Datasets¶

Datasets ¶

Interact with datasets in the Driverless AI server.

create ¶

create(
    data: str | DataFrame,
    data_source: str = "upload",
    data_source_config: dict[str, str] = None,
    force: bool = False,
    name: str = None,
    description: str | None = None,
) -> Dataset

Creates a dataset in the Driverless AI server.

Parameters:

data (str | DataFrame) –

Path to the data file(s), or path to a directory, or a SQL query, or a pandas Dataframe.
data_source (str, default: 'upload' ) –

Name of the connector to import data from. Use driverlessai.connectors.list() to get enabled connectors in the server.
data_source_config (dict[str, str], default: None ) –

A dictionary of configuration options for advanced connectors.
force (bool, default: False ) –

Whether to create the dataset even when a dataset already exists with the same name.
name (str, default: None ) –

Name for the created dataset.
description (str | None, default: None ) –

Description for the created dataset. (only available from Driverless AI version 1.10.7 onwards)

Returns:

Dataset –

Created dataset.

Example

dataset = client.datasets.create(
    data="s3://h2o-public-test-data/smalldata/iris/iris.csv",
    data_source="s3",
    name="iris-data",
    description="My Iris dataset",
)

create_async ¶

create_async(
    data: str | DataFrame,
    data_source: str = "upload",
    data_source_config: dict[str, str] = None,
    force: bool = False,
    name: str = None,
    description: str | None = None,
) -> DatasetJob

Launches the creation of a dataset in the Driverless AI server.

Parameters:

data (str | DataFrame) –

Path to the data file(s), or path to a directory, or a SQL query, or a pandas Dataframe.
data_source (str, default: 'upload' ) –

Name of the connector to import data from. Use driverlessai.connectors.list() to get enabled connectors in the server.
data_source_config (dict[str, str], default: None ) –

A dictionary of configuration options for advanced connectors.
force (bool, default: False ) –

Whether to create the dataset even when a dataset already exists with the same name.
name (str, default: None ) –

Name for the created dataset.
description (str | None, default: None ) –

Description for the created dataset. (only available from Driverless AI version 1.10.7 onwards)

Returns:

DatasetJob –

Started the dataset job.

Example

dataset = client.datasets.create(
    data="SELECT * FROM creditcard",
    data_source="jdbc",
    data_source_config=dict(
        jdbc_jar="/opt/jdbc-drivers/mysql/mysql-connector-java-8.0.23.jar",
        jdbc_driver="com.mysql.cj.jdbc.Driver",
        jdbc_url="jdbc:mysql://localhost:3306/datasets",
        jdbc_username="root",
        jdbc_password="root"
    ),
    name="creditcard",
    description="Sample creditcard data",
    force=True,
)

get ¶

get(key: str) -> Dataset

Retrieves a dataset in the Driverless AI server. If the dataset only exists in H2O Storage then it will be imported into the server first.

Parameters:

key (str) –

The unique ID of the dataset.

Returns:

Dataset –

The dataset corresponding to the key.

Example

key = "e7de8630-dbfb-11ea-9f69-0242ac110002"
dataset = client.datasets.get(key=key)

get_by_name ¶

get_by_name(name: str) -> Dataset | None

Retrieves a dataset by its display name from the Driverless AI server.

Parameters:

name (str) –

Name of the dataset.

Returns:

Dataset | None –

The dataset with the specified name if it exists, otherwise None.

Beta API

A beta API that is subject to future changes.

gui ¶

gui() -> Hyperlink

Returns the full URL to the Datasets page in the Driverless AI server.

Returns:

Hyperlink –

The full URL to the Datasets page.

list ¶

list(start_index: int = 0, count: int = None) -> Sequence[Dataset]

Retrieves datasets in the Driverless AI server.

Parameters:

start_index (int, default: 0 ) –

The index of the first dataset to retrieve.
count (int, default: None ) –

The maximum number of datasets to retrieve. If None, retrieves all available datasets.

Returns:

Sequence[Dataset] –

Datasets.

DatasetJob ¶

Monitor the creation of a dataset in the Driverless AI server.

key `property` ¶

key: str

Universally unique key of the entity.

Returns:

str –

name `property` ¶

name: str

Name of the entity.

Returns:

str –

is_complete ¶

is_complete() -> bool

Whether the job has been completed successfully.

