hello.fiftyone.core module

hello.fiftyone.core.add_sample_field(dataset, field_name, ftype)[source]

Adds a new sample field or embedded field to the dataset, if necessary.

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • field_name – the field name or embedded.field.name

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

hello.fiftyone.core.change_tag(dataset, files, data_path='data', add=None, rm=None)[source]

Change the tags of some samples in this dataset.

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • files (list) – a list of archived dataset file

  • data_path (str, optional) – defaults to “data”

  • add (str, optional) – defaults to None

  • rm (str, optional) – defaults to None

hello.fiftyone.core.clear_sample_field(dataset, field_name)[source]

Clears the values of the field from all samples in the dataset.

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • field_name – the field name or embedded.field.name

hello.fiftyone.core.clone_sample_field(dataset, field_name, new_field_name)[source]

Clones the given sample field into a new field of the dataset.

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • field_name – the field name or embedded.field.name

  • new_field_name – the new field name or embedded.field.name

hello.fiftyone.core.count_values(dataset, field_or_expr, sort_by='label')[source]
hello.fiftyone.core.delete_sample_field(dataset, field_name, error_level=0)[source]

Deletes the field from all samples in the dataset.

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • field_name – the field name or embedded.field.name

  • error_level (int, optional) – the error level to use

hello.fiftyone.core.filter_segmentation_samples(out_dir, data_root, classes, mask_targets, threshold=0.05, splits=['train', 'val'], img_dir='data', ann_dir='labels', img_suffix='.jpg', seg_map_suffix='.png')[source]

Filter samples, based on area of interest ratio.

>>> <data_root>/
>>> ├── objectInfo150.txt
>>> ├── sceneCategories.txt
>>> ├── train
>>> │   ├── data
>>> │   └── labels
>>> └── val
>>>     ├── data
>>>     └── labels
Parameters:

  • out_dir (str) – _description_

  • data_root (str) – _description_

  • classes (list[str]) – _description_

  • mask_targets (dict[int, str]) – _description_

  • threshold (float, optional) – _description_. Defaults to 0.05

  • splits (list, optional) – _description_. Defaults to [“train”, “val”]

  • img_dir (str, optional) – _description_. Defaults to “data”

  • ann_dir (str, optional) – _description_. Defaults to “labels”

  • img_suffix (str, optional) – _description_. Defaults to “.jpg”

  • seg_map_suffix (str, optional) – _description_. Defaults to “.png”

Returns:

a str

hello.fiftyone.core.gen_detections_mapping(old_classes, new_classes)[source]

Generate detections mapping.

Parameters:

  • old_classes (list) – ['c0', 'c1', 'c2', 'c3', 'c4', 'c5']

  • new_classes (list) – ['c0', 'c1', 'c2', ['c3', 'c4', 'c5']]

Returns:

a dict

hello.fiftyone.core.gen_segmentation_mapping(old_classes, new_classes)[source]

Generate segmentation mapping.

Parameters:

  • old_classes (list) – ['c0', 'c1', 'c2', 'c3', 'c4', 'c5', 'be ignored']

  • new_classes (list) – ['c0', 'c1', 'c2', ['c3', 'c4', 'c5'], 'be ignored']

Returns:

a dict

hello.fiftyone.core.has_sample_field(dataset, field_name)[source]

Determines whether the collection has a sample field with the given name.

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • field_name – the field name

Returns:

True/False

hello.fiftyone.core.map_labels(dataset, mapping, field_name='ground_truth')[source]
hello.fiftyone.core.merge_datasets(dataset, others, in_field=None, out_field=None, **kwargs)[source]

Merges the given samples into this dataset.

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • others – a list of fiftyone.core.dataset.Dataset

  • in_field (str) – the name of the input label field

  • out_field (str) – the name of the output label field, which will be created if necessary

  • **kwargs – optional keyword arguments to pass to merge_samples() <https://voxel51.com/docs/fiftyone/api/fiftyone.core.dataset.html#fiftyone.core.dataset.Dataset.merge_samples>

hello.fiftyone.core.merge_labels(dataset, in_field, out_field)[source]

Merges the labels from the given input field into the given output field of the collection.

If this collection is a dataset, the input field is deleted after the merge.

If this collection is a view, the input field will still exist on the underlying dataset but will only contain the labels not present in this view.

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • in_field (str) – the name of the input label field

  • out_field (str) – the name of the output label field, which will be created if necessary

hello.fiftyone.core.merge_samples(datasets, **kwargs)[source]
hello.fiftyone.core.random_split(dataset, splits=None, seed=51)[source]

Adds the split tags to all samples in this dataset.

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • splits (dict, optional) – defaults to None

  • seed (int, optional) – defaults to 51

Returns:

a DatasetView

hello.fiftyone.core.remap_detections_dataset(dataset, new_classes=None, field_name='ground_truth', background=None, least_one=False)[source]

Steps: map labels -> check dataset.classes -> filter valid samples

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • new_classes (None) – refer to gen_label_mapping()

  • field_name (str, optional) – defaults to “ground_truth”

  • background (str, optional) – defaults to None

Returns:

a fiftyone.core.dataset.Dataset

hello.fiftyone.core.remap_segmentation_dataset(dataset, new_classes=None, field_name='ground_truth', ignore_index=255, least_one=False)[source]

Steps: map labels -> check dataset.mask_targets -> filter valid samples

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • new_classes (None) – refer to gen_mask_mapping()

  • field_name (str, optional) – defaults to “ground_truth”

  • ignore_index (int, optional) – defaults to 255

Returns:

a fiftyone.core.dataset.Dataset

hello.fiftyone.core.rename_sample_field(dataset, field_name, new_field_name)[source]

Renames the sample field to the given new name.

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • field_name – the field name or embedded.field.name

  • new_field_name – the new field name or embedded.field.name

hello.fiftyone.core.save_tags(dataset, out_file)[source]
hello.fiftyone.core.split_dataset(dataset, splits=None, limit=3000, seed=51, field_name='ground_truth', from_field=None)[source]

Adds the split tags to all samples in this dataset.

Parameters:

  • dataset – a fiftyone.core.dataset.Dataset

  • splits (dict, optional) – defaults to None

  • limit (int, optional) – defaults to 3000

  • seed (int, optional) – defaults to 51

  • field_name (str, optional) – defaults to “ground_truth”

  • from_field (str, optional) – defaults to None

Returns:

a DatasetView

hello.fiftyone.core.tag_from(dataset, by_dir=None, by_json=None, by_dict=None, by_list=None, is_in=None)[source]
hello.fiftyone.core.tag_from_text(dataset, text_file, tag_map='synsets.txt', remove_prefix=False)[source]

Tag dataset from a text file(cvat export: ImageNet 1.0).

Examples:

000471_0622023132018.jpg 0
000472_0622023132018.jpg 1
000473_0622023132019.jpg 1
000474_0622023132019.jpg 0 1
000475_0622023132019.jpg 0
000476_0622023132019.jpg 0
hello.fiftyone.core.update_dataset_default(dataset, classes, background=None, ignore_index=255)[source]