pandantic

pandantic

Defines types for Pandas Series and DataFrame serialization and deserialization.

PandasDataFrame module-attribute

PandasDataFrame = Annotated[
    DataFrame, BeforeValidator(to_dataframe), WrapSerializer(from_dataframe)
]

Type alias for Pandas DataFrame objects that can be serialized and deserialized. This type alias is used to define the expected data structure when working with Pandas DataFrame objects in a way that allows them to be both serialized and deserialized using the DataFrameSerDe class. The PandasDataFrame type alias is defined as an Annotated type, which means it has two type arguments: pd.DataFrame, which specifies the expected data structure for this type, and a BeforeValidator and PlainSerializer, respectively. This allows you to use Pandas DataFrame objects in your code while ensuring that they can be serialized and deserialized using the DataFrameSerDe class.

PandasSeries module-attribute

PandasSeries = Annotated[Series, BeforeValidator(to_series), WrapSerializer(from_series)]

Type alias for Pandas Series objects that can be serialized and deserialized. This type alias is used to define the expected data structure when working with Pandas Series objects in a way that allows them to be both serialized and deserialized using the SeriesSerDe class. The PandasSeries type alias is defined as an Annotated type, which means it has two type arguments: pd.Series, which specifies the expected data structure for this type, and a BeforeValidator and PlainSerializer, respectively. This allows you to use Pandas Series objects in your code while ensuring that they can be serialized and deserialized using the SeriesSerDe class.

DataFrameSerDe

Bases: BaseModel

Class to serialize and deserialize Pandas DataFrame objects.

Attributes:

Name	Type	Description
columns	NDArray[Shape['*'], Any]	Column names of the dataframe
values	NDArray[Shape['*, *'], Any]	Values of the dataframe
index	NDArray[Shape['*'], Any]	Index of the dataframe
index_name	int | str | None	Name of the index
columns_name	list[int | str | None]	Names of the columns

Methods: as_dataframe(): Convert to Pandas DataFrame object from_dict(data: dict) -> DataFrameSerDe: Create a new instance from a dictionary model_validate_json(json: str | bytes, loc: Any = None, prop: Any = None, cls: Any = None, **kwargs: Any) -> Self: Validate and deserialize JSON data into the model.

as_dataframe

as_dataframe()

Convert to Pandas DataFrame object.

Returns:

Type	Description
DataFrame	Pandas DataFrame object from stored data

Source code in src/data_handlers/custom_serde_definitions/pandantic.py

def as_dataframe(self):
    """
    Convert to Pandas DataFrame object.

    Returns:
        (pd.DataFrame): Pandas DataFrame object from stored data
    """
    return pd.DataFrame(
        self.values, 
        columns=pd.Index(self.columns, name=self.columns_name), 
        index=pd.Index(self.index, name=self.index_name)
    )

from_dataframe classmethod

from_dataframe(df: DataFrame, nxt: SerializerFunctionWrapHandler, info: FieldSerializationInfo)

Create a new instance from a Pandas DataFrame object.

Parameters:

Name	Type	Description	Default
df	DataFrame	Pandas DataFrame object to convert.	required

Returns:

Type	Description
cls	New instance created from the Pandas DataFrame object.

Source code in src/data_handlers/custom_serde_definitions/pandantic.py

@classmethod
def from_dataframe(cls, df: pd.DataFrame, nxt: SerializerFunctionWrapHandler, info: FieldSerializationInfo):
    """
    Create a new instance from a Pandas DataFrame object.

    Args:
        df (pd.DataFrame): Pandas DataFrame object to convert.

    Returns:
        (cls): New instance created from the Pandas DataFrame object.
    """
    return cls(
        columns=df.columns.values,
        values=df.values,
        index=df.index.values,
        index_name=df.index.name,
        columns_name=df.columns.name
    )

from_dict classmethod

from_dict(d: dict)

Create a new instance from a dictionary.

Parameters:

Name	Type	Description	Default
d	dict	Dictionary containing the data.	required

Returns:

Type	Description
DataFrame	New DataFrame created from the dictionary.

Source code in src/data_handlers/custom_serde_definitions/pandantic.py

@classmethod
def from_dict(cls, d: dict):
    """
    Create a new instance from a dictionary.

    Args:
        d (dict): Dictionary containing the data.

    Returns:
        (pd.DataFrame): New DataFrame created from the dictionary.
    """
    try:
        return pd.DataFrame(
            d['values'],
            columns=pd.Index(d['columns'], name=d['columns_name']),
            index=pd.Index(d['index'], name=d['index_name'])
        )
    except:
        raise ValueError(d)

to_dataframe classmethod

to_dataframe(input: DataFrame | dict | str)

Convert input to a Pandas DataFrame object.