Returns:

bool –

True if the job has been completed successfully, otherwise False.

is_running ¶

is_running() -> bool

Whether the job has been scheduled or is running, finishing, or syncing.

Returns:

bool –

True if the job has not completed yet, otherwise False.

result ¶

result(silent: bool = False) -> Dataset

Awaits the job's completion before returning the created dataset.

Parameters:

silent (bool, default: False ) –

Whether to display status updates or not.

Returns:

Dataset –

Created dataset by the job.

status ¶

status(verbose: int = 0) -> str

Returns the status of the job.

Parameters:

verbose (int, default: 0 ) –
- 0: A short description.
- 1: A short description with a progress percentage.
- 2: A detailed description with a progress percentage.

Returns:

str –

Current status of the job.

Dataset ¶

A dataset in the Driverless AI server.

columns `property` ¶

columns: list[str]

Column names of the dataset.

Returns:

list[str] –

creation_timestamp `property` ¶

creation_timestamp: float

Creation timestamp of the dataset in seconds since the epoch (POSIX timestamp).

Returns:

float –

data_source `property` ¶

data_source: str

Original data source of the dataset.

Returns:

str –

description `property` ¶

description: str | None

Description of the dataset.

Returns:

str | None –

file_path `property` ¶

file_path: str

Path to the dataset bin file in the Driverless AI server.

Returns:

str –

file_size `property` ¶

file_size: int

Size in bytes of the dataset bin file in the Driverless AI server.

Returns:

int –

key `property` ¶

key: str

Universally unique key of the entity.

Returns:

str –

log `property` ¶

log: DatasetLog

Log of the dataset.

Returns:

DatasetLog –

name `property` ¶

name: str

Name of the entity.

Returns:

str –

shape `property` ¶

shape: tuple[int, int]

Dimensions of the dataset in (rows, cols) format.

Returns:

tuple[int, int] –

column_summaries ¶

column_summaries(columns: list[str] = None) -> DatasetColumnSummaryCollection

Returns a collection of column summaries.

Parameters:

columns (list[str], default: None ) –

Names of columns to include.

Returns:

DatasetColumnSummaryCollection –

Summaries of the columns of the dataset.

delete ¶

delete() -> None

Permanently deletes the dataset from the Driverless AI server.

download ¶

download(
    dst_dir: str = ".",
    dst_file: str | None = None,
    file_system: AbstractFileSystem | None = None,
    overwrite: bool = False,
    timeout: float = 30,
) -> str

Downloads the dataset as a CSV file.

Parameters:

dst_dir (str, default: '.' ) –

The path where the CSV file will be saved.
dst_file (str | None, default: None ) –

The name of the CSV file (overrides the default file name).
file_system (AbstractFileSystem | None, default: None ) –

FSSPEC-based file system to download to instead of the local file system.
overwrite (bool, default: False ) –

Whether to overwrite or not if a file already exists.
timeout (float, default: 30 ) –

Connection timeout in seconds.

Returns:

str –

Path to the downloaded CSV file.

Example

dataset = client.datasets.create(
    data="s3://h2o-public-test-data/smalldata/parser/avro/weather_snappy-compression.avro",
    data_source="s3",
)
dataset.download()

export ¶

export(**kwargs: Any) -> str

Exports the dataset as a CSV file from the Driverless AI server. Note that the export location is configured in the server. Refer to the Driverless AI docs for more information.

Other Parameters:

storage_destination (str) –

Exporting destination. Possible values are file_system, s3, bitbucket, or azure.
username (str) –

BitBucket username.
password (str) –

BitBucket password.
branch (str) –

BitBucket branch.
user_note (str) –

BitBucket commit message.

Returns:

str –

Relative path to the exported CSV in the export location.

get_used_in_experiments ¶

get_used_in_experiments() -> dict[str, list[Experiment]]

Retrieves the completed experiments where the dataset has been used as the training, testing, or validation dataset.

Returns:

dict[str, list[Experiment]] –

A dictionary with three keys, train, test, and validation, each containing a list of experiments.

Driverless AI version requirement

Requires Driverless AI server 1.10.6 or higher.

gui ¶

gui() -> Hyperlink

Returns the full URL to the dataset details page in the Driverless AI server.

Returns:

Hyperlink –

URL to the dataset details page.

head ¶

head(num_rows: int = 5) -> Table

Returns the column headers and the first n number of rows of the dataset.

