in neuroglancer’s precomputed annotations format from a pandas DataFrame.

    • Supports all five annotation types:

      • point

      • line

      • axis_aligned_bounding_box

      • ellipsoid

      • polyline

    • Per-annotation properties (numeric, enum/categorical, and rgb/rgba color).

    • Per-annotation relationships (lists of related segment IDs, used by neuroglancer to filter annotations by segment).

    • Written to all “index” types (annotation id, related segment, and multi-level spatial grid)

    • Sharded output, written in parallel via tensorstore.

    • Note: writing directly to cloud storage is not yet supported; outputs must be written to a local filesystem (you must upload to cloud storage afterwards).

Installation#

Packages are available from both PyPI and conda-forge.

pip install ngsidekick

conda install -c conda-forge ngsidekick

For additional features:

pip install ngsidekick[gcs]  # For Google Cloud Storage support

conda install -c conda-forge ngsidekick google-cloud-storage
Contents