tableserializer package

Subpackages

• tableserializer.serializer module
  • Submodules
  • tableserializer.serializer.metadata module
    • MetadataSerializer
      • MetadataSerializer.serialize_metadata()
    • PairwiseMetadataSerializer
      • PairwiseMetadataSerializer.serialize_metadata()
    • JSONMetadataSerializer
      • JSONMetadataSerializer.serialize_metadata()
  • tableserializer.serializer.schema module
    • SchemaSerializer
      • SchemaSerializer.serialize_schema()
    • ColumnNameSchemaSerializer
      • ColumnNameSchemaSerializer.serialize_schema()
    • SQLSchemaSerializer
      • SQLSchemaSerializer.serialize_schema()
  • tableserializer.serializer.serializer module
    • Serializer
      • Serializer.serialize()
  • tableserializer.serializer.table module
    • RawTableSerializer
      • RawTableSerializer.serialize_raw_table()
    • JSONRawTableSerializer
      • JSONRawTableSerializer.serialize_raw_table()
    • MarkdownRawTableSerializer
      • MarkdownRawTableSerializer.serialize_raw_table()
    • CSVRawTableSerializer
      • CSVRawTableSerializer.serialize_raw_table()
    • LatexRawTableSerializer
      • LatexRawTableSerializer.serialize_raw_table()
  • Module contents
• tableserializer.table module
  • Submodules
  • tableserializer.table.preprocessor module
    • TablePreprocessor
      • TablePreprocessor.process()
    • ColumnDroppingPreprocessor
      • ColumnDroppingPreprocessor.process()
    • StringTruncationPreprocessor
      • StringTruncationPreprocessor.process()
  • tableserializer.table.row_sampler module
    • RowSampler
      • RowSampler.sample()
    • RandomRowSampler
      • RandomRowSampler.sample()
    • FirstRowSampler
      • FirstRowSampler.sample()
    • KMeansRowSampler
      • KMeansRowSampler.sample()
  • tableserializer.table.table module
    • Table
      • Table.as_list_of_lists()
      • Table.as_list_of_dicts()
      • Table.as_dataframe()
  • Module contents

Submodules

tableserializer.recipe module

class tableserializer.recipe.SerializationRecipe(recipe: str, identifier: str = None)

Bases: object

A SerializationRecipe details the structure of the table serialization.

Parameters:

• recipe (str) – String representation of the overall structure of the serialization with placeholders that are dynamically filled in on a per-table basis.

• identifier (Optional[str]) – Optional identifier for the table.

cook_recipe(schema_contents: str = None, metadata_contents: str = None, table_contents: str = None) → str

Fill-in the values for the placeholder values.

Parameters:

• schema_contents (str) – Schema contents to fill in.

• metadata_contents (str) – Metadata contents to fill in.

• table_contents (str) – Table contents to fill in.

Returns:

Table serialization according to the schema.

Return type:

str

get_raw_recipe() → str

Get the raw recipe.

Returns:

The raw recipe as a string.

Return type:

str

get_fields() → List[str]

Get a list of all fields defined in the recipe.

Returns:

List of names of all fields defined in the recipe.

Return type:

List[str]

tableserializer.kitchen module

class tableserializer.kitchen.ExperimentalSerializerKitchen

Bases: object

Central class for managing serialization components and custom extensions for experiments.

register_schema_serializer_class(schema_serializer_class: Type[SchemaSerializer]) → None

Register a custom schema serializer class to the kitchen.

Parameters:

schema_serializer_class (Type[SchemaSerializer]) – Schema serializer class to register.

Return type:

None

register_raw_table_serializer_class(table_serializer_class: Type[RawTableSerializer]) → None

Register a custom raw table serializer class to the kitchen.

Parameters:

table_serializer_class (Type[RawTableSerializer]) – Raw table serializer class to register.

Return type:

None

register_metadata_serializer_class(metadata_serializer_class: Type[MetadataSerializer]) → None

Register a custom metadata serializer class to the kitchen.

Parameters:

metadata_serializer_class – Metadata serializer class to register.

Return type:

None

register_row_sampler_class(row_sampler_class: Type[RowSampler]) → None

Register a custom row sampler class to the kitchen.

Parameters:

row_sampler_class – Row sampler class to register.

Return type:

None

register_table_preprocessor_class(table_preprocessor_class: Type[TablePreprocessor]) → None

Register a custom table preprocessor class to the kitchen.

Parameters:

table_preprocessor_class – Table preprocessor class to register.

Return type:

None

create_schema_serializer(schema_serializer_name: str, **kwargs: Any) → SchemaSerializer

Create a SchemaSerializer for the given schema serializer name. This assumes that a SchemaSerializer with the supplied name is registered.

Parameters:

• schema_serializer_name (str) – Name of the registered schema serializer class that should be instantiated.

