fiftyone.core.view

Dataset views.

Copyright 2017-2020, Voxel51, Inc.

voxel51.com

Classes

DatasetView(dataset)

A view into a fiftyone.core.dataset.Dataset.

class fiftyone.core.view.DatasetView(dataset)

Bases: fiftyone.core.collections.SampleCollection

A view into a fiftyone.core.dataset.Dataset.

Dataset views represent ordered collections of subsets of samples in a dataset.

Operations on dataset views are designed to be chained together to yield the desired subset of the dataset, which is then iterated over to directly access the sample views. Each stage in the pipeline defining a dataset view is represented by a fiftyone.core.stages.ViewStage instance.

The stages of a dataset view specify:

• what subset of samples (and their order) should be included

• what “parts” (fields and their elements) of the sample should be included

Samples retrieved from dataset views are returns as fiftyone.core.sample.SampleView objects, as opposed to fiftyone.core.sample.Sample objects, since they may contain a subset of the sample’s content.

Example use:

# Print paths for 5 random samples from the test split of a dataset
view = dataset.match_tag("test").take(5)
for sample in view:
    print(sample.filepath)

Parameters

dataset – a fiftyone.core.dataset.Dataset

property media_type

The media type of the underlying dataset.

property name

The name of the view.

property dataset_name

The name of the underlying dataset.

property info

The fiftyone.core.dataset.Dataset.info() dict of the underlying dataset.

property stages

The list of fiftyone.core.stages.ViewStage instances in this view’s pipeline.

summary()

Returns a string summary of the view.

Returns

a string summary

iter_samples()

Returns an iterator over the samples in the view.

Returns

an iterator over fiftyone.core.sample.SampleView instances

get_field_schema(ftype=None, embedded_doc_type=None, include_private=False)

Returns a schema dictionary describing the fields of the samples in the view.

Parameters

• ftype (None) – an optional field type to which to restrict the returned schema. Must be a subclass of fiftyone.core.fields.Field

• embedded_doc_type (None) – an optional embedded document type to which to restrict the returned schema. Must be a subclass of fiftyone.core.odm.BaseEmbeddedDocument

• include_private (False) – whether to include fields that start with _ in the returned schema

Returns

an OrderedDict mapping field names to field types

get_frame_field_schema(ftype=None, embedded_doc_type=None, include_private=False)

Returns a schema dictionary describing the fields of the frames of the samples in the view.

Only applicable for video datasets.

Parameters

• ftype (None) – an optional field type to which to restrict the returned schema. Must be a subclass of fiftyone.core.fields.Field

• embedded_doc_type (None) – an optional embedded document type to which to restrict the returned schema. Must be a subclass of fiftyone.core.odm.BaseEmbeddedDocument

• include_private (False) – whether to include fields that start with _ in the returned schema

Returns

a dictionary mapping field names to field types, or None if the dataset is not a video dataset

get_tags()

Returns the list of unique tags of samples in the view.

Returns

a list of tags

create_index(field)

Creates a database index on the given field, enabling efficient sorting on that field.

Parameters

field – the name of the field to index

aggregate(pipeline=None)

Calls the view’s current MongoDB aggregation pipeline.

Parameters

pipeline (None) – an optional aggregation pipeline (list of dicts) to append to the view’s pipeline before calling it

Returns

an iterable over the aggregation result

to_dict(rel_dir=None)

Returns a JSON dictionary representation of the view.

Parameters

rel_dir (None) – a relative directory to remove from the filepath of each sample, if possible. The path is converted to an absolute path (if necessary) via os.path.abspath(os.path.expanduser(rel_dir)). The typical use case for this argument is that your source data lives in a single directory and you wish to serialize relative, rather than absolute, paths to the data within that directory

Returns

a JSON dict

add_stage(stage)

Applies the given fiftyone.core.stages.ViewStage to the collection.

