Using Sample Parsers¶

This page describes how to use the SampleParser interface to add samples to your FiftyOne datasets.

The SampleParser interface provides native support for loading samples in a variety of common formats, and it can be easily extended to import datasets in custom formats, allowing you to automate the dataset loading process.

Warning

Using the SampleParser interface is not recommended. You’ll likely prefer adding samples manually or using dataset importers to load data into FiftyOne.

Adding samples to datasets¶

Basic recipe¶

The basic recipe for using the SampleParser interface to add samples to a Dataset is to create a parser of the appropriate type and then pass the parser along with an iterable of samples to the appropriate Dataset method.

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# An iterable of samples and an UnlabeledImageSampleParser to parse them
samples = ...
sample_parser = foud.ImageSampleParser  # for example

# Add the image samples to the dataset
dataset.add_images(samples, sample_parser)

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# An iterable of samples and a LabeledImageSampleParser to parse them
samples = ...
sample_parser = foud.ImageClassificationSampleParser  # for example

# Add the labeled image samples to the dataset
dataset.add_labeled_images(samples, sample_parser)

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# An iterable of samples and an UnlabeledVideoSampleParser to parse them
samples = ...
sample_parser = foud.VideoSampleParser  # for example

# Add the video samples to the dataset
dataset.add_images(samples, sample_parser)

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# An iterable of samples and a LabeledVideoSampleParser to parse them
samples = ...
sample_parser = foud.FiftyOneVideoLabelsSampleParser  # for example

# Add the labeled video samples to the dataset
dataset.add_labeled_videos(samples, sample_parser)

Note

A typical use case is that samples in the above recipe is a torch.utils.data.Dataset or an iterable generated by tf.data.Dataset.as_numpy_iterator().

Adding unlabeled images¶

FiftyOne provides a few convenient ways to add unlabeled images in FiftyOne datasets.

Adding a directory of images¶

Use Dataset.add_images_dir() to add a directory of images to a dataset:

import fiftyone as fo

dataset = fo.Dataset()

# A directory of images to add
images_dir = "/path/to/images"

# Add images to the dataset
dataset.add_images_dir(images_dir)

Adding a glob pattern of images¶

Use Dataset.add_images_patt() to add a glob pattern of images to a dataset:

import fiftyone as fo

dataset = fo.Dataset()

# A glob pattern of images to add
images_patt = "/path/to/images/*.jpg"

# Add images to the dataset
dataset.add_images_patt(images_patt)

Adding images using a SampleParser¶

Use Dataset.add_images() to add an iterable of unlabeled images that can be parsed via a specified UnlabeledImageSampleParser to a dataset.

Example

FiftyOne provides an ImageSampleParser that handles samples that contain either an image that can be converted to numpy format via np.asarray() of the path to an image on disk.

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# An iterable of images or image paths and the UnlabeledImageSampleParser
# to use to parse them
samples = ...
sample_parser = foud.ImageSampleParser

# Add images to the dataset
dataset.add_images(samples, sample_parser)

Adding labeled images¶

Use Dataset.add_labeled_images() to add an iterable of samples that can be parsed via a specified LabeledImageSampleParser to a dataset.

Example

FiftyOne provides an ImageClassificationSampleParser that handles samples that contain (image_or_path, target) tuples, where:

image_or_path is either an image that can be converted to numpy format via np.asarray() or the path to an image on disk
target is either a class ID or a label string

The snippet below adds an iterable of image classification data in the above format to a dataset:

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# An iterable of `(image_or_path, target)` tuples and the
# LabeledImageSampleParser to use to parse them
samples = ...
sample_parser = foud.ImageClassificationSampleParser

# Add labeled images to the dataset
dataset.add_labeled_images(samples, sample_parser)

Adding unlabeled videos¶

FiftyOne provides a few convenient ways to add unlabeled videos in FiftyOne datasets.

