nthistle/n5-label-multisets-spark

Spark Utilities for Label Multisets

Utilities for label multisets. Currently available:

• convert integer type n5 dataset into label multiset dataset
• Downsample label multiset n5 dataset

Compile

Currently, these SNAPSHOT dependencies are required:

• imglib2-algorithm parallelize-over-blocks branch

To compile for use on local machine, run

mvn -Denforcer.skip=true -PfatWithSpark clean package

To compile for deployment on the Janelia cluster, run

mvn -Denforcer.skip=true -Pfat clean package

Run

To run locally, run

java \
    -Dspark.master=local[*] \
    -jar target/n5-label-multisets-spark-0.1.0-SNAPSHOT-shaded.jar COMMAND [--help/-h] [COMMAND ARG...]

Convert

To run locally, run

java \
    -Dspark.master=local[*] \
    -jar target/n5-label-multisets-spark-0.1.0-SNAPSHOT-shaded.jar convert [--help/-h] [ARG...]

Downsample

To run locally, run

java \
    -Dspark.master=local[*] \
    -jar target/n5-label-multisets-spark-0.1.0-SNAPSHOT-shaded.jar downsample [--help/-h] [ARG...]