Parameters:

num_rows (int, default: 5 ) –

Number of rows to retrieve.

Returns:

Table –

A Table containing the retrieved rows.

Example

dataset = client.datasets.create(
    data="s3://h2o-public-test-data/smalldata/iris/iris.csv",
    data_source="s3",
)
# Print the headers and first 10 rows
print(dataset.head(num_rows=10))

merge_by_rows ¶

merge_by_rows(other_dataset: Dataset, new_dataset_name: str) -> Dataset

Merges the specified dataset into this dataset. Note that the other dataset must have the same columns.

Parameters:

other_dataset (Dataset) –

The dataset that will be merged into this.
new_dataset_name (str) –

Name of the resulting dataset.

Returns:

Dataset –

Merged dataset.

Driverless AI version requirement

Requires Driverless AI server 1.10.6 or higher.

modify_by_code ¶

modify_by_code(code: str, names: list[str] = None) -> dict[str, Dataset]

Creates new dataset(s) by modifying the dataset using a Python script.

In the Python script

The original dataset is available as variable X with type datatable.Frame.
Newly created dataset(s) should be returned as a datatable.Frame, or a pandas.DataFrame, or a numpy.ndarray, or a list of those.

Parameters:

code (str) –

Python script that modifies X.
names (list[str], default: None ) –

Names for the new dataset(s).

Returns:

dict[str, Dataset] –

A dictionary of newly created datasets with names as keys.

Example

# Import the iris dataset.
dataset = client.datasets.create(
    data="s3://h2o-public-test-data/smalldata/iris/iris.csv",
    data_source="s3",
)

# Create a new dataset only with the first 4 columns.
new_dataset = dataset.modify_by_code(
    code="return X[:, :4]",
    names=["new_dataset"],
)

# Split on 5th column to create 2 datasets.
new_datasets = dataset.modify_by_code(
    code="return [X[:, :5], X[:, 5:]]",
    names=["new_dataset_1", "new_dataset_2"],
)

modify_by_code_preview ¶

modify_by_code_preview(code: str) -> Table

Returns a preview of new dataset(s) created by modifying the dataset using a Python script.

In the Python script

The original dataset is available as variable X with type datatable.Frame.
Newly created dataset(s) should be returned as a datatable.Frame, or a pandas.DataFrame, or a numpy.ndarray, or a list of those (only the first dataset in the list is shown in the preview).

Parameters:

code (str) –

Python script that modifies X.

Returns:

Table –

A table containing the first 10 rows of the new dataset.

Example

# Import the iris dataset.
dataset = client.datasets.create(
    data="s3://h2o-public-test-data/smalldata/iris/iris.csv",
    data_source="s3",
)

# A new dataset only with the first 4 columns.
table = dataset.modify_by_code_preview("return X[:, :4]")
print(table)

modify_by_recipe ¶

modify_by_recipe(
    recipe: str | DataRecipe = None, names: list[str] = None
) -> dict[str, Dataset]

Creates new dataset(s) by modifying the dataset using a data recipe.

Parameters:

recipe (str | DataRecipe, default: None ) –

The path to the recipe file, or the url for the recipe, or the data recipe.
names (list[str], default: None ) –

Names for the new dataset(s).

Returns:

dict[str, Dataset] –

A dictionary of newly created datasets with names as keys.

Example

# Import the airlines dataset.
dataset = client.datasets.create(
    data="s3://h2o-public-test-data/smalldata/airlines/allyears2k_headers.zip",
    data_source="s3",
)

# Modify the original dataset with a recipe.
new_datasets = dataset.modify_by_recipe(
    recipe="https://github.com/h2oai/driverlessai-recipes/blob/master/data/airlines_multiple.py",
    names=["new_airlines1", "new_airlines2"],
)

redescribe ¶

redescribe(description: str) -> Dataset

Changes the description of the dataset.

Parameters:

description (str) –

New description.

Returns:

Dataset –

This dataset.

Driverless AI version requirement

Requires Driverless AI server 1.10.7 or higher.

rename ¶

rename(name: str) -> Dataset

Changes the display name of the dataset.

Parameters:

name (str) –

New display name.

Returns:

Dataset –

This dataset.

set_datetime_format ¶

set_datetime_format(columns: dict[str, str]) -> None

Sets/updates the datetime format for columns of the dataset.