Parameters

stage – a fiftyone.core.stages.ViewStage

Returns

a fiftyone.core.view.DatasetView

Raises

fiftyone.core.stages.ViewStageError – if the stage was not a valid stage for this collection

compute_metadata(overwrite=False)

Populates the metadata field of all samples in the collection.

Any samples with existing metadata are skipped, unless overwrite == True.

Parameters

overwrite (False) – whether to overwrite existing metadata

draw_labels(anno_dir, label_fields=None, overwrite=False, annotation_config=None)

Renders annotated versions of the samples in the collection with label field(s) overlaid to the given directory.

The filenames of the sample data are maintained, unless a name conflict would occur in anno_dir, in which case an index of the form "-%d" % count is appended to the base filename.

Images are written in format fo.config.default_image_ext.

Parameters

• anno_dir – the directory to write the annotated files

• label_fields (None) – a list of fiftyone.core.labels.Label fields to render. By default, all fiftyone.core.labels.Label fields are drawn

• overwrite (False) – whether to delete anno_dir if it exists before rendering the labels

• annotation_config (None) – an fiftyone.utils.annotations.AnnotationConfig specifying how to render the annotations

Returns

the list of paths to the labeled images

exclude(sample_ids)

Excludes the samples with the given IDs from the collection.

Examples:

import fiftyone as fo

dataset = fo.load_dataset(...)

#
# Exclude a single sample from a dataset
#

view = dataset.exclude("5f3c298768fd4d3baf422d2f")

#
# Exclude a list of samples from a dataset
#

view = dataset.exclude([
    "5f3c298768fd4d3baf422d2f",
    "5f3c298768fd4d3baf422d30"
])

Parameters

sample_ids – a sample ID or iterable of sample IDs

Returns

a fiftyone.core.view.DatasetView

exclude_fields(field_names)

Excludes the fields with the given names from the returned fiftyone.core.sample.SampleView instances.

Note that default fields cannot be excluded.

Examples:

import fiftyone as fo

dataset = fo.load_dataset(...)

#
# Exclude a field from all samples in a dataset
#

view = dataset.exclude_fields("predictions")

#
# Exclude a list of fields from all samples in a dataset
#

view = dataset.exclude_fields(["ground_truth", "predictions"])

Parameters

field_names – a field name or iterable of field names to exclude

Returns

a fiftyone.core.view.DatasetView

exists(field, bool=True)

Returns a view containing the samples that have (or do not have) a non-None value for the given field.

Examples:

import fiftyone as fo

dataset = fo.load_dataset(...)

#
# Only include samples that have a value in their `predictions`
# field
#

view = dataset.exists("predictions")

#
# Only include samples that do NOT have a value in their
# `predictions` field
#

view = dataset.exists("predictions", False)

Parameters

• field – the field

• bool (True) – whether to check if the field exists (True) or does not exist (False)

Returns

a fiftyone.core.view.DatasetView

export(export_dir=None, dataset_type=None, dataset_exporter=None, label_field=None, label_prefix=None, labels_dict=None, overwrite=False, **kwargs)

Exports the samples in the collection to disk.

Provide either export_dir and dataset_type or dataset_exporter to perform an export.

See this guide for more details about exporting datasets in custom formats by defining your own DatasetExporter.

Parameters

• export_dir (None) – the directory to which to export the samples in format dataset_type

• dataset_type (None) – the fiftyone.types.dataset_types.Dataset type to write. If not specified, the default type for label_field is used

• dataset_exporter (None) – a fiftyone.utils.data.exporters.DatasetExporter to use to export the samples

• label_field (None) – the name of the label field to export, if applicable. If none of label_field, label_prefix, and labels_dict are specified and the requested output type is a labeled dataset, the first field of compatible type for the output format is used

• label_prefix (None) – a label field prefix; all fields whose name starts with the given prefix will be exported (with the prefix removed when constructing the label dicts). This parameter can only be used when the exporter can handle dictionaries of labels

• labels_dict (None) – a dictionary mapping label field names to keys to use when constructing the label dict to pass to the exporter. This parameter can only be used when the exporter can handle dictionaries of labels

• overwrite (False) – when an export_dir is provided, whether to delete the existing directory before performing the export

• **kwargs – optional keyword arguments to pass to dataset_type.get_dataset_exporter_cls(export_dir, **kwargs)

filter_classifications(field, filter)

Filters the classifications of the given fiftyone.core.labels.Classifications field.

