feat(nifti): Add nifti volume loader to cornerstone 3D repo

[sedghi]

🎉 The work on this PR was made possible thanks to the generous funding from Mayo Clinic, underlining their dedication to fostering open-source community development.

👏 A hearty round of applause for James A. Petts (@JamesAPetts) and Davide Punzo (@Punzo), whose diligent work on the nifti loader proved invaluable.

Context

The nifti format is widely recognized for storing medical images and is the preferred choice for some research labs. This PR introduces a new package, @cornerstonejs/nifti-volume-loader, crafted to fetch a nifti file from a URL and load it as a volume within cornerstone3D.

We have included two new demonstrations:

• BasicNifti: Loads and renders a nifti file.
• NiftiWithTools: Loads a nifti file, enables tools, and displays a progress bar during the rendering process.

Given that the metadata for the nifti file is incorporated within the file header, and there is no standard procedure for servers to return the metadata, both metadata and pixelData will be fetched from the entire binary blob from the URL via an XHR request.

API

1. Prepend nifti: to the URL for the volume

const niftiURL = 'https://ohif-assets.s3.us-east-2.amazonaws.com/nifti/MRHead.nii.gz';
const volumeId = 'nifti:' + niftiURL;

2.1 create rendering engine and enable viewports like every other use case in cs3d

1. Create and Cache the volume, this will also load the volume as opposed to our dicom volume loader which needs a final .load (since in dicom we can fetch the metadata, create the volume, and stream data in)

  const volume = await volumeLoader.createAndCacheVolume(volumeId);

1. assign it to viewports

setVolumesForViewports(
    renderingEngine,
    [{ volumeId, callback: setCtTransferFunctionForVolumeActor }],
    viewportIds
  );

Changes & Results

Add nifti loader

Testing

Test the two examples with these data

1. https://ohif-assets.s3.us-east-2.amazonaws.com/nifti/CTACardio.nii
2. https://ohif-assets.s3.us-east-2.amazonaws.com/nifti/CTACardio.nii.gz
3. https://ohif-assets.s3.us-east-2.amazonaws.com/nifti/MRBrainTumor1.nii.gz
4. https://ohif-assets.s3.us-east-2.amazonaws.com/nifti/MRHead.nii.gz

Checklist

PR

• [] My Pull Request title is descriptive, accurate and follows the
  semantic-release format and guidelines.

Code

• [] My code has been well-documented (function documentation, inline comments,
  etc.)
• [] I have run the yarn build:update-api to update the API documentation, and have
  committed the changes to this PR. (Read more here https://www.cornerstonejs.org/docs/contribute/update-api)

Public Documentation Updates

• [] The documentation page has been updated as necessary for any public API
  additions or removals.

Tested Environment

• [] "OS:
• [] "Node version:
• [] "Browser:

[netlify]

✅ Deploy Preview for cornerstone-3d-docs canceled.

Name	Link
🔨 Latest commit	f294015
🔍 Latest deploy log	https://app.netlify.com/sites/cornerstone-3d-docs/deploys/64f9d1428542430008e4763e

[jbocce]

If cancel does not actually do anything, then why bother including it. If you keep it, consider removing the commented code.

[sedghi]

since it is part of the interface

[jbocce]

Perhaps remove the commented code then?

[jbocce]

Looks good so far. See my comments.

[wayfarer3130]

Do you really need to define yet another matrix format? How about affineMatrix as a function producing a gl-matrix mat4 instance instead, taking four arrays like this as the arguments, and producing the standard mat4 representation.

[sedghi]

As discussed in the meeting, I will do it later

[longuto]

Why not continue with this PR

[sedghi]

@longuto it is still under review (@wayfarer3130)

[wayfarer3130]

Is it necessary to have different events from the dicom volume loader? Really they should be the same events and just distinguish the type of data being loaded. It might be necessary to have them be slightly different because of bulk loading, but ideally making them identical would be useful.

[sedghi]

I think so... , since there is no streaming happening?

[wayfarer3130]

Could you write out LPS/RAS? At least once

[wayfarer3130]

This will throw a NPE if it isn't NIFTI data - probably throw an error if it isn't nifti data specifically, and otherwise continue handling it.

[wayfarer3130]

Things like this feel like they could be put into a function rather than making this function so long.

[wayfarer3130]

And a fair bit of this could also be put into functions that provide specific bits of data/usage, and can have their own documentation and maybe even their own unit tests.

[wayfarer3130]

You are modifying this in place - would it be safer to make a copy? What if the data is exposed as a Uint8array and the scaling no longer works? This just feels like it might cause problems, but the memory usage is definitely bigger if you copy the data instead.

[sedghi]

This is the same behaviour in the dicom image loader

[wayfarer3130]

Didn't we find that referencing a global tsconfig file has issues with linking?

[sedghi]

no that was something else that I fixed in webpack

[wayfarer3130]

There are a number of small comments/issues, nothing that HAS to be addressed, and I know this really should get in soon, so I'm ok with you reviewing the comments and deciding which ones to address.

[jbocce]

Looks good.