Parameters:

columns (dict[str, str]) –

The dictionary where the key is the column name and the value is a valid datetime format.

Example

# Import the Eurodate dataset.
dataset = client.datasets.create(
    data="s3://h2o-public-test-data/smalldata/jira/v-11-eurodate.csv",
    data_source="s3",
)

# Set the date time format for column ‘ds5'
dataset.set_datetime_format({"ds5": "%d-%m-%y %H:%M"})

set_logical_types ¶

set_logical_types(columns: dict[str, str | list[str]]) -> None

Sets/updates the logical data types of the columns of the dataset. The logical type of columns is primarily used to determine which transformers to try on the column's data.

Possible logical types:

categorical
date
datetime
id
numerical
text

Parameters:

columns (dict[str, str | list[str]]) –

A dictionary where the key is the column name and the value is the logical type or a list of logical types for the column Use None to unset all logical types.

Example

# Import the prostate dataset.
dataset = client.datasets.create(
    data="s3://h2o-public-test-data/smalldata/prostate/prostate.csv",
    data_source="s3",
)

# Set the logical types
dataset.set_logical_types(
    {"ID": "id", "AGE": ["categorical", "numerical"], "RACE": None}
)

split_to_train_test ¶

split_to_train_test(
    train_size: float = 0.5,
    train_name: str = None,
    test_name: str = None,
    target_column: str = None,
    fold_column: str = None,
    time_column: str = None,
    seed: int = 1234,
) -> dict[str, Dataset]

Splits the dataset into train and test datasets in the Driverless AI server.

Parameters:

train_size (float, default: 0.5 ) –

Proportion of the rows to put to the train dataset. Rest will be in the test dataset.
train_name (str, default: None ) –

Name for the train dataset.
test_name (str, default: None ) –

Name for the test dataset.
target_column (str, default: None ) –

Column to use for stratified sampling.
fold_column (str, default: None ) –

Column to ensure grouped splitting.
time_column (str, default: None ) –

Column for time-based splitting.
seed (int, default: 1234 ) –

A random seed for reproducibility.

Note

Only one of target_column, fold_column, or time_column can be passed at a time.

Returns:

dict[str, Dataset] –

A dictionary with keys train_dataset and test_dataset, containing the respective dataset.

Example

# Import the iris dataset.
dataset = client.datasets.create(
    data="s3://h2o-public-test-data/smalldata/iris/iris.csv",
    data_source="s3",
)

# Split the iris dataset into train/test sets.
splits = dataset.split_to_train_test(train_size=0.7)
train_dataset = splits["train_dataset"]
test_dataset = splits["test_dataset"]

split_to_train_test_async ¶

split_to_train_test_async(
    train_size: float = 0.5,
    train_name: str = None,
    test_name: str = None,
    target_column: str = None,
    fold_column: str = None,
    time_column: str = None,
    seed: int = 1234,
) -> DatasetSplitJob

Launches the splitting of the dataset into train and test datasets in the Driverless AI server.

Parameters:

train_size (float, default: 0.5 ) –

Proportion of the rows to put to the train dataset. Rest will be in the test dataset.
train_name (str, default: None ) –

Name for the train dataset.
test_name (str, default: None ) –

Name for the test dataset.
target_column (str, default: None ) –

Column to use for stratified sampling.
fold_column (str, default: None ) –

Column to ensure grouped splitting.
time_column (str, default: None ) –

Column for time-based splitting.
seed (int, default: 1234 ) –

A random seed for reproducibility.

Note

Only one of target_column, fold_column, or time_column can be passed at a time.

Returns:

DatasetSplitJob –

Started dataset split job.

summarize ¶

summarize() -> DatasetSummary

Summarizes the dataset using a GPT configured in the Driverless AI server.

Returns:

DatasetSummary –

Dataset summary.

Driverless AI version requirement

Requires Driverless AI server 1.10.6 or higher.

Beta API

A beta API that is subject to future changes.

summarize_async ¶

summarize_async() -> DatasetSummarizeJob

Launches the summarization of the dataset using a GPT configured in the Driverless AI server.

Returns:

DatasetSummarizeJob –

Started summarization job.

Driverless AI version requirement

Requires Driverless AI server 1.10.6 or higher.

Beta API

A beta API that is subject to future changes.

tail ¶

tail(num_rows: int = 5) -> Table

Returns the column headers and the last n number of rows of the dataset.