Adding a directory of videos¶

Use Dataset.add_videos_dir() to add a directory of videos to a dataset:

import fiftyone as fo

dataset = fo.Dataset()

# A directory of videos to add
videos_dir = "/path/to/videos"

# Add videos to the dataset
dataset.add_videos_dir(videos_dir)

Adding a glob pattern of videos¶

Use Dataset.add_videos_patt() to add a glob pattern of videos to a dataset:

import fiftyone as fo

dataset = fo.Dataset()

# A glob pattern of videos to add
videos_patt = "/path/to/videos/*.mp4"

# Add videos to the dataset
dataset.add_videos_patt(videos_patt)

Adding videos using a SampleParser¶

Use Dataset.add_videos() to add an iterable of unlabeled videos that can be parsed via a specified UnlabeledVideoSampleParser to a dataset.

Example

FiftyOne provides a VideoSampleParser that handles samples that directly contain the path to the video on disk.

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# An iterable of video paths and the UnlabeledVideoSampleParser to use to
# parse them
samples = ...
sample_parser = foud.VideoSampleParser

# Add videos to the dataset
dataset.add_videos(samples, sample_parser)

Adding labeled videos¶

Use Dataset.add_labeled_videos() to add an iterable of samples that can be parsed via a specified LabeledVideoSampleParser to a dataset.

Example

FiftyOne provides a VideoLabelsSampleParser that handles samples that contain (video_path, video_labels_or_path) tuples, where:

video_path is the path to a video on disk
video_labels_or_path is an eta.core.video.VideoLabels instance, a serialized dict representation of one, or the path to one on disk

The snippet below adds an iterable of labeled video samples in the above format to a dataset:

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# An iterable of `(video_path, video_labels_or_path)` tuples and the
# LabeledVideoSampleParser to use to parse them
samples = ...
sample_parser = foud.VideoLabelsSampleParser

# Add labeled videos to the dataset
dataset.add_labeled_videos(samples, sample_parser)

Ingesting samples into datasets¶

Creating FiftyOne datasets typically does not create copies of the source media, since Sample instances store the filepath to the media, not the media itself.

However, in certain circumstances, such as loading data from binary sources like TFRecords or creating a FiftyOne dataset from unorganized and/or temporary files on disk, it can be desirable to ingest the raw media for each sample into a common backing location.

FiftyOne provides support for ingesting samples and their underlying source media in both common formats and can be extended to import datasets in custom formats.

Basic recipe¶

The basic recipe for ingesting samples and their source media into a Dataset is to create a SampleParser of the appropriate type of sample that you’re loading and then pass the parser along with an iterable of samples to the appropriate Dataset method.

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# The iterable of samples and the UnlabeledImageSampleParser to use
# to parse them
samples = ...
sample_parser = foud.ImageSampleParser  # for example

# A directory in which the images will be written; If `None`, a default directory
# based on the dataset's `name` will be used
dataset_dir = ...

# Ingest the labeled image samples into the dataset
# The source images are copied into `dataset_dir`
dataset.ingest_images(samples, sample_parser, dataset_dir=dataset_dir)

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# The iterable of samples and the LabeledImageSampleParser to use
# to parse them
samples = ...
sample_parser = foud.ImageClassificationSampleParser  # for example

# A directory in which the images will be written; If `None`, a default directory
# based on the dataset's `name` will be used
dataset_dir = ...

# Add the labeled image samples to the dataset
dataset.add_labeled_images(samples, sample_parser, dataset_dir=dataset_dir)

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# The iterable of samples and the UnlabeledVideoSampleParser to use
# to parse them
samples = ...
sample_parser = foud.VideoSampleParser  # for example

# A directory in which the videos will be written; If `None`, a default directory
# based on the dataset's `name` will be used
dataset_dir = ...

# Ingest the labeled video samples into the dataset
# The source videos are copied into `dataset_dir`
dataset.ingest_videos(samples, sample_parser, dataset_dir=dataset_dir)

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# The iterable of samples and the LabeledVideoSampleParser to use
# to parse them
samples = ...
sample_parser = foud.VideoLabelsSampleParser  # for example

# A directory in which the videos will be written; If `None`, a default directory
# based on the dataset's `name` will be used
dataset_dir = ...

# Add the labeled video samples to the dataset
dataset.add_labeled_videos(samples, sample_parser, dataset_dir=dataset_dir)

Note

A typical use case is that samples in the above recipe is a torch.utils.data.Dataset or an iterable generated by tf.data.Dataset.as_numpy_iterator().

