packflow.utils.normalize package

Submodules

packflow.utils.normalize.base module

class packflow.utils.normalize.base.TypeConversionHandler[source]

Bases: ABC

available()[source]

Returns if the type conversion handler is available.

Return type:

bool

abstract convert(obj)[source]

Logic to convert the type to a native, json-serializable Python type.

Parameters:

obj (Any) – The object to convert

Returns:

Any python native type that is JSON serializable

Return type:

object

abstract is_type(obj)[source]

Check if the object is the proper type for the converter

Return type:

bool

abstract property package_name: str

The name of the required package to attempt to import

packflow.utils.normalize.handlers module

class packflow.utils.normalize.handlers.NumpyTypeHandler[source]

Bases: TypeConversionHandler

convert(obj)[source]

Logic to convert the type to a native, json-serializable Python type.

Parameters:

obj (Any) – The object to convert

Returns:

Any python native type that is JSON serializable

Return type:

object

is_type(obj)[source]

Check if the object is the proper type for the converter

Return type:

bool

property package_name: str

The name of the required package to attempt to import

class packflow.utils.normalize.handlers.PandasDataFrameHandler[source]

Bases: TypeConversionHandler

convert(obj)[source]

Logic to convert the type to a native, json-serializable Python type.

Parameters:

obj (Any) – The object to convert

Returns:

Any python native type that is JSON serializable

Return type:

object

is_type(obj)[source]

Check if the object is the proper type for the converter

Return type:

bool

property package_name: str

The name of the required package to attempt to import

class packflow.utils.normalize.handlers.PandasSeriesHandler[source]

Bases: TypeConversionHandler

convert(obj)[source]

Logic to convert the type to a native, json-serializable Python type.

Parameters:

obj (Any) – The object to convert

Returns:

Any python native type that is JSON serializable

Return type:

object

is_type(obj)[source]

Check if the object is the proper type for the converter

Return type:

bool

property package_name: str

The name of the required package to attempt to import

class packflow.utils.normalize.handlers.PillowImageHandler[source]

Bases: TypeConversionHandler

convert(obj)[source]

Logic to convert the type to a native, json-serializable Python type.

Parameters:

obj (Any) – The object to convert

Returns:

Any python native type that is JSON serializable

Return type:

object

is_type(obj)[source]

Check if the object is the proper type for the converter

Return type:

bool

property package_name: str

The name of the required package to attempt to import

class packflow.utils.normalize.handlers.TorchScalarHandler[source]

Bases: TypeConversionHandler

convert(obj)[source]

Logic to convert the type to a native, json-serializable Python type.

Parameters:

obj (Any) – The object to convert

Returns:

Any python native type that is JSON serializable

Return type:

object

is_type(obj)[source]

Check if the object is the proper type for the converter

Return type:

bool

property package_name: str

The name of the required package to attempt to import

class packflow.utils.normalize.handlers.TorchTensorHandler[source]

Bases: TypeConversionHandler

convert(obj)[source]

Logic to convert the type to a native, json-serializable Python type.

Parameters:

obj (Any) – The object to convert

Returns:

Any python native type that is JSON serializable

Return type:

object

is_type(obj)[source]

Check if the object is the proper type for the converter

Return type:

bool

property package_name: str

The name of the required package to attempt to import

packflow.utils.normalize.normalize module

packflow.utils.normalize.normalize.ensure_native_types(obj)[source]

Converts any non-native data types to ensure JSON serialization is possible.

Parameters:

obj (Any) – the object to convert to a native data type

Return type:

A scalar or object with all-native Python data types

packflow.utils.normalize.normalize.ensure_valid_output(output, parent_key='output')[source]

Coerces outputs to native Python types and correct formats.

Parameters:
  • output (Iterable[Any]) – Model outputs to be cleaned

  • parent_key (str) – Default ‘output’. Used when the output is not a dictionary type. Ex: Output of [0, 1] would be wrapped to {f’{parent_key}’: [0, 1]}

Return type:

List[Dict[str, Any]]

Module contents