Labelbox Annotation Plugin

A plugin that contains utilities for integrating FiftyOne with the Labelbox annotation integration backend described here.

Installation

Local Python environment:

fiftyone plugins download \
    https://github.com/ehofesmann/fiftyone_labelbox

Refer to the main README for more information about managing downloaded plugins and developing plugins locally.

Set the FIFTYONE_LABELBOX_API_KEY=XXXXX environment variable following these instructions.

Install in FiftyOne Teams environment:

1. Download the zip of this repository:

2. Install it through your FiftyOne Teams App.

   (Alternative): Upload it through the FiftyOne Teams management SDK.

3. Add the FIFTYONE_LABELBOX_API_KEY secret in your FiftyOne Teams App.

Usage

In the FiftyOne App, press ` or click the Browse operations action to open the Operators list. Select one of the operators mentioned below.

This plugin can also be used entirely through Python by loading the relevant methods directly from the plugin through foo.get_operator().

Operators

request_annotations

You can use this operator to create annotation tasks for the current dataset or view.

This operator is essentially a wrapper around the request annotations Python workflow:

import fiftyone.operators as foo
annotate = foo.get_operator("@ehofesmann/labelbox/request_annotations")
annotate(
    dataset_or_view,
    anno_key,
    label_schema=...,
    ...
)

where the operator's form allows you to configure the annotation key, annotation backend, label schema, and any other applicable fields for your annotation backend.

load_annotations

You can use this operator to load annotations for existing runs back onto your dataset.

This operator is essentially a wrapper around the load annotations Python workflow:

import fiftyone.operators as foo
load_annotations = foo.get_operator("@ehofesmann/labelbox/load_annotations")

load_annotations(dataset, anno_key, ...)

where the operator's form allows you to configure the annotation key and related options.

delete_annotation_run

You can use this operator to delete annotation runs.

This operator is essentially a wrapper around delete_annotation_run():

import fiftyone.operators as foo
delete_annotation_run = foo.get_operator("@ehofesmann/labelbox/delete_annotation_run")

delete_annotation_run(dataset, anno_key, cleanup=True)