Ingesting unlabeled images¶

Use Dataset.ingest_images() to ingest an iterable of unlabeled images that can be parsed via a specified UnlabeledImageSampleParser into a dataset.

The has_image_path property of the parser may either be True or False. If the parser provides image paths, the source images will be directly copied from their source locations into the backing directory for the dataset; otherwise, the image will be read in-memory via get_image() and then written to the backing directory.

Example

FiftyOne provides an ImageSampleParser that handles samples that contain either an image that can be converted to numpy format via np.asarray() of the path to an image on disk.

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# An iterable of images or image paths and the UnlabeledImageSampleParser
# to use to parse them
samples = ...
sample_parser = foud.ImageSampleParser

# A directory in which the images will be written; If `None`, a default directory
# based on the dataset's `name` will be used
dataset_dir = ...

# Ingest the images into the dataset
# The source images are copied into `dataset_dir`
dataset.ingest_images(samples, sample_parser, dataset_dir=dataset_dir)

Ingesting labeled images¶

Use Dataset.ingest_labeled_images() to ingest an iterable of samples that can be parsed via a specified LabeledImageSampleParser into a dataset.

The has_image_path property of the parser may either be True or False. If the parser provides image paths, the source images will be directly copied from their source locations into the backing directory for the dataset; otherwise, the image will be read in-memory via get_image() and then written to the backing directory.

Example

FiftyOne provides an ImageClassificationSampleParser that handles samples that contain (image_or_path, target) tuples, where:

image_or_path is either an image that can be converted to numpy format via np.asarray() or the path to an image on disk
target is either a class ID or a label string

The snippet below ingests an iterable of image classification data in the above format intoa a FiftyOne dataset:

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# An iterable of `(image_or_path, target)` tuples and the
# LabeledImageSampleParser to use to parse them
samples = ...
sample_parser = foud.ImageClassificationSampleParser  # for example

# A directory in which the images will be written; If `None`, a default directory
# based on the dataset's `name` will be used
dataset_dir = ...

# Ingest the labeled images into the dataset
# The source images are copied into `dataset_dir`
dataset.ingest_labeled_images(samples, sample_parser, dataset_dir=dataset_dir)

Ingesting unlabeled videos¶

Use Dataset.ingest_videos() to ingest an iterable of unlabeled videos that can be parsed via a specified UnlabeledVideoSampleParser into a dataset.

The source videos will be directly copied from their source locations into the backing directory for the dataset.

Example

FiftyOne provides a VideoSampleParser that handles samples that directly contain the paths to videos on disk.

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# An iterable of videos or video paths and the UnlabeledVideoSampleParser
# to use to parse them
samples = ...
sample_parser = foud.VideoSampleParser

# A directory in which the videos will be written; If `None`, a default directory
# based on the dataset's `name` will be used
dataset_dir = ...

# Ingest the videos into the dataset
# The source videos are copied into `dataset_dir`
dataset.ingest_videos(samples, sample_parser, dataset_dir=dataset_dir)

Ingesting labeled videos¶

Use Dataset.ingest_labeled_videos() to ingest an iterable of samples that can be parsed via a specified LabeledVideoSampleParser into a dataset.

The source videos will be directly copied from their source locations into the backing directory for the dataset.

Example

FiftyOne provides a VideoLabelsSampleParser that handles samples that contain (video_path, video_labels_or_path) tuples, where:

video_path is the path to a video on disk
video_labels_or_path is an eta.core.video.VideoLabels instance, a serialized dict representation of one, or the path to one on disk

The snippet below ingests an iterable of labeled videos in the above format into a FiftyOne dataset:

import fiftyone as fo
import fiftyone.utils.data as foud

dataset = fo.Dataset()

# An iterable of `(video_path, video_labels_or_path)` tuples and the
# LabeledVideoSampleParser to use to parse them
samples = ...
sample_parser = foud.VideoLabelsSampleParser  # for example

# A directory in which the videos will be written; If `None`, a default directory
# based on the dataset's `name` will be used
dataset_dir = ...

