Implement alignment service

[jonchang]

Description

Implement the alignment service. This can now be accessed through the ImageAlignment.align method and supports two backends (four-point alignment and image homography)

Screenshots (if applicable)

Related Issues

#226

Checklist

• The title of this PR is descriptive and concise.
• My changes follow the style guidelines of this project.
• I have added or updated test cases to cover my changes.
• I've let the team know about this PR by linking it in the review channel

[jonchang]

This will likely need to be run jointly with segmentation, as the segmentation templates currently have 'burned-in' color boxes and that needs to be excluded somehow prior to alignment

[jonchang]

Running alongside segmentation now, needs more aggressive filtering to pass tests as some of the test files are very small and don't have enough distinguishing features to properly do image homography

[jonchang]

two issues so far-

• needs to handle case where we are relying on small, already segmented images to test the ocr pipeline
• needs to handle the case where the test data already has segmentation areas 'burned in' to the image rather than getting passed in a reference plus coordinates for segment areas

[jonchang]

• reorganized the code, all backends are now in ocr.services.alignment.backends
• added tests that check both backends for alignment capability

[jonchang]

Thanks Andrew!