Elements of <field>.classifications for which filter returns False are omitted from the field.

Examples:

import fiftyone as fo
from fiftyone import ViewField as F

dataset = fo.load_dataset(...)

#
# Only include classifications in the `predictions` field whose
# `confidence` greater than 0.8
#

view = dataset.filter_classifications(
    "predictions", F("confidence") > 0.8
)

#
# Only include classifications in the `predictions` field whose
# `label` is "cat" or "dog"
#

view = dataset.filter_classifications(
    "predictions", F("label").is_in(["cat", "dog"])
)

Parameters

• field – the fiftyone.core.labels.Classifications field

• filter – a fiftyone.core.expressions.ViewExpression or MongoDB expression that returns a boolean describing the filter to apply

Returns

a fiftyone.core.view.DatasetView

filter_detections(field, filter)

Filters the detections of the given fiftyone.core.labels.Detections field.

Elements of <field>.detections for which filter returns False are omitted from the field.

Examples:

import fiftyone as fo
from fiftyone import ViewField as F

dataset = fo.load_dataset(...)

#
# Only include detections in the `predictions` field whose
# `confidence` is greater than 0.8
#

view = dataset.filter_detections(
    "predictions", F("confidence") > 0.8
)

#
# Only include detections in the `predictions` field whose `label`
# is "cat" or "dog"
#

view = dataset.filter_detections(
    "predictions", F("label").is_in(["cat", "dog"])
)

#
# Only include detections in the `predictions` field whose bounding
# box area is smaller than 0.2
#

# bbox is in [top-left-x, top-left-y, width, height] format
bbox_area = F("bounding_box")[2] * F("bounding_box")[3]

view = dataset.filter_detections("predictions", bbox_area < 0.2)

Parameters

• field – the fiftyone.core.labels.Detections field

• filter –

  a fiftyone.core.expressions.ViewExpression or MongoDB expression that returns a boolean describing the filter to apply

Returns

a fiftyone.core.view.DatasetView

filter_field(field, filter)

Filters the values of the given field of the samples.

Values of field for which filter returns False are replaced with None.

Examples:

import fiftyone as fo
from fiftyone import ViewField as F

dataset = fo.load_dataset(...)

#
# Only include classifications in the `predictions` field (assume
# it is a `Classification` field) whose `label` is "cat"
#

view = dataset.filter_field("predictions", F("label") == "cat")

#
# Only include classifications in the `predictions` field (assume
# it is a `Classification` field) whose `confidence` is greater
# than 0.8
#

view = dataset.filter_field("predictions", F("confidence") > 0.8)

Parameters

• field – the field to filter

• filter –

  a fiftyone.core.expressions.ViewExpression or MongoDB expression that returns a boolean describing the filter to apply

Returns

a fiftyone.core.view.DatasetView

filter_keypoints(field, filter)

Filters the keypoints of the given fiftyone.core.labels.Keypoints field.

Elements of <field>.keypoints for which filter returns False are omitted from the field.

Examples:

import fiftyone as fo
from fiftyone import ViewField as F
from fiftyone.core.stages import FilterKeypoints

dataset = fo.load_dataset(...)

