Ontology SDK
-
Print
-
DarkLight
To reach the ontologies repository press here
To reach the ontologies entity press here
To learn more about Dataloop's recipe & ontology go to What are Recipe & Ontology?
For a UI usage of recipe go to Ontology or Instructions pages.
Highlighted text require your input
Relations
Prep
import dtlpy as dl
if dl.token_expired():
dl.login()
project = dl.projects.get(project_name='project_name')
dataset = project.datasets.get(dataset_name='dataset_name')Add and Edit Ontology
By Dataset Entity
View Dataset's Labels
# as objects labels = dataset.labels # as instance map labels = dataset.instance_map
Add Labels by Dataset
To reach the "add a label by dataset" function full reference, press here.
Please verify that you are not creating labels that already exist.
Add one Label
# Add one label dataset.add_label(label_name='person')
Add Multiple Labels
# Add multiple dataset.add_labels(label_list=['person', 'animal', 'object'])
Add a single label with a specific color
# Add single label with specific color dataset.add_label(label_name='person', color=(34, 6, 231))
Add a single label with a specific color and attributes
# Add single label with specific color and attributes dataset.add_label(label_name='person', color=(34, 6, 231), attributes=['big', 'small'])
Add Labels using Label object
# Create Labels list using Label object
labels = [
dl.Label(tag='Donkey', color=(1, 1, 1)),
dl.Label(tag='Mammoth', color=(34, 56, 7)),
dl.Label(tag='Bird', color=(100, 14, 150))
]
#Add Labels to Dataset
dataset.add_labels(label_list=labels)
# or you can also create a recipe from the label list
recipe = dataset.recipes.create(recipe_name='My-Recipe-name', labels=labels)Add a Label with children and attributes
# Add label with children and attributes
label = dl.Label(
tag='Fish',
attributes=['carnivore', 'herbivores'],
color=(34, 6, 231),
children=[
dl.Label(
tag='Shark',
color=(34, 6, 231),
attributes=['baby', 'old'],
),
dl.Label(
tag='Salmon',
color=(34, 6, 231),
attributes=['pink', 'norwegian'],
)
]
)
dataset.add_labels(label_list=label)Add multiple Labels with children and attributes
# Add multiple labels with children and attributes
# Create Labels list
labels = [ dl.Label(
tag='Fish',
attributes=['carnivore', 'herbivores'],
color=(34, 6, 231),
children=[
dl.Label(
tag='Shark',
color=(34, 6, 231),
attributes=['baby', 'old'],
),
dl.Label(
tag='Salmon',
color=(34, 6, 231),
attributes=['pink', 'norwegian'],
)
]
),
dl.Label(
tag='Meat',
attributes=['carnivore', 'herbivores'],
color=(34, 6, 231),
children=[
dl.Label(
tag='Beef',
color=(34, 6, 231),
attributes=['Rib', 'Brisket'],
),
dl.Label(
tag='Lamb',
color=(34, 6, 231),
attributes=['baby', 'old'],
)
]
)
]
#Add Labels to Dataset
dataset.add_labels(label_list=labels)
# or you can also create a recipe from the label list
recipe = dataset.recipes.create(recipe_name='My-Recipe-name', labels=labels)Add hierarchy labels with nested
Different options for hierarchy label creation.
Option A:
#Option A #add father label labels = dataset.add_label(label_name="animal",color=(123,134,64),attributes=["Farm","Home"]) #add child label labels = dataset.add_label(label_name="animal.Dog",color=(45,34,164),attributes=["Big","Small"]) #add grandchild label labels = dataset.add_label(label_name="animal.Dog.poodle",attributes=["Black","Weight"])
Option B:
#Option B:only if you dont have attributes #parent and grandparent (animal and dog) will be generate automaticly labels = dataset.add_label(label_name="animal.Dog.poodle")
Option C:
#Option C: with the Big Dict
nested_labels = [
{'label_name': 'animal.Dog',
'color': '#220605',
'children': [{'label_name': 'poodle',
'color': '#298345'},
{'label_name': 'labrador',
'color': '#298651'}]},
{'label_name': 'animal.cat',
'color': '#287605',
'children': [{'label_name': 'Persian',
'color': '#298345'},
{'label_name': 'Balinese',
'color': '#298651'}]}
]Add Labels to a dataset:
labels = dataset.add_label(label_name=nested_label)
Create a Recipe From from a Label list
Use Add Labels using Label object and Add multiple labels with children and attributes scripts to understand how to create a labels list using Label object.