# Ingest the labeled videos into the dataset
# The source videos are copied into `dataset_dir`
dataset.ingest_labeled_videos(samples, sample_parser, dataset_dir=dataset_dir)

Built-in SampleParser classes¶

The table below lists the common data formats for which FiftyOne provides built-in SampleParser implementations. You can also write a custom SampleParser to automate the parsing of samples in your own custom data format.

You can use a SampleParser to add samples to datasets and ingest samples into datasets.

SampleParser	Description
`ImageSampleParser`	A sample parser that parses raw image samples.
`VideoSampleParser`	A sample parser that parses raw video samples.
`ImageClassificationSampleParser`	Generic parser for image classification samples whose labels are represented as `Classification` instances.
`ImageDetectionSampleParser`	Generic parser for image detection samples whose labels are represented as `Detections` instances.
`ImageLabelsSampleParser`	Generic parser for image detection samples whose labels are stored in ETA ImageLabels format.
`FiftyOneImageClassificationSampleParser`	Parser for samples in FiftyOne image classification datasets. See `FiftyOneImageClassificationDataset` for format details.
`FiftyOneImageDetectionSampleParser`	Parser for samples in FiftyOne image detection datasets. See `FiftyOneImageDetectionDataset` for format details.
`FiftyOneImageLabelsSampleParser`	Parser for samples in FiftyOne image labels datasets. See `FiftyOneImageLabelsDataset` for format details.
`FiftyOneVideoLabelsSampleParser`	Parser for samples in FiftyOne video labels datasets. See `FiftyOneVideoLabelsDataset` for format details.
`TFImageClassificationSampleParser`	Parser for image classification samples stored as TFRecords.
`TFObjectDetectionSampleParser`	Parser for image detection samples stored in TF Object Detection API format.

Writing a custom SampleParser¶

FiftyOne provides a variety of built-in SampleParser classes to parse data in common formats. However, if your samples are stored in a custom format, you can provide a custom SampleParser class and provide it to FiftyOne when adding or ingesting samples into your datasets.

The SampleParser interface provides a mechanism for defining methods that parse a data sample that is stored in a particular (external to FiftyOne) format and return various elements of the sample in a format that FiftyOne understands.

SampleParser itself is an abstract interface; the concrete interface that you should implement is determined by the type of samples that you are importing. For example, LabeledImageSampleParser defines an interface for parsing information from a labeled image sample, such as the path to the image on disk, the image itself, metadata about the image, and the label (e.g., classification or object detections) associated with the image.

To define a custom parser for unlabeled images, implement the UnlabeledImageSampleParser interface.

The pseudocode below provides a template for a custom UnlabeledImageSampleParser:

import fiftyone.utils.data as foud

class CustomUnlabeledImageSampleParser(foud.UnlabeledImageSampleParser):
    """Custom parser for unlabeled image samples."""

    @property
    def has_image_path(self):
        """Whether this parser produces paths to images on disk for samples
        that it parses.
        """
        # Return True or False here
        pass

    @property
    def has_image_metadata(self):
        """Whether this parser produces
        :class:`fiftyone.core.metadata.ImageMetadata` instances for samples
        that it parses.
        """
        # Return True or False here
        pass

    def get_image(self):
        """Returns the image from the current sample.

        Returns:
            a numpy image
        """
        # Return the image in `self.current_sample` here
        pass

    def get_image_path(self):
        """Returns the image path for the current sample.

        Returns:
            the path to the image on disk
        """
        # Return the image path for `self.current_sample` here, or raise
        # an error if `has_image_path == False`
        pass

    def get_image_metadata(self):
        """Returns the image metadata for the current sample.

        Returns:
            a :class:`fiftyone.core.metadata.ImageMetadata` instance
        """
        # Return the image metadata for `self.current_sample` here, or
        # raise an error if `has_image_metadata == False`
        pass

When Dataset.add_images() is called with a custom UnlabeledImageSampleParser, the import is effectively performed via the pseudocode below:

import fiftyone as fo

dataset = fo.Dataset(...)
samples = ...
sample_parser = CustomUnlabeledImageSampleParser(...)

for sample in samples:
    sample_parser.with_sample(sample)

    image_path = sample_parser.get_image_path()

    if sample_parser.has_image_metadata:
        metadata = sample_parser.get_image_metadata()
    else:
        metadata = None

    sample = fo.Sample(filepath=image_path, metadata=metadata)

    dataset.add_sample(sample)

The base SampleParser interface provides a with_sample() method that ingests the next sample and makes it available via the current_sample property of the parser. Subsequent calls to the parser’s get_XXX() methods return information extracted from the current sample.