#
# Only include keypoints in the `predictions` field whose `label`
# is "face"
#

stage = FilterKeypoints("predictions", F("label") == "face")
view = dataset.add_stage(stage)

#
# Only include keypoints in the `predictions` field with at least
# 10 points
#

stage = FilterKeypoints("predictions", F("points").length() >= 10)
view = dataset.add_stage(stage)

Parameters

• field – the fiftyone.core.labels.Keypoints field

• filter –

  a fiftyone.core.expressions.ViewExpression or MongoDB expression that returns a boolean describing the filter to apply

Returns

a fiftyone.core.view.DatasetView

filter_polylines(field, filter)

Filters the polylines of the given fiftyone.core.labels.Polylines field.

Elements of <field>.polylines for which filter returns False are omitted from the field.

Examples:

import fiftyone as fo
from fiftyone import ViewField as F
from fiftyone.core.stages import FilterPolylines

dataset = fo.load_dataset(...)

#
# Only include polylines in the `predictions` field that are filled
#

stage = FilterPolylines("predictions", F("filled"))
view = dataset.add_stage(stage)

#
# Only include polylines in the `predictions` field whose `label`
# is "lane"
#

stage = FilterPolylines("predictions", F("label") == "lane")
view = dataset.add_stage(stage)

#
# Only include polylines in the `predictions` field with at least
# 10 vertices
#

num_vertices = F("points").map(F().length()).sum()
stage = FilterPolylines("predictions", num_vertices >= 10)
view = dataset.add_stage(stage)

Parameters

• field – the fiftyone.core.labels.Polylines field

• filter –

  a fiftyone.core.expressions.ViewExpression or MongoDB expression that returns a boolean describing the filter to apply

Returns

a fiftyone.core.view.DatasetView

first()

Returns the first sample in the collection.

Returns

a fiftyone.core.sample.Sample or fiftyone.core.sample.SampleView

Raises

ValueError – if the collection is empty

head(num_samples=3)

Returns a list of the first few samples in the collection.

If fewer than num_samples samples are in the collection, only the available samples are returned.

Parameters

num_samples (3) – the number of samples

Returns

a list of fiftyone.core.sample.Sample objects

last()

Returns the last sample in the collection.

Returns

a fiftyone.core.sample.Sample or fiftyone.core.sample.SampleView

Raises

ValueError – if the collection is empty

limit(limit)

Returns a view with at most the given number of samples.

Examples:

import fiftyone as fo

dataset = fo.load_dataset(...)

#
# Only include the first 10 samples in the view
#

view = dataset.limit(10)

Parameters

limit – the maximum number of samples to return. If a non-positive number is provided, an empty view is returned

Returns

a fiftyone.core.view.DatasetView

classmethod list_view_stages()

Returns a list of all available methods on this collection that apply fiftyone.core.stages.ViewStage operations that return fiftyone.core.view.DatasetView instances.

Returns

a list of SampleCollection method names

make_unique_field_name(root='')

Makes a unique field name with the given root name for the collection.

Parameters

root – an optional root for the output field name

Returns

the field name

match(filter)

Filters the samples in the collection by the given filter.

Samples for which filter returns False are omitted.

Examples:

import fiftyone as fo
from fiftyone import ViewField as F

dataset = fo.load_dataset(...)

#
# Only include samples whose `filepath` ends with ".jpg"
#

view = dataset.match(F("filepath").ends_with(".jpg"))

#
# Only include samples whose `predictions` field (assume it is a
# `Classification` field) has `label` of "cat"
#

view = dataset.match(F("predictions").label == "cat"))

#
# Only include samples whose `predictions` field (assume it is a
# `Detections` field) has at least 5 detections
#

view = dataset.match(F("predictions").detections.length() >= 5)

#
# Only include samples whose `predictions` field (assume it is a
# `Detections` field) has at least one detection with area smaller
# than 0.2
#

# bbox is in [top-left-x, top-left-y, width, height] format
pred_bbox = F("predictions.detections.bounding_box")
pred_bbox_area = pred_bbox[2] * pred_bbox[3]

view = dataset.match((pred_bbox_area < 0.2).length() > 0)

Parameters

filter –

a fiftyone.core.expressions.ViewExpression or MongoDB expression that returns a boolean describing the filter to apply

Returns

a fiftyone.core.view.DatasetView

match_tag(tag)

Returns a view containing the samples that have the given tag.

