EmbeddedDocument field with a label type (ex: ground_truth of type Detections) should not have other typed fields of type primitives.

[manivoxel51]

What changes are proposed in this pull request?

An embedded document field with a label list type (ex: ground_truth of type Detections) should not have other typed fields (ground_truth.annotator for example) of type primitives.

Note this can be extended to more use cases but a good safeguard for now.

import fiftyone as fo

dataset = fo.load_datase("quickstart")
sample = dataset.first()
sample["ground_truth.annotator"] = "foo"
sample.save()

dataset.add_dynamic_sample_fields()
dataset.save()

// launch app

before fix

after fix

How is this patch tested? If it is not, please explain why.

(Details)

Release Notes

Is this a user-facing change that should be mentioned in the release notes?

• No. You can skip the rest of this section.
• Yes. Give a description of this change to be included in the release
  notes for FiftyOne users.

(Details in 1-2 sentences. You can just refer to another PR with a description
if this PR is part of a larger change.)

What areas of FiftyOne does this PR affect?

• App: FiftyOne application changes
• Build: Build and test infrastructure changes
• Core: Core fiftyone Python library changes
• Documentation: FiftyOne documentation changes
• Other

[lkaneda]

in your screenshot after fix, the field annotator got moved to live at the Detection level. did you change your code to make that happen? or are you showing it nested incorrectly?

[codecov]

Codecov Report

All modified lines are covered by tests ✅

❗ No coverage uploaded for pull request base (release/v0.22.1@2d75b2e). Click here to learn what that means.

Additional details and impacted files

@@                Coverage Diff                 @@
##             release/v0.22.1    #3577   +/-   ##
==================================================
  Coverage                   ?   16.02%           
==================================================
  Files                      ?      594           
  Lines                      ?    72861           
  Branches                   ?      912           
==================================================
  Hits                       ?    11676           
  Misses                     ?    61185           
  Partials                   ?        0           

Flag	Coverage Δ
app	16.02% <0.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[manivoxel51]

in your screenshot after fix, the field annotator got moved to live at the Detection level. did you change your code to make that happen? or are you showing it nested incorrectly?

good point! the two shouldn't really be compared. I added a nested annotator to detections field (in the second screenshare) which is ok to show I believe. The fix is to remove the annotator as in the first screenshare

[lkaneda]

By showing it though, doesn't it imply that it can be visualized like the other fields (and that it would appear in the side bar). This is how I understand it.

[manivoxel51]

By showing it though, doesn't it imply that it can be visualized like the other fields (and that it would appear in the side bar). This is how I understand it.

right! the annotator within ground_truth.detections is showing in the sidebar and in field visibility no? The other annotator is removed.

[lkaneda]

sweet - thanks for clarifying. lgtm!

[benjaminpkane]

I made a tweak to this because I think we only need to include the list field itself. This ensures all fields are excluded, includings list fields, dict fields, etc.

For example

dataset = foz.load_zoo_dataset("quickstart")
first = dataset.first()
first.detections["other_list"] = ["data"]
first.save()

We want other_list to be omitted from schema settings, which should be handled now.

I am seeing an issue with frame fields (see screenshot). The nested detection fields are missing detections.detections in the path. Is there an easy fix for that?

[manivoxel51]

I made a tweak to this because I think we only need to include the list field itself. This ensures all fields are excluded, includings list fields, dict fields, etc.

For example

dataset = foz.load_zoo_dataset("quickstart")
first = dataset.first()
first.detections["other_list"] = ["data"]
first.save()

We want other_list to be omitted from schema settings, which should be handled now.

I am seeing an issue with frame fields (see screenshot). The nested detection fields are missing detections.detections in the path. Is there an easy fix for that?

thanks for testing Ben! I'll address your comment out later today

[benjaminpkane]

LGTM, last comment is unrelated