Parameters:

num_rows (int, default: 5 ) –

Number of rows to retrieve.

Returns:

Table –

A Table containing the retrieved rows.

Example

dataset = client.datasets.create(
    data="s3://h2o-public-test-data/smalldata/iris/iris.csv",
    data_source="s3",
)
# Print the headers and last 10 rows
print(dataset.tail(num_rows=10))

DatasetLog ¶

A dataset log file in the Driverless AI server.

file_name `property` ¶

file_name: str

Filename of the log file.

Returns:

str –

download ¶

download(
    dst_dir: str = ".",
    dst_file: str | None = None,
    file_system: AbstractFileSystem | None = None,
    overwrite: bool = False,
    timeout: float = 30,
) -> str

Downloads the log file.

Parameters:

dst_dir (str, default: '.' ) –

The path where the log file will be saved.
dst_file (str | None, default: None ) –

The name of the log file (overrides the default file name).
file_system (AbstractFileSystem | None, default: None ) –

FSSPEC-based file system to download to instead of the local file system.
overwrite (bool, default: False ) –

Whether to overwrite or not if a file already exists.
timeout (float, default: 30 ) –

Connection timeout in seconds.

Returns:

str –

Path to the downloaded log file.

head ¶

head(num_lines: int = 50) -> str

Returns the first n lines of the log file.

Parameters:

num_lines (int, default: 50 ) –

Number of lines to retrieve.

tail ¶

tail(num_lines: int = 50) -> str

Returns the last n lines of the log file.

Parameters:

num_lines (int, default: 50 ) –

Number of lines to retrieve.

DatasetColumnSummaryCollection ¶

A collection of column summaries of a dataset.

A column summary can be retrieved,

by the column index dataset.column_summaries()[0]
by the column name dataset.column_summaries()["C1"]
Or slice it to get multiple summaries dataset.column_summaries()[0:3]

DatasetColumnSummary ¶

Summary of a column in a dataset.

Example

c1_summary = dataset.column_summaries()["C1"]
# Print the summary for a histogram along with column statistics.
print(c1_summary)

Sample output:

--- C1 ---

 4.3|███████
    |█████████████████
    |██████████
    |████████████████████
    |████████████
    |███████████████████
    |█████████████
    |████
    |████
 7.9|████

Data Type: real
Logical Types: ['categorical', 'numerical']
Datetime Format:
Count: 150
Missing: 0
Mean: 5.84
SD: 0.828
Min: 4.3
Max: 7.9
Unique: 35
Freq: 10

count `property` ¶

count: int

Non-missing values count in the column.

Returns:

int –

data_type `property` ¶

data_type: str

Raw data type of the column as detected by the Driverless AI server when the dataset was imported.

Returns:

str –

datetime_format `property` ¶

datetime_format: str

Datetime format of the column. See also Dataset.set_datetime_format.

Returns:

str –

freq `property` ¶

freq: int

Count of most frequent value in the column.

Returns:

int –

logical_types `property` ¶

logical_types: list[str]

User defined data types for the column to be used by Driverless AI server. This precedes the data_type. See also Dataset.set_logical_types.

Returns:

list[str] –

max `property` ¶

max: bool | float | int | None

Maximum value in the column if it contains binary/numeric data.

Returns:

bool | float | int | None –

mean `property` ¶

mean: float | None

Mean value of the column if it contains binary/numeric data.

Returns:

float | None –

min `property` ¶

min: bool | float | int | None

Minimum value in the column if it contains binary/numeric data.

Returns:

bool | float | int | None –

missing `property` ¶

missing: int

Missing values count in the column.

Returns:

int –

name `property` ¶

name: str

Column name.

Returns:

str –

sd `property` ¶

sd: float | None

Standard deviation of the column if it contains binary/numeric data.

Returns:

float | None –

unique `property` ¶

unique: int

Unique values count of the column.

Returns:

int –

DatasetSplitJob ¶

Monitor splitting of a dataset in the Driverless AI server.

key `property` ¶

key: str

Universally unique key of the entity.

Returns:

str –

name `property` ¶

name: str

Name of the entity.

Returns:

str –

is_complete ¶

is_complete() -> bool

Whether the job has been completed successfully.

Returns:

bool –

True if the job has been completed successfully, otherwise False.

is_running ¶

is_running() -> bool

Whether the job has been scheduled or is running, finishing, or syncing.

