JaneliaSciComp/cellpose
modulePerforms a 3-D segmentation using cellpose distributed on a dask cluster
dask segmentation cellpose
Module Information
Tools
Inputs
| Name | Type | Description |
|---|---|---|
| meta | map | Groovy map containing: id, input_image_dataset, output_image_name |
| image | file | Path to the image container |
| image_subpath | string | Relative path to the image dataset if needed (for N5 or zarr containers) |
| models_path | directory | Optional path to the cellpose models cache directory. If this is not specified pass in as empty list. |
| model_name | string | The name of the cellpose model. |
| output_dir | directory | Directory location of the output image container |
| labels | string | Name of the labels container created at the location specified by output_dir |
| labels_subpath | string | Relative labels dataset subpath |
| working_dir | directory | Location of the temporary files created during the segmentation. This must have enough space to hold the temporary labels before they are reassigned during the merge process |
| dask_scheduler | string | Dask scheduler IP |
| dask_config | file | Optional path to the dask config. If this is not specified pass in as empty list. |
| preprocessing_config | file | Optional YAML file holding the configuration parameters for various preprocessing operations. So far the only preprocessing available is a gaussian filter but there may be more in the future |
| logging_config | file | Optional log config. |
| cellpose_cpus | integer | Number of cpus needed to run the main cellpose program |
| cellpose_mem_in_gb | string | Memory needed for the main cellpose program |
Outputs
| Name | Type | Description |
|---|---|---|
| meta | map | Groovy map containing: id, input_image_dataset, output_image_name |
| input_image_fullpath | directory | Full path to the input image container |
| image_subpath | string | Subpath to the input image dataset |
| output_label_images | string | One or more image names containing segmentation results. If the output is a zarr or N5 container there's only one output_label_image, but if the output type is a tiff and the segmented image is large there can be multiple tiff files |
| output_labels_subpath | string | this is the labels dataset if the output is a zarr or N5 container |
| versions | file | File containing software versions |
Quick Start
Include this module in your Nextflow pipeline:
include { CELLPOSE } from 'https://github.com/JaneliaSciComp/nextflow-modules/tree/main/modules/janelia/cellpose'