open-dicom / csa_header Goto Github PK

pip update is suddenly raising an exception in #11.

Worth implementing a data store with Pooch or similar? For example, on lines of : https://github.com/nipraxis/nipraxis/blob/main/nipraxis/_fetcher.py

Then you could do something like this in the README:

import csa_header
# Fetch example DICOM file from data repository.
eg_fname = csa_header.data.fetch_file('an_example.dcm')
dcm = pydicom.dcmread(eg_fname)

or the tests:

def test_something():
    eg_fname = csa_header.data.fetch_file('an_example.dcm')
    eg_dcm = pydicom.dcmread(eg_fname)
    raw_csa = eg_dcm.get((0x29, 0x1020)).value
    parsed_csa = CsaHeader(raw_csa).read()
    assert (parsed_csa['NumberOfPrescans' == 
        {'index': 1, 'VR': 'IS', 'VM': 1, 'value': 0})

Issue a Zenodo DOI after a first release and add to the README.

Am I right in thinking that the index elements in a CSA header are always sequential?

For example, from the README:

{
    'NumberOfPrescans': {'index': 1, 'VR': 'IS', 'VM': 1, 'value': 0},
    'TransmitterCalibration': {'index': 2, 'VR': 'DS', 'VM': 1, 'value': 247.102},
    'PhaseGradientAmplitude': {'index': 3, 'VR': 'DS', 'VM': 1, 'value': 0.0},
    'ReadoutGradientAmplitude': {'index': 4, 'VR': 'DS', 'VM': 1, 'value': 0.0},
    'SelectionGradientAmplitude': {'index': 5, 'VR': 'DS', 'VM': 1, 'value': 0.0},
...
}

Are these index entries necessary? I realize they are in the CSA header structure, but, given Python dictionary insertion order, does:

indices = [v['index'] for k, v in parsed_csa.items()]

ever differ from np.arange(len(parsed_csa))?

I'd like to include csa_header in my package (to parse CSA headers, obviously) but cannot do so because there is no pypi release. The alternative is to install (it's parent?) dicom_parser, but that has more dependencies and seems to be behind in terms of commits to the csa lib. What's the status of this project, is it a fork of dicom_parser? Is it still actively developed, i.e. is there a pypi release planned?

Nibabel has a get_csa_header function : https://github.com/nipy/nibabel/blob/master/nibabel/nicom/csareader.py#L31

It implements the logic of finding the image or series CSA header from the DICOM file. I see from that file that there is logic for finding where the relevant private elements start: https://github.com/nipy/nibabel/blob/master/nibabel/nicom/utils.py#L7. . So, for example, instead of doing this:

raw_csa = dcm.get((0x29, 0x1020)).value
CsaHeader(raw_csa).read()

you can do this:

get_csa_header(dcm, 'image')

Following the suggestion made by @matthew-brett in #11, it would be good to integrate doctests.

Following #16, pooch should be integrated to support fetching test files from dicom-gallery repositories.

Where should this repo sit in relationship to Nibabel or pydicom? Should it be an optional dependency? Won't that mean that lots of DICOM libraries could have this as an optional dependency?

Currently, it is inconsistent between 3.7+ and 3.9+.

Add CONTRIBUTING and CODE_OF_CONDUCT files.

Test coverage should be greater than 90%.

It might be convenient to move NiBabel's summary of CSA headers, as well as any other useful information, to this repository's Wiki tab (I think a complete Sphinx setup might be a bit of an overkill, happy to discuss).

• Get hatch run lint:typing to run successfully.
• Update CI with an appropriate check to make sure it continues to run successfully.