Parameters:

Name	Type	Description	Default
input	DataFrame | dict | str	Input data to convert.	required

Returns:

Type	Description
DataFrame	Pandas DataFrame object from input data.

Source code in src/data_handlers/custom_serde_definitions/pandantic.py

@classmethod
def to_dataframe(cls, input: pd.DataFrame | dict | str):
    """
    Convert input to a Pandas DataFrame object.

    Args:
        input (pd.DataFrame | dict | str): Input data to convert.

    Returns:
        (pd.DataFrame): Pandas DataFrame object from input data.
    """
    if isinstance(input, str):
        return cls.model_validate_json(input).as_dataframe()
    elif isinstance(input, dict):
        return cls.from_dict(input)
    elif isinstance(input, pd.DataFrame):
        return input
    else:
        # breakpoint()
        raise ValueError(input)

SeriesSerDe

Bases: BaseModel

Class to serialize and deserialize Pandas Series objects.

Attributes:

Name	Type	Description
name	int | str | None	Name of the series
values	NDArray[Shape['*'], Any]	Values of the series
index	NDArray[Shape['*'], Any]	Index of the series
index_name	int | str | None	Name of the index

Methods: as_series(): Convert to Pandas Series object from_dict(data: dict) -> SeriesSerDe: Create a new instance from a dictionary model_validate_json(json: str | bytes, loc: Any = None, prop: Any = None, cls: Any = None, **kwargs: Any) -> Self: Validate and deserialize JSON data into the model.

as_series

as_series()

Convert to Pandas Series object.

Returns:

Type	Description
Series	Pandas Series object from stored data

Source code in src/data_handlers/custom_serde_definitions/pandantic.py

def as_series(self):
    """
    Convert to Pandas Series object.

    Returns:
        (pd.Series): Pandas Series object from stored data
    """
    return pd.Series(self.values, name=self.name, index=pd.Index(self.index, name=self.index_name))

from_dict classmethod

from_dict(d: dict)

Create a new instance from a dictionary.

Parameters:

Name	Type	Description	Default
d	dict	Dictionary containing the data.	required

Returns:

Type	Description
cls	New instance created from the dictionary.

Source code in src/data_handlers/custom_serde_definitions/pandantic.py

@classmethod
def from_dict(cls, d: dict):
    """
    Create a new instance from a dictionary.

    Args:
        d (dict): Dictionary containing the data.

    Returns:
        (cls): New instance created from the dictionary.
    """
    try:
        return pd.Series(
            d['values'], 
            name=d['name'],
            index=pd.Index(
                d['index'], 
                name=d['index_name'],
            ),
        )
    except:
        raise ValueError(d)

from_series classmethod

from_series(series: Series, nxt: SerializerFunctionWrapHandler, info: FieldSerializationInfo)

Create a new instance from a Pandas Series object.

Parameters:

Name	Type	Description	Default
series	Series	Pandas Series object to convert.	required

Returns:

Type	Description
cls	New instance created from the Pandas Series object.

Source code in src/data_handlers/custom_serde_definitions/pandantic.py

@classmethod
def from_series(cls, series: pd.Series, nxt: SerializerFunctionWrapHandler, info: FieldSerializationInfo):
    """
    Create a new instance from a Pandas Series object.

    Args:
        series (pd.Series): Pandas Series object to convert.

    Returns:
        (cls): New instance created from the Pandas Series object.
    """
    return cls(
        name=series.name,
        values=series.values,
        index=series.index.values,
        index_name=series.index.name,
    )

to_series classmethod

to_series(input: Series | dict | str)

Convert input to a Pandas Series object.

Parameters:

Name	Type	Description	Default
input	Series | dict | str	Input data to convert.	required

Returns:

Type	Description
Series	Pandas Series object from input data.

Source code in src/data_handlers/custom_serde_definitions/pandantic.py

@classmethod
def to_series(cls, input: pd.Series | dict | str):
    """
    Convert input to a Pandas Series object.

    Args:
        input (pd.Series | dict | str): Input data to convert.

    Returns:
        (pd.Series): Pandas Series object from input data.
    """
    if isinstance(input, str):
        print('str')
        return cls.model_validate_json(input).as_series()
    elif isinstance(input, dict):
        print('dict')
        return cls.from_dict(input)
    elif isinstance(input, pd.Series):
        print('series')
        return input
    else:
        raise ValueError(input)

np_encoder

np_encoder(object)

JSON encoder function for numpy types.

Parameters:

Name	Type	Description	Default
object	Any	Object to encode	required

Returns:

Name	Type	Description
Any		Encoded object

Source code in src/data_handlers/custom_serde_definitions/pandantic.py

def np_encoder(object):
    """
    JSON encoder function for numpy types.

    Args:
        object (Any): Object to encode

    Returns:
        Any: Encoded object
    """
    if isinstance(object, np.generic):
        return object.item()
    else:
        return object