Execute a darfix workflow without GUI

import os
from orangecontrib.darfix import tutorials

from CLI using ewoks

You can execute darfix workflows as any ewoks workflow using the ewoks execute command. Parameter values can be provided with the --parameter options (or the -p alias). For example, we want to execute a workflow named my_darfix_workflow.ows by setting two parameters: the detector HDF5 dataset and the HDF5 positioner dataset (aka metadata). For this, we can use the following command:

ewoks execute /home/esrf/payno/Documents/my_darfix_workflow.ows --parameter filenames=silx:///data/scisoft/darfix/datasets/bliss_hdf5/Silicon_111_reflection_0003/Silicon_111_reflection_0003.h5?path=/1.1/instrument/pco_ff/data --parameter metadata_url=silx:///data/scisoft/darfix/datasets/bliss_hdf5/Silicon_111_reflection_0003/Silicon_111_reflection_0003.h5?path=/1.1/instrument/positioners

Warning: the HDF5 dataset must be provided as silx URL

From python

step 1: Define task arguments which are not defined in the workflow

in this example we will focus about HDF5 dataset.

root_dir = "/tmp/darfix"
os.makedirs(root_dir, exist_ok=True)
hdf5_file = os.path.join(tutorials.__path__[0], "hdf5_dataset", "strain.hdf5")
assert os.path.exists(hdf5_file)

step2: Execute the workflow

For more information on executing ewoks workflows: https://workflow.gitlab-pages.esrf.fr/ewoks/ewoks/

from ewoks import execute_graph
graph = os.path.join(tutorials.__path__[0], "darfix_example_hdf.ows")
results = execute_graph(
    graph,
    inputs=[
        {"name": "raw_input_file", "value": hdf5_file,},
        {"name": "raw_detector_data_path", "value": "/1.1/instrument/my_detector/data",},
        {"name": "raw_positioners_group_path", "value": "/1.1/instrument/positioners"},
        {"name": "in_memory", "value": False,},
    ],
    output_tasks=True,
)

WARNING:/usr/local/lib/python3.9/site-packages/darfix/tasks/roi.py:Cannot apply a ROI if origin ([]) or size ([]) is empty. Dataset is unchanged.

Computing moments |████████████████████████████████████████████████████████████████████████████████████████████████████| 100.0%

Inspect the results

for node_id, task in results.items():
    assert task.succeeded, f"task {node_id} failed"
    print(task.get_output_values())

{'dataset': Dataset(dataset=<darfix.core.dataset.ImageDataset object at 0x7fc3bc09c670>, indices=None, bg_indices=None, bg_dataset=None)}
{}
{'dataset': Dataset(dataset=<darfix.core.dataset.ImageDataset object at 0x7fc3bc09c670>, indices=None, bg_indices=None, bg_dataset=None)}
{'dataset': Dataset(dataset=<darfix.core.dataset.ImageDataset object at 0x7fc3bc09c670>, indices=None, bg_indices=None, bg_dataset=None)}
{'dataset': Dataset(dataset=<darfix.core.dataset.ImageDataset object at 0x7fc3bc09c670>, indices=None, bg_indices=None, bg_dataset=None)}
{'dataset': Dataset(dataset=<darfix.core.dataset.ImageDataset object at 0x7fc3bc09c670>, indices=None, bg_indices=None, bg_dataset=None)}
{'zsum': array([[178., 214., 218., ..., 217., 208., 221.],
       [215., 219., 212., ..., 227., 225., 222.],
       [377., 223., 221., ..., 217., 217., 211.],
       ...,
       [167., 188., 197., ..., 191., 199., 190.],
       [185., 184., 181., ..., 178., 183., 191.],
       [176., 198., 203., ..., 181., 195., 187.]])}
{'dataset': Dataset(dataset=<darfix.core.dataset.ImageDataset object at 0x7fc3bc09c670>, indices=None, bg_indices=None, bg_dataset=None)}

Note

The same can be done for EDF datasets. In this case the, metadata_url parameter should be omitted and the filenames parameter will contain the list of the EDF image paths:

from glob import glob
root_dir = "/tmp/darfix"
os.makedirs(root_dir, exist_ok=True)
filenames = glob(os.path.join(tutorials.__path__[0], "*.edf"))