Skip to content

Writer

Defines abstract writer behaviour.

Writer

Bases: ABC

Abstract Writer class.

Source code in mkdocs/lakehouse_engine/packages/io/writer.py 
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
class Writer(ABC):
    """Abstract Writer class."""

    def __init__(
        self, output_spec: OutputSpec, df: DataFrame, data: OrderedDict | None = None
    ) -> None:
        """Construct Writer instances.

        Args:
            output_spec: output specification to write data.
            df: dataframe to write.
            data: list of all dfs generated on previous steps before writer.
        """
        self._logger = LoggingHandler(self.__class__.__name__).get_logger()
        self._output_spec = output_spec
        self._df = df
        self._data = data or OrderedDict()

    @abstractmethod
    def write(self) -> OrderedDict | None:
        """Abstract write method."""
        raise NotImplementedError

    @staticmethod
    def write_transformed_micro_batch(**kwargs: Any) -> Callable:
        """Define how to write a streaming micro batch after transforming it.

        This function must define an inner function that manipulates a streaming batch,
        and then return that function. Look for concrete implementations of this
        function for more clarity.

        Args:
            kwargs: any keyword arguments.

        Returns:
            A function to be executed in the foreachBatch spark write method.
        """

        def inner(batch_df: DataFrame, batch_id: int) -> None:
            logger = LoggingHandler(__name__).get_logger()
            logger.warning("Skipping transform micro batch... nothing to do.")

        return inner

    @classmethod
    def get_transformed_micro_batch(
        cls,
        output_spec: OutputSpec,
        batch_df: DataFrame,
        batch_id: int,
        data: OrderedDict | None,
    ) -> DataFrame:
        """Get the result of the transformations applied to a micro batch dataframe.

        Args:
            output_spec: output specification associated with the writer.
            batch_df: batch dataframe (given from streaming foreachBatch).
            batch_id: if of the batch (given from streaming foreachBatch).
            data: list of all dfs generated on previous steps before writer
                to be available on micro batch transforms.

        Returns:
            The transformed dataframe.
        """
        transformed_df = batch_df
        data = data or OrderedDict()
        if output_spec.with_batch_id:
            transformed_df = transformed_df.withColumn("lhe_batch_id", lit(batch_id))

        for transformer in output_spec.streaming_micro_batch_transformers or []:
            transformed_df = transformed_df.transform(
                TransformerFactory.get_transformer(transformer, data)
            )

        return transformed_df

    @classmethod
    def get_streaming_trigger(cls, output_spec: OutputSpec) -> dict:
        """Define which streaming trigger will be used.

        Args:
            output_spec: output specification.

        Returns:
            A dict containing streaming trigger.
        """
        trigger: dict[str, Any] = {}

        if output_spec.streaming_available_now:
            trigger["availableNow"] = output_spec.streaming_available_now
        elif output_spec.streaming_once:
            trigger["once"] = output_spec.streaming_once
        elif output_spec.streaming_processing_time:
            trigger["processingTime"] = output_spec.streaming_processing_time
        elif output_spec.streaming_continuous:
            trigger["continuous"] = output_spec.streaming_continuous
        else:
            raise NotImplementedError(
                "The requested output spec streaming trigger is not supported."
            )

        return trigger

    @staticmethod
    def run_micro_batch_dq_process(df: DataFrame, dq_spec: list[DQSpec]) -> DataFrame:
        """Run the data quality process in a streaming micro batch dataframe.

        Iterates over the specs and performs the checks or analysis depending on the
        data quality specification provided in the configuration.

        Args:
            df: the dataframe in which to run the dq process on.
            dq_spec: data quality specification.

        Returns: the validated dataframe.
        """
        from lakehouse_engine.dq_processors.dq_factory import DQFactory

        validated_df = df
        for spec in dq_spec:
            validated_df = DQFactory.run_dq_process(spec, df)

        return validated_df

__init__(output_spec, df, data=None)

Construct Writer instances.

Parameters:

Name Type Description Default
output_spec OutputSpec

output specification to write data.

 required
df DataFrame

dataframe to write.

 required
data OrderedDict | None

list of all dfs generated on previous steps before writer.

 None
Source code in mkdocs/lakehouse_engine/packages/io/writer.py 
19
20
21
22
23
24
25
26
27
28
29
30
31
32
def __init__(
    self, output_spec: OutputSpec, df: DataFrame, data: OrderedDict | None = None
) -> None:
    """Construct Writer instances.

    Args:
        output_spec: output specification to write data.
        df: dataframe to write.
        data: list of all dfs generated on previous steps before writer.
    """
    self._logger = LoggingHandler(self.__class__.__name__).get_logger()
    self._output_spec = output_spec
    self._df = df
    self._data = data or OrderedDict()

get_streaming_trigger(output_spec) classmethod

Define which streaming trigger will be used.