The UnlabeledImageSampleParser interface provides a has_image_path property that declares whether the sample parser can return the path to the current sample’s image on disk via get_image_path(). Similarly, the has_image_metadata property that declares whether the sample parser can return an ImageMetadata for the current sample’s image via get_image_metadata().

By convention, all UnlabeledImageSampleParser implementations must make the current sample’s image available via get_image().

To define a custom parser for labeled images, implement the LabeledImageSampleParser interface.

The pseudocode below provides a template for a custom LabeledImageSampleParser:

import fiftyone.utils.data as foud

class CustomLabeledImageSampleParser(foud.LabeledImageSampleParser):
    """Custom parser for labeled image samples."""

    @property
    def has_image_path(self):
        """Whether this parser produces paths to images on disk for samples
        that it parses.
        """
        # Return True or False here
        pass

    @property
    def has_image_metadata(self):
        """Whether this parser produces
        :class:`fiftyone.core.metadata.ImageMetadata` instances for samples
        that it parses.
        """
        # Return True or False here
        pass

    @property
    def label_cls(self):
        """The :class:`fiftyone.core.labels.Label` class(es) returned by this
        parser.

        This can be any of the following:

        -   a :class:`fiftyone.core.labels.Label` class. In this case, the
            parser is guaranteed to return labels of this type
        -   a list or tuple of :class:`fiftyone.core.labels.Label` classes. In
            this case, the parser can produce a single label field of any of
            these types
        -   a dict mapping keys to :class:`fiftyone.core.labels.Label` classes.
            In this case, the parser will return label dictionaries with keys
            and value-types specified by this dictionary. Not all keys need be
            present in the imported labels
        -   ``None``. In this case, the parser makes no guarantees about the
            labels that it may return
        """
        # Return the appropriate value here
        pass

    def get_image(self):
        """Returns the image from the current sample.

        Returns:
            a numpy image
        """
        # Return the image in `self.current_sample` here
        pass

    def get_image_path(self):
        """Returns the image path for the current sample.

        Returns:
            the path to the image on disk
        """
        # Return the image path for `self.current_sample` here, or raise
        # an error if `has_image_path == False`
        pass

    def get_image_metadata(self):
        """Returns the image metadata for the current sample.

        Returns:
            a :class:`fiftyone.core.metadata.ImageMetadata` instance
        """
        # Return the image metadata for `self.current_sample` here, or
        # raise an error if `has_image_metadata == False`
        pass

    def get_label(self):
        """Returns the label for the current sample.

        Returns:
            a :class:`fiftyone.core.labels.Label` instance, or a dictionary
            mapping field names to :class:`fiftyone.core.labels.Label`
            instances, or ``None`` if the sample is unlabeled
        """
        # Return the label for `self.current_sample` here
        pass

When Dataset.add_labeled_images() is called with a custom LabeledImageSampleParser, the import is effectively performed via the pseudocode below:

import fiftyone as fo

dataset = fo.Dataset(...)

samples = ...
sample_parser = CustomLabeledImageSampleParser(...)
label_field = ...

if isinstance(label_field, dict):
    label_key = lambda k: label_field.get(k, k)
elif label_field is not None:
    label_key = lambda k: label_field + "_" + k
else:
    label_field = "ground_truth"
    label_key = lambda k: k

for sample in samples:
    sample_parser.with_sample(sample)

    image_path = sample_parser.get_image_path()

    if sample_parser.has_image_metadata:
        metadata = sample_parser.get_image_metadata()
    else:
        metadata = None

    label = sample_parser.get_label()

    sample = fo.Sample(filepath=image_path, metadata=metadata)

    if isinstance(label, dict):
        sample.update_fields({label_key(k): v for k, v in label.items()})
    elif label is not None:
        sample[label_field] = label

    dataset.add_sample(sample)

The base SampleParser interface provides a with_sample() method that ingests the next sample and makes it available via the current_sample property of the parser. Subsequent calls to the parser’s get_XXX() methods return information extracted from the current sample.