• kwargs (Any) – Constructor arguments for instantiating the SchemaSerializer class.

Returns:

SchemaSerializer instance.

Return type:

SchemaSerializer

create_table_serializer(raw_table_serializer_name: str, **kwargs: Any) → RawTableSerializer

Create a RawTableSerializer for the given table serializer name. This assumes that a RawTableSerializer with the supplied name is registered.

Parameters:

• raw_table_serializer_name (str) – Name of the registered RawTableSerializer class that should be instantiated.

• kwargs (Any) – Constructor arguments for instantiating the RawTableSerializer class.

Returns:

RawTableSerializer instance.

Return type:

RawTableSerializer

create_metadata_serializer(metadata_serializer_name: str, **kwargs: Any) → MetadataSerializer

Create a MetadataSerializer for the given metadata serializer name. This assumes that a MetadataSerializer with the supplied name is registered.

Parameters:

• metadata_serializer_name (str) – Name of the registered MetadataSerializer class that should be instantiated.

• kwargs (Any) – Constructor arguments for instantiating the MetadataSerializer class.

Returns:

MetadataSerializer instance.

Return type:

MetadataSerializer

create_row_sampler(row_sampler_name: str, rows_to_sample: int = 10, **kwargs: Any) → RowSampler

Create a RowSampler for the given row sampler name. This assumes that a RowSampler with the supplied name is registered.

Parameters:

• row_sampler_name (str) – Name of the registered RowSampler class that should be instantiated.

• rows_to_sample (int) – Number of rows to sample.

• kwargs (Any) – Constructor arguments for instantiating the RowSampler class.

Returns:

RowSampler instance.

Return type:

RowSampler

create_table_preprocessor(table_preprocessor_name: str, **kwargs: Any) → TablePreprocessor

Create a TablePreprocessor for the given table preprocessor name. This assumes that a TablePreprocessor with the supplied name is registered.

Parameters:

• table_preprocessor_name (str) – Name of the registered TablePreprocessor class that should be instantiated.

• kwargs (Any) – Constructor arguments for instantiating the TablePreprocessor class.

Returns:

TablePreprocessor instance.

Return type:

TablePreprocessor

static jar_up_as_json(serializer: Serializer) → str

Create a JSON representation of the given serializer. This representation captures the full configuration of the serializer, allowing instantiating an equal serializer.

Parameters:

serializer (Serializer) – Serializer to jar up as JSON.

Returns:

JSON representation of the given serializer.

Return type:

str

unjar_from_json(serializer_json: str) → Serializer

Create a serializer instance from a JSON representation of a serializer.

Parameters:

serializer_json (str) – JSON representation of a serializer.

Returns:

Serializer instance.

Return type:

Serializer

create_serializers(recipes: List[SerializationRecipe], metadata_serializers: List[MetadataSerializer], schema_serializers: List[SchemaSerializer], table_serializers: List[RawTableSerializer], row_samplers: List[RowSampler], table_preprocessor_constellations: List[List[TablePreprocessor]]) → List[Serializer]

Create serializers with different parameter configurations.

Parameters:

• recipes (List[SerializationRecipe]) – List of recipes for which to create serializer instances.

• metadata_serializers (List[MetadataSerializer]) – List of metadata serializers for which to create serializer instances.

• schema_serializers (List[SchemaSerializer]) – List of schema serializers for which to create serializer instances.

• table_serializers (List[RawTableSerializer]) – List of raw table serializers for which to create serializer instances.

• row_samplers (List[RowSampler]) – List of row samplers for which to create serializer instances.

• table_preprocessor_constellations (List[List[TablePreprocessor]]) – List of table preprocessor constellations for which to create serializer instances.

Returns:

A list of serializers for all possible combinations of components.

Return type:

List[Serializer]

save_serializer_experiment_configurations(serializers: List[Serializer], base_folder: str) → None

Create a folder structure within a base folder and save configurations of the provided serializers into this structure.

Parameters:

• serializers (List[Serializer]) – Serializers that are saved.

• base_folder (str) – Base folder that configuration files will be saved into.

Returns:

None

Return type:

None

get_serializers_from_dir(base_folder: str) → List[Tuple[str, Serializer]]

Get a list of serializers that have been saved in a base folder.

Parameters:

base_folder – Base folder where

Returns:

A tuple (experiment_folder, )

run_experiments_with_serializers(base_folder: str, experiment_callback: Callable) → None

Provide a callback function and run experiments over all serializers saved in the base folder.

Parameters:

• base_folder (str) – Base folder that serializer configuration files reside in.

• experiment_callback (Callable) – Callback function that is invoked with a tuple (experiment_folder, serializer) (Tuple[str, Serializer]) for each experiment. The callback function is provided the experiment folder as string value and the associated Serializer.

Returns:

None

Return type:

None

Module contents