Returns:

bool –

True if the job has not completed yet, otherwise False.

result ¶

result(silent: bool = False) -> dict[str, Dataset]

Awaits the job's completion before returning the split datasets.

Parameters:

silent (bool, default: False ) –

Whether to display status updates or not.

Returns:

dict[str, Dataset] –

A dictionary with keys train_dataset and test_dataset, containing the respective dataset created by the job.

status ¶

status(verbose: int = None) -> str

Returns the status of the job.

Parameters:

verbose (int, default: None ) –

Ignored.

Returns:

str –

Current status of the job.

DatasetSummarizeJob ¶

Monitor the creation of a dataset summary in the Driverless AI server.

key `property` ¶

key: str

Universally unique key of the entity.

Returns:

str –

name `property` ¶

name: str

Name of the entity.

Returns:

str –

is_complete ¶

is_complete() -> bool

Whether the job has been completed successfully.

Returns:

bool –

True if the job has been completed successfully, otherwise False.

is_running ¶

is_running() -> bool

Whether the job has been scheduled or is running, finishing, or syncing.

Returns:

bool –

True if the job has not completed yet, otherwise False.

result ¶

result(silent: bool = False) -> DatasetSummary

Awaits the job's completion before returning the created dataset summary.

Parameters:

silent (bool, default: False ) –

Whether to display status updates or not.

Returns:

DatasetSummary –

Created dataset summary by the job.

status ¶

status(verbose: int = 0) -> str

Returns the status of the job.

Parameters:

verbose (int, default: 0 ) –
- 0: A short description.
- 1: A short description with a progress percentage.
- 2: A detailed description with a progress percentage.

Returns:

str –

Current status of the job.

DatasetSummary `dataclass` ¶

A summary of a dataset.

provider `instance-attribute` ¶

provider: str

GPT provider that generated the dataset summary.

Returns:

str –

summary `instance-attribute` ¶

summary: str

Dataset summary.

Returns:

str –

Datasets¶

Datasets ¶

create ¶

create_async ¶

get ¶

get_by_name ¶

gui ¶

list ¶

DatasetJob ¶

key property ¶

name property ¶

is_complete ¶

is_running ¶

result ¶

status ¶

Dataset ¶

columns property ¶

creation_timestamp property ¶

data_source property ¶

description property ¶

file_path property ¶

file_size property ¶

key property ¶

log property ¶

name property ¶

shape property ¶

column_summaries ¶

delete ¶

download ¶

export ¶

get_used_in_experiments ¶

gui ¶

head ¶

merge_by_rows ¶

modify_by_code ¶

modify_by_code_preview ¶

modify_by_recipe ¶

redescribe ¶

rename ¶

set_datetime_format ¶

set_logical_types ¶

split_to_train_test ¶

split_to_train_test_async ¶

summarize ¶

summarize_async ¶

tail ¶

DatasetLog ¶

file_name property ¶

download ¶

head ¶

tail ¶

DatasetColumnSummaryCollection ¶

DatasetColumnSummary ¶

count property ¶

data_type property ¶

datetime_format property ¶

freq property ¶

logical_types property ¶

max property ¶

mean property ¶

min property ¶

missing property ¶

name property ¶

sd property ¶

unique property ¶

DatasetSplitJob ¶

key property ¶

name property ¶

is_complete ¶

is_running ¶

result ¶

status ¶

DatasetSummarizeJob ¶

key property ¶

name property ¶

is_complete ¶

is_running ¶

result ¶

status ¶

DatasetSummary dataclass ¶

key `property` ¶

name `property` ¶

columns `property` ¶

creation_timestamp `property` ¶

data_source `property` ¶

description `property` ¶

file_path `property` ¶

file_size `property` ¶

key `property` ¶

log `property` ¶

name `property` ¶

shape `property` ¶

file_name `property` ¶

count `property` ¶

data_type `property` ¶

datetime_format `property` ¶

freq `property` ¶

logical_types `property` ¶

max `property` ¶

mean `property` ¶

min `property` ¶

missing `property` ¶

name `property` ¶

sd `property` ¶

unique `property` ¶

key `property` ¶

name `property` ¶

key `property` ¶

name `property` ¶

DatasetSummary `dataclass` ¶

provider `instance-attribute` ¶

summary `instance-attribute` ¶