The LabeledImageSampleParser interface provides a has_image_path property that declares whether the sample parser can return the path to the current sample’s image on disk via get_image_path(). Similarly, the has_image_metadata property that declares whether the sample parser can return an ImageMetadata for the current sample’s image via get_image_metadata(). Additionality, the label_cls property of the parser declares the type of label(s) that the parser will produce.

By convention, all LabeledImageSampleParser implementations must make the current sample’s image available via get_image() , and they must make the current sample’s label available via get_label().

To define a custom parser for unlabeled videos, implement the UnlabeledVideoSampleParser interface.

The pseudocode below provides a template for a custom UnlabeledVideoSampleParser:

import fiftyone.utils.data as foud

class CustomUnlabeledVideoSampleParser(foud.UnlabeledVideoSampleParser):
    """Custom parser for unlabeled video samples."""

    @property
    def has_video_metadata(self):
        """Whether this parser produces
        :class:`fiftyone.core.metadata.VideoMetadata` instances for samples
        that it parses.
        """
        # Return True or False here
        pass

    def get_video_path(self):
        """Returns the video path for the current sample.

        Returns:
            the path to the video on disk
        """
        # Return the video path for `self.current_sample` here
        pass

    def get_video_metadata(self):
        """Returns the video metadata for the current sample.

        Returns:
            a :class:`fiftyone.core.metadata.VideoMetadata` instance
        """
        # Return the video metadata for `self.current_sample` here, or
        # raise an error if `has_video_metadata == False`
        pass

When Dataset.add_videos() is called with a custom UnlabeledVideoSampleParser, the import is effectively performed via the pseudocode below:

import fiftyone as fo

dataset = fo.Dataset(...)
samples = ...
sample_parser = CustomUnlabeledVideoSampleParser(...)

for sample in samples:
    sample_parser.with_sample(sample)

    video_path = sample_parser.get_video_path()

    if sample_parser.has_image_metadata:
        metadata = sample_parser.get_image_metadata()
    else:
        metadata = None

    sample = fo.Sample(filepath=video_path, metadata=metadata)

    dataset.add_sample(sample)

The base SampleParser interface provides a with_sample() method that ingests the next sample and makes it available via the current_sample property of the parser. Subsequent calls to the parser’s get_XXX() methods return information extracted from the current sample.

The UnlabeledVideoSampleParser interface provides a get_video_path() to get the video path for the current sample. The has_video_metadata property that declares whether the sample parser can return a VideoMetadata for the current sample’s video via get_video_metadata().

To define a custom parser for labeled videos, implement the LabeledVideoSampleParser interface.

The pseudocode below provides a template for a custom LabeledVideoSampleParser:

import fiftyone.utils.data as foud

class CustomLabeledVideoSampleParser(foud.LabeledVideoSampleParser):
    """Custom parser for labeled video samples."""

    @property
    def has_video_metadata(self):
        """Whether this parser produces
        :class:`fiftyone.core.metadata.VideoMetadata` instances for samples
        that it parses.
        """
        # Return True or False here
        pass

    @property
    def label_cls(self):
        """The :class:`fiftyone.core.labels.Label` class(es) returned by this
        parser within the sample-level labels that it produces.

        This can be any of the following:

        -   a :class:`fiftyone.core.labels.Label` class. In this case, the
            parser is guaranteed to return sample-level labels of this type
        -   a list or tuple of :class:`fiftyone.core.labels.Label` classes. In
            this case, the parser can produce a single sample-level label field
            of any of these types
        -   a dict mapping keys to :class:`fiftyone.core.labels.Label` classes.
            In this case, the parser will return sample-level label
            dictionaries with keys and value-types specified by this
            dictionary. Not all keys need be present in the imported labels
        -   ``None``. In this case, the parser makes no guarantees about the
            sample-level labels that it may return
        """
        # Return the appropriate value here
        pass

    @property
    def frame_labels_cls(self):
        """The :class:`fiftyone.core.labels.Label` class(es) returned by this
        parser within the frame labels that it produces.