Examples:

import fiftyone as fo

dataset = fo.load_dataset(...)

#
# Only include samples that have the "test" tag
#

view = dataset.match_tag("test")

Parameters

tag – a tag

Returns

a fiftyone.core.view.DatasetView

match_tags(tags)

Returns a view containing the samples that have any of the given tags.

To match samples that must contain multiple tags, chain multiple match_tag() or match_tags() calls together.

Examples:

import fiftyone as fo

dataset = fo.load_dataset(...)

#
# Only include samples that have either the "test" or "validation"
# tag
#

view = dataset.match_tags(["test", "validation"])

Parameters

tags – an iterable of tags

Returns

a fiftyone.core.view.DatasetView

mongo(pipeline)

Adds a view stage defined by a raw MongoDB aggregation pipeline.

See MongoDB aggregation pipelines for more details.

Examples:

import fiftyone as fo

dataset = fo.load_dataset(...)

#
# Extract a view containing the 6th through 15th samples in the
# dataset
#

view = dataset.mongo([{"$skip": 5}, {"$limit": 10}])

#
# Sort by the number of detections in the `precictions` field of
# the samples (assume it is a `Detections` field)
#

view = dataset.mongo([
    {
        "$addFields": {
            "_sort_field": {
                "$size": {
                    "$ifNull": ["$predictions.detections", []]
                }
            }
        }
    },
    {"$sort": {"_sort_field": -1}},
    {"$unset": "_sort_field"}
])

Parameters

pipeline – a MongoDB aggregation pipeline (list of dicts)

Returns

a fiftyone.core.view.DatasetView

select(sample_ids)

Returns a view containing only the samples with the given IDs.

Examples:

import fiftyone as fo

dataset = fo.load_dataset(...)

#
# Select the samples with the given IDs from the dataset
#

view = dataset.select([
    "5f3c298768fd4d3baf422d34",
    "5f3c298768fd4d3baf422d35",
    "5f3c298768fd4d3baf422d36",
])

#
# Create a view containing the currently selected samples in the
# App
#

session = fo.launch_app(dataset=dataset)

# Select samples in the App...

view = dataset.select(session.selected)

Parameters

sample_ids – a sample ID or iterable of sample IDs

Returns

a fiftyone.core.view.DatasetView

select_fields(field_names=None)

Selects the fields with the given names as the only fields present in the returned fiftyone.core.sample.SampleView instances. All other fields are excluded.

Note that default sample fields are always selected and will be added if not included in field_names.

Examples:

import fiftyone as fo

dataset = fo.load_dataset(...)

#
# Include only the default fields on each sample
#

view = dataset.select_fields()

#
# Include only the `ground_truth` field (and the default fields) on
# each sample
#

view = dataset.select_fields("ground_truth")

Parameters

field_names (None) – a field name or iterable of field names to select. If not specified, just the default fields will be selected

Returns

a DatasetView

shuffle(seed=None)

Randomly shuffles the samples in the collection.

Examples:

import fiftyone as fo

dataset = fo.load_dataset(...)

#
# Return a view that contains a randomly shuffled version of the
# samples in the dataset
#

view = dataset.shuffle()

#
# Shuffle the samples with a set random seed
#

view = dataset.shuffle(seed=51)

Parameters

seed (None) – an optional random seed to use when shuffling the samples

Returns

a fiftyone.core.view.DatasetView

skip(skip)

Omits the given number of samples from the head of the collection.

Examples:

import fiftyone as fo

dataset = fo.load_dataset(...)

#
# Omit the first 10 samples from the dataset
#

view = dataset.skip(10)

Parameters

skip – the number of samples to skip. If a non-positive number is provided, no samples are omitted

Returns

a fiftyone.core.view.DatasetView

sort_by(field_or_expr, reverse=False)

Sorts the samples in the collection by the given field or expression.

