Update installation with optional dependencies

Author: MaximLippeveld

README.md

@@ -3,68 +3,3 @@
 ![main workflow badge](https://github.com/ScalableImagingPipeline/dask-pipeline/actions/workflows/main.yml/badge.svg) [![Documentation Status](https://readthedocs.org/projects/scalable-cytometry-image-processing/badge/?version=latest)](https://scalable-cytometry-image-processing.readthedocs.io/en/latest/?badge=latest) [![codecov](https://codecov.io/gh/ScalableCytometryImageProcessing/SCIP/branch/master/graph/badge.svg?token=6RYKJ8CLU0)](https://codecov.io/gh/ScalableCytometryImageProcessing/SCIP)
 
 Check the docs for installation and usage instructions.
-
-## Installation
-
-We recommend using [mamba](https://github.com/mamba-org/mamba) and pip to install SCIP.
-
-1. Create a new python environment:
- - Python 3.8: `mamba create -n scip python=3.8`
- - Python 3.9: `mamba create -n scip python=3.9`
-1. Activate environment: `conda activate scip`
-1. Install mpi4py: `mamba install -c conda-forge mpi4py`
-1. Install SCIP: `pip install .` or `pip install -e .` for development
-1. (Optional) Install development dependencies: `pip install -r requirements.txt`
-
-## Usage
-
-SCIP is called from the command line using the `scip` command. The help output is shown below.
-
-```
-Usage: scip [OPTIONS] OUTPUT CONFIG [PATHS]...
-
-  Intro documentation
-
-Options:
-  -d, --port INTEGER              dask dashboard port
-  --debug                         sets logging level to debug
-  --mode [local|jobqueue|mpi]     In which mode to run Dask
-  -j, --n-workers INTEGER         Number of workers in the LocalCluster or per
-                                  node
-  -n, --n-nodes INTEGER           Number of nodes started
-  -c, --n-cores INTEGER RANGE     Number of cores available per node in the
-                                  cluster  [x>=1]
-  -t, --n-threads INTEGER RANGE   Number of threads per worker process  [x>=1]
-  -m, --memory INTEGER RANGE      Amount of memory available per node in the
-                                  cluster  [x>=1]
-  -w, --walltime TEXT             Expected required walltime for the job to
-                                  finish
-  -p, --project TEXT              Project name for HPC cluster
-  -e, --job-extra TEXT            Extra arguments for job submission
-  --headless                      If set, the program will never ask for user
-                                  input
-  -s, --partition-size INTEGER RANGE
-                                  Set partition size  [x>=1]
-  --limit INTEGER                 Limit the number of events to load from each
-                                  file or directory
-  --timing FILE
-  --report / --no-report
-  --reach-limit                   If set the dataset will be repeated until
-                                  the limit is reached.
-  --gpu INTEGER RANGE             Specify the amount of available GPUs  [x>=0]
-  -l, --local-directory DIRECTORY
-  -V, --version                   Display version information
-  --help                          Show this message and exit.
-
-```
-
-### mode
-
-SCIP can run in three different modes: local, mpi or jobqueue.
- - In local mode, SCIP spins up a Dask `LocalCluster`. This mode can be used for execution on a laptop or desktop, for example.
- - In mpi mode, Dask workers and the scheduler are setup using the `dask-mpi` package. This is the ideal mode for use on high performance clusters. Note that for this mode SCIP should be run using `mpirun` or `mpiexec`.
- - Finally, in jobqueue mode, a PBSCluster is set up using the `dask-jobqueue` package. This mode can also be used for execution on a high performance cluster if MPI is not available, for instance.
-
-### OUTPUT, CONFIG and PATHS
-
-`OUTPUT` should be passed as a path to a directory where SCIP can write its output. `CONFIG` should be passed as the path to a YAML-file containing the configuration for the execution of SCIP. The files [scip_czi.yml](scip_czi.yml), [scip_multi_tiff.yml](scip_multi_tiff.yml) and [scip_tiff.yml](scip_tiff.yml) contain example configurations. `PATHS` should point to one or more images or paths containing images.

README.md.template

@@ -4,17 +4,21 @@ SCIP: Scalable Cytometry Image Processing
 Installation
 ------------
 
-SCIP requires an MPI implementation to be present on the path. If no implementation is present,
-mpi4py can be installed in a conda environment.
+SCIP can be installed in three ways:
+- with default functionality,
+- with additional MPI support for running SCIP in MPI mode,
+- and with additional CZI support for loading CZI files.
 
-We recommend using `mamba <https://github.com/mamba-org/mamba>`_ as a drop-in replacement for conda.
+To install SCIP:
 
-1. Create a new python environment: ``mamba create -n scip python=3.{8,9}`` (choose a version)
-2. Activate environment: ``conda activate scip``
-3. (Optional) Install mpi4py: ``mamba install -c conda-forge mpi4py``
-4. Download SCIP from Github
-5. Install SCIP: ``pip install .`` or ``pip install -e .`` for development
-6. (Optional) Install development dependencies: ``pip install -r requirements.txt``
+1. Download the latest release from Github, or clone the repository.
+2. Enter the repository directory.
+3. Run ``pip install .`` if you only need default functionality. Run ``pip install .[czi]`` for CZI support, and ``pip install .[mpi]`` for MPI support.
+4.  (Optional) Install development dependencies: ``pip install -r requirements.txt``.
+
+MPI support requires an MPI implementation to be present on the path. Please refer
+to instructions for your operating system to install MPI. Another option is to install, mpi4py
+in a conda environment.
 
 Usage
 -----

docs/source/index.rst

@@ -23,20 +23,19 @@
         'graphviz',
         'pyarrow',
         'umap-learn',
-        'dask_mpi',
-        'aicsimageio',
-        'aicspylibczi',
-        'mpi4py',
         'bokeh',
         'zarr',
-        'fastparquet',
-        'cellpose'
+        'fastparquet'
     ],
     entry_points={
         'console_scripts': [
             'scip=scip.main:cli'
         ]
     },
+    extras_require={
+        "mpi": ['dask_mpi', 'mpi4py'],
+        "czi": ['cellpose', 'aicsimageio', 'aicspylibczi']
+    },
     python_requires=">=3.8",
     package_data={
         'scip': ['utils/logging.yml']

setup.py

@@ -17,7 +17,6 @@
 
 from scip.loading import multiframe_tiff
 from scip.masking import threshold, util
-from functools import partial
 
 
 def test_bounding_box(images_folder, cluster):