Skip to content

fractal-analytics-platform/fractal-plantseg-tasks

Repository files navigation

fractal-plantseg-tasks

Collection of Fractal task with the PlantSeg segmentation pipeline.

Tasks

  1. Plantseg Workflow: The main PlantSeg segmentation pipeline. For detailed information on PlantSeg, please refer to the PlantSeg repository.
  2. Tiff Converters: A Basic converter to convert a simple tiff file in OME-Zarr format.
  3. HDF5 Converters: A Basic converter to convert a simple hdf5 file in OME-Zarr format.

Installation and Deployment

  • Install the mamba package manager

  • Download the installation script from this repository

curl -O https://raw.githubusercontent.com/fractal-analytics-platform/fractal-plantseg-tasks/main/create_env_script.sh
  • The scrip might require some small modifications.
VERSION="v0.1.0" # Version of the package to install (by default the latest version)
COMMMAND="mamba" # Command to use to create the environment (mamba or conda) 
CUDA="12.1" # Available options: 12.1, 11.8 or CPU (default is 12.1)
# Location of the environment
# If ENVPREFIX is not NULL, the environment will be created with the prefix $ENVPREFIX/$ENVNAME 
# If ENVPREFIX is NULL, the environment will be created in the default location
ENVPREFIX="NULL" 
  • Install the package using the installation script
bash create_env_script.sh

The installation script will create a conda environment with the name fractal-plantseg-tasks and install the package in the environment. It will also download the correct __FRACTAL_MANIFEST__.json file.

  • In the fractal web interface add the task to the workflow as a "local env" task.
  • Plantseg will download the necessary models on the first run. The default location for the models and data is ~/.plantseg_models. If your system has a very limited home directory, you can set the environment variable PLANTSEG_HOME to a different location.

About

A collection of fractal tasks to run plantseg workflows

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published