        This can be any of the following:

        -   a :class:`fiftyone.core.labels.Label` class. In this case, the
            parser is guaranteed to return frame labels of this type
        -   a list or tuple of :class:`fiftyone.core.labels.Label` classes. In
            this case, the parser can produce a single frame label field of any
            of these types
        -   a dict mapping keys to :class:`fiftyone.core.labels.Label` classes.
            In this case, the parser will return frame label dictionaries with
            keys and value-types specified by this dictionary. Not all keys
            need be present in each frame
        -   ``None``. In this case, the parser makes no guarantees about the
            frame labels that it may return
        """
        # Return the appropriate value here
        pass

    def get_video_path(self):
        """Returns the video path for the current sample.

        Returns:
            the path to the video on disk
        """
        # Return the video path for `self.current_sample` here
        pass

    def get_video_metadata(self):
        """Returns the video metadata for the current sample.

        Returns:
            a :class:`fiftyone.core.metadata.VideoMetadata` instance
        """
        # Return the video metadata for `self.current_sample` here, or
        # raise an error if `has_video_metadata == False`
        pass

    def get_label(self):
        """Returns the sample-level labels for the current sample.

        Returns:
            a :class:`fiftyone.core.labels.Label` instance, or a dictionary
            mapping field names to :class:`fiftyone.core.labels.Label`
            instances, or ``None`` if the sample has no sample-level labels
        """
        # Return the sample labels for `self.current_sample` here
        pass

    def get_frame_labels(self):
        """Returns the frame labels for the current sample.

        Returns:
            a dictionary mapping frame numbers to dictionaries that map label
            fields to :class:`fiftyone.core.labels.Label` instances for each
            video frame, or ``None`` if the sample has no frame labels
        """
        # Return the frame labels for `self.current_sample` here
        pass

When Dataset.add_labeled_videos() is called with a custom LabeledVideoSampleParser, the import is effectively performed via the pseudocode below:

import fiftyone as fo

dataset = fo.Dataset(...)
samples = ...
sample_parser = CustomLabeledVideoSampleParser(...)
label_field = ...

if isinstance(label_field, dict):
    label_key = lambda k: label_field.get(k, k)
elif label_field is not None:
    label_key = lambda k: label_field + "_" + k
else:
    label_field = "ground_truth"
    label_key = lambda k: k

for sample in samples:
    sample_parser.with_sample(sample)

    video_path = sample_parser.get_video_path()

    if sample_parser.has_video_metadata:
        metadata = sample_parser.get_video_metadata()
    else:
        metadata = None

    label = sample_parser.get_label()
    frames = sample_parser.get_frame_labels()

    sample = fo.Sample(filepath=video_path, metadata=metadata)

    if isinstance(label, dict):
        sample.update_fields({label_key(k): v for k, v in label.items()})
    elif label is not None:
        sample[label_field] = label

    if frames is not None:
        frame_labels = {}

        for frame_number, _label in frames.items():
            if isinstance(_label, dict):
                frame_labels[frame_number] = {
                    label_key(k): v for k, v in _label.items()
                }
            elif _label is not None:
                frame_labels[frame_number] = {label_field: _label}

        sample.frames.merge(frame_labels)

    dataset.add_sample(sample)

The base SampleParser interface provides a with_sample() method that ingests the next sample and makes it available via the current_sample property of the parser. Subsequent calls to the parser’s get_XXX() methods return information extracted from the current sample.

The LabeledVideoSampleParser interface provides a get_video_path() to get the video path for the current sample. The has_video_metadata property that declares whether the sample parser can return a VideoMetadata for the current sample’s video via get_video_metadata().

The label_cls property of the parser declares the type of sample-level label(s) that the parser may produce (if any). The frame_labels_cls property of the parser declares the type of frame-level label(s) that the parser may produce (if any). By convention, all LabeledVideoSampleParser implementations must make the current sample’s sample-level labels available via get_label() and its frame-level labels available via get_frame_labels().