When sorting by an expression, field_or_expr can either be a fiftyone.core.expressions.ViewExpression or a MongoDB expression that defines the quantity to sort by.

Examples:

import fiftyone as fo
from fiftyone import ViewField as F

dataset = fo.load_dataset(...)

#
# Sorts the samples in descending order by the `confidence` of
# their `predictions` field (assume it is a `Classification` field)
#

view = dataset.sort_by("predictions.confidence", reverse=True)

#
# Sorts the samples in ascending order by the number of detections
# in their `predictions` field (assume it is a `Detections` field)
# whose bounding box area is at most 0.2
#

# bbox is in [top-left-x, top-left-y, width, height] format
pred_bbox = F("predictions.detections.bounding_box")
pred_bbox_area = pred_bbox[2] * pred_bbox[3]

view = dataset.sort_by((pred_bbox_area < 0.2).length())

Parameters

• field_or_expr – the field or expression to sort by

• reverse (False) – whether to return the results in descending order

Returns

a fiftyone.core.view.DatasetView

tail(num_samples=3)

Returns a list of the last few samples in the collection.

If fewer than num_samples samples are in the collection, only the available samples are returned.

Parameters

num_samples (3) – the number of samples

Returns

a list of fiftyone.core.sample.Sample objects

take(size, seed=None)

Randomly samples the given number of samples from the collection.

Examples:

import fiftyone as fo

dataset = fo.load_dataset(...)

#
# Take 10 random samples from the dataset
#

view = dataset.take(10)

#
# Take 10 random samples from the dataset with a set seed
#

view = dataset.take(10, seed=51)

Parameters

• size – the number of samples to return. If a non-positive number is provided, an empty view is returned

• seed (None) – an optional random seed to use when selecting the samples

Returns

a fiftyone.core.view.DatasetView

to_json(rel_dir=None, pretty_print=False)

Returns a JSON string representation of the collection.

The samples will be written as a list in a top-level samples field of the returned dictionary.

Parameters

• rel_dir (None) – a relative directory to remove from the filepath of each sample, if possible. The path is converted to an absolute path (if necessary) via os.path.abspath(os.path.expanduser(rel_dir)). The typical use case for this argument is that your source data lives in a single directory and you wish to serialize relative, rather than absolute, paths to the data within that directory

• pretty_print (False) – whether to render the JSON in human readable format with newlines and indentations

Returns

a JSON string

validate_field_type(field_name, ftype, embedded_doc_type=None, subfield=None)

Validates that the collection has a field of the given type.

Parameters

• field_name – the field name

• ftype – the expected field type. Must be a subclass of fiftyone.core.fields.Field

• embedded_doc_type (None) – the fiftyone.core.odm.BaseEmbeddedDocument type of the field. Used only when ftype is an embedded fiftyone.core.fields.EmbeddedDocumentField

• subfield (None) – the type of the contained field. Used only when ftype is a fiftyone.core.fields.ListField or fiftyone.core.fields.DictField

Raises

ValueError – if the field does not exist or does not have the expected type

validate_fields_exist(field_or_fields)

Validates that the collection has fields with the given names.

If field_or_fields contains an embedded field name such as field_name.document.field, only the root field_name is checked for existence.

Parameters

field_or_fields – a field name or iterable of field names

Raises

ValueError – if one or more of the fields do not exist

write_json(json_path, rel_dir=None, pretty_print=False)

Writes the colllection to disk in JSON format.

Parameters

• json_path – the path to write the JSON

• rel_dir (None) – a relative directory to remove from the filepath of each sample, if possible. The path is converted to an absolute path (if necessary) via os.path.abspath(os.path.expanduser(rel_dir)). The typical use case for this argument is that your source data lives in a single directory and you wish to serialize relative, rather than absolute, paths to the data within that directory

• pretty_print (False) – whether to render the JSON in human readable format with newlines and indentations