Parameters:

Name Type Description Default
output_spec OutputSpec

output specification.

 required

Returns:

Type Description
dict

A dict containing streaming trigger.
Source code in mkdocs/lakehouse_engine/packages/io/writer.py 
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
@classmethod
def get_streaming_trigger(cls, output_spec: OutputSpec) -> dict:
    """Define which streaming trigger will be used.

    Args:
        output_spec: output specification.

    Returns:
        A dict containing streaming trigger.
    """
    trigger: dict[str, Any] = {}

    if output_spec.streaming_available_now:
        trigger["availableNow"] = output_spec.streaming_available_now
    elif output_spec.streaming_once:
        trigger["once"] = output_spec.streaming_once
    elif output_spec.streaming_processing_time:
        trigger["processingTime"] = output_spec.streaming_processing_time
    elif output_spec.streaming_continuous:
        trigger["continuous"] = output_spec.streaming_continuous
    else:
        raise NotImplementedError(
            "The requested output spec streaming trigger is not supported."
        )

    return trigger

get_transformed_micro_batch(output_spec, batch_df, batch_id, data) classmethod

Get the result of the transformations applied to a micro batch dataframe.

Parameters:

Name Type Description Default
output_spec OutputSpec

output specification associated with the writer.

 required
batch_df DataFrame

batch dataframe (given from streaming foreachBatch).

 required
batch_id int

if of the batch (given from streaming foreachBatch).

 required
data OrderedDict | None

list of all dfs generated on previous steps before writer to be available on micro batch transforms.

 required

Returns:

Type Description
DataFrame

The transformed dataframe.
Source code in mkdocs/lakehouse_engine/packages/io/writer.py 
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
@classmethod
def get_transformed_micro_batch(
    cls,
    output_spec: OutputSpec,
    batch_df: DataFrame,
    batch_id: int,
    data: OrderedDict | None,
) -> DataFrame:
    """Get the result of the transformations applied to a micro batch dataframe.

    Args:
        output_spec: output specification associated with the writer.
        batch_df: batch dataframe (given from streaming foreachBatch).
        batch_id: if of the batch (given from streaming foreachBatch).
        data: list of all dfs generated on previous steps before writer
            to be available on micro batch transforms.

    Returns:
        The transformed dataframe.
    """
    transformed_df = batch_df
    data = data or OrderedDict()
    if output_spec.with_batch_id:
        transformed_df = transformed_df.withColumn("lhe_batch_id", lit(batch_id))

    for transformer in output_spec.streaming_micro_batch_transformers or []:
        transformed_df = transformed_df.transform(
            TransformerFactory.get_transformer(transformer, data)
        )

    return transformed_df

run_micro_batch_dq_process(df, dq_spec) staticmethod

Run the data quality process in a streaming micro batch dataframe.

Iterates over the specs and performs the checks or analysis depending on the data quality specification provided in the configuration.

Parameters:

Name Type Description Default
df DataFrame

the dataframe in which to run the dq process on.

 required
dq_spec list[DQSpec]

data quality specification.

 required

Returns: the validated dataframe.

Source code in mkdocs/lakehouse_engine/packages/io/writer.py 
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
@staticmethod
def run_micro_batch_dq_process(df: DataFrame, dq_spec: list[DQSpec]) -> DataFrame:
    """Run the data quality process in a streaming micro batch dataframe.

    Iterates over the specs and performs the checks or analysis depending on the
    data quality specification provided in the configuration.

    Args:
        df: the dataframe in which to run the dq process on.
        dq_spec: data quality specification.

    Returns: the validated dataframe.
    """
    from lakehouse_engine.dq_processors.dq_factory import DQFactory

    validated_df = df
    for spec in dq_spec:
        validated_df = DQFactory.run_dq_process(spec, df)

    return validated_df

write() abstractmethod

Abstract write method.

Source code in mkdocs/lakehouse_engine/packages/io/writer.py 
34
35
36
37
@abstractmethod
def write(self) -> OrderedDict | None:
    """Abstract write method."""
    raise NotImplementedError

write_transformed_micro_batch(**kwargs) staticmethod

Define how to write a streaming micro batch after transforming it.

This function must define an inner function that manipulates a streaming batch, and then return that function. Look for concrete implementations of this function for more clarity.

Parameters:

Name Type Description Default
kwargs Any

any keyword arguments.

 {}

Returns:

Type Description
Callable

A function to be executed in the foreachBatch spark write method.
Source code in mkdocs/lakehouse_engine/packages/io/writer.py 
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
@staticmethod
def write_transformed_micro_batch(**kwargs: Any) -> Callable:
    """Define how to write a streaming micro batch after transforming it.

    This function must define an inner function that manipulates a streaming batch,
    and then return that function. Look for concrete implementations of this
    function for more clarity.

    Args:
        kwargs: any keyword arguments.

    Returns:
        A function to be executed in the foreachBatch spark write method.
    """

    def inner(batch_df: DataFrame, batch_id: int) -> None:
        logger = LoggingHandler(__name__).get_logger()
        logger.warning("Skipping transform micro batch... nothing to do.")

    return inner