#dont forget to create labels list using different scripts recipe = dataset.recipes.create(recipe_name='My-Recipe-name', labels=labels)
Delete Labels by Dataset
dataset.delete_labels(label_names=['Cat', 'Dog'])
Update Label Features
dataset.update_label(label_name='Cat', color="#000080") #update existing label , if not exist fails dataset.update_label(label_name='Cat', color="#fcba03", upsert=True) #update label, if not exist add it
By Ontology Entity
Prep
📘 Please make sure you already run the top page prep script.
# Get recipe from list recipe = dataset.recipes.list()[0] # Or get specific recipe: recipe = dataset.recipes.get(recipe_id='id') # Get ontology from list ontology = recipe.ontologies.list()[0] # Or get specific ontology: ontology = recipe.ontologies.get(ontology_id='id')
Labels
Labels is a python list, as such, you can use python list methods like sort(). Just make sure to use ontology.update() afterwards.
📘 You can also add labels by the ontology entity
ontology.add_labels(label_list=['Shark', 'whale', 'Animal.Donkey'],update_ontology=True)
Child Labels
childLabel = ontology.labels[x].children[y].children[z] #If you want to reach the pug label in the example childLabel = ontology.labels[0].children[0].children[3]
Attributes
❗️ If you wish to add attributes to a child labels well as all the labels he inherits from, use append attribute to an existing label example.
Add attributes to the ontology
🚧 This script is for adding attributes to an ontology that does not already has attributes, if the ontology already has attributes use "Update attributes " instead.
ontology.attributes = ['NewAttribute', 'NewAttribute2'] #Or for child label ontology.labels[x].children[y].attributes = ['NewAttribute', 'NewAttribute2'] ontology.update()
Update attributes of the ontology
1. View your current attributes
ontology.attributes #Or for child label ontology.labels[x].children[y].attributes
Output:
['CurrentFirstAttribute', 'CurrentSecondAttribute']
2. Add new attributes to an existing one
ontology.attributes = ['CurrentFirstAttribute', 'CurrentSecondAttribute', 'NewAttribute', 'NewAttribute2'] #Or for child labels ontology.labels[x].children[y].attributes = ['CurrentFirstAttribute', 'CurrentSecondAttribute', 'NewAttribute', 'NewAttribute2'] ontology.update()
By Annotation Entity
Prep
📘 Please make sure you already run the top page prep script
item = dataset.items.get(filepath='/your-image-file-path.jpg') annotation = item.annotations.get(annotation_id='your-annotation-id-number')
Change annotation label to a different label
To change multiple annotations labels, use Change Annotation's Label To A New Label tutorial.
annotation.label = "label2" annotation.update()
Edit Annotation's Attribute
annotation.attributes = ['leftEye'] annotation = annotation.update()
Link dataset labels to a different dataset
🚧 Please make sure you run the prep script with the dataset you wish to get the labels from.
- Notice that if you delete a recipe it will delete it for all its linked datasets
Link dataset labels to a new dataset
new_dataset = project.datasets.create(dataset_name='new_dataset_with_labels',
labels=dataset.labels)Link dataset labels to an existing dataset
new_dataset = project.datasets.create(dataset_name='new_dataset_without_labels') # Get from a list or recipes recipe = new_dataset.recipes.list()[0] # Or get recipe by id recipe = new_dataset.recipes.get(recipe_id='recipe_id') # Get from the list of ontologies ontology = recipe.ontologies.list()[0] # Or get ontology by id ontology = recipe.ontologies.get(ontology='ontology_id') # Add the labels to the dataset ontology.add_labels(label_list=dataset.labels) ontology.update()
Link dataset ontology to a different dataset
🚧 Please make sure you run the prep script with the dataset you wish to get the labels from.
- Notice that if you delete a recipe it will delete it for all its linked datasets
Link dataset ontology to a new dataset
new_dataset = project.datasets.create(dataset_name='new_dataset_with_ontology',
ontology_ids=dataset.ontology_ids)Link dataset ontology to an existing dataset's recipe
new_dataset = project.datasets.create(dataset_name='new_dataset_without_ontology') # get recipe new_dataset_recipe = new_dataset.recipes.list()[0] # Copy from a different dataset new_dataset_recipe.ontologyIds = dataset.ontology_ids # Update the new dataset new_dataset_recipe.update()
Was This Article Helpful?