Skip to content

Convert Edge Xpert Json to Process Control Data Model

EdgeXOPCUAJsonToPCDMTransformer

Bases: TransformerInterface

Converts a Spark Dataframe column containing a json string created by EdgeX to the Process Control Data Model.

Example

from rtdip_sdk.pipelines.transformers import EdgeXOPCUAJsonToPCDMTransformer

edge_opcua_json_to_pcdm_transformer = EdgeXOPCUAJsonToPCDMTransformer(
    data=df,
    souce_column_name="body",
    status_null_value="Good",
    change_type_value="insert"
)

result = edge_opcua_json_to_pcdm_transformer.transform()

Parameters:

Name Type Description Default
data DataFrame

Dataframe containing the column with EdgeX data

required
source_column_name str

Spark Dataframe column containing the OPC Publisher Json OPC UA data

required
status_null_value optional str

If populated, will replace 'Good' in the Status column with the specified value.

'Good'
change_type_value optional str

If populated, will replace 'insert' in the ChangeType column with the specified value.

'insert'
Source code in src/sdk/python/rtdip_sdk/pipelines/transformers/spark/edgex_opcua_json_to_pcdm.py
 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
class EdgeXOPCUAJsonToPCDMTransformer(TransformerInterface):
    """
    Converts a Spark Dataframe column containing a json string created by EdgeX to the Process Control Data Model.

    Example
    --------
    ```python
    from rtdip_sdk.pipelines.transformers import EdgeXOPCUAJsonToPCDMTransformer

    edge_opcua_json_to_pcdm_transformer = EdgeXOPCUAJsonToPCDMTransformer(
        data=df,
        souce_column_name="body",
        status_null_value="Good",
        change_type_value="insert"
    )

    result = edge_opcua_json_to_pcdm_transformer.transform()
    ```

    Parameters:
        data (DataFrame): Dataframe containing the column with EdgeX data
        source_column_name (str): Spark Dataframe column containing the OPC Publisher Json OPC UA data
        status_null_value (optional str): If populated, will replace 'Good' in the Status column with the specified value.
        change_type_value (optional str): If populated, will replace 'insert' in the ChangeType column with the specified value.
    """

    data: DataFrame
    source_column_name: str
    status_null_value: str
    change_type_value: str
    tagname_field: str

    def __init__(
        self,
        data: DataFrame,
        source_column_name: str,
        status_null_value: str = "Good",
        change_type_value: str = "insert",
        tagname_field="resourceName",
    ) -> None:
        self.data = data
        self.source_column_name = source_column_name
        self.status_null_value = status_null_value
        self.change_type_value = change_type_value
        self.tagname_field = tagname_field

    @staticmethod
    def system_type():
        """
        Attributes:
            SystemType (Environment): Requires PYSPARK
        """
        return SystemType.PYSPARK

    @staticmethod
    def libraries():
        libraries = Libraries()
        return libraries

    @staticmethod
    def settings() -> dict:
        return {}

    def pre_transform_validation(self):
        return True

    def post_transform_validation(self):
        return True

    def transform(self) -> DataFrame:
        """
        Returns:
            DataFrame: A dataframe with the specified column converted to PCDM
        """
        df = (
            self.data.withColumn(
                self.source_column_name,
                from_json(self.source_column_name, EDGEX_SCHEMA),
            )
            .select("*", explode("{}.readings".format(self.source_column_name)))
            .selectExpr(
                "explode({}.readings.{}) as TagName".format(
                    self.source_column_name, self.tagname_field
                ),
                "to_utc_timestamp(to_timestamp((col.origin / 1000000000)), current_timezone()) as EventTime",
                "col.value as Value",
                "col.valueType as ValueType",
            )
            .withColumn("Status", lit(self.status_null_value))
            .withColumn("ChangeType", lit(self.change_type_value))
            .withColumn(
                "ValueType",
                (
                    when(col("ValueType") == "Int8", "integer")
                    .when(col("ValueType") == "Int16", "integer")
                    .when(col("ValueType") == "Int32", "integer")
                    .when(col("ValueType") == "Int64", "integer")
                    .when(col("ValueType") == "Uint8", "integer")
                    .when(col("ValueType") == "Uint16", "integer")
                    .when(col("ValueType") == "Uint32", "integer")
                    .when(col("ValueType") == "Uint64", "integer")
                    .when(col("ValueType") == "Float32", "float")
                    .when(col("ValueType") == "Float64", "float")
                    .when(col("ValueType") == "Bool", "bool")
                    .otherwise("string")
                ),
            )
        )

        return df.select(
            "TagName", "EventTime", "Status", "Value", "ValueType", "ChangeType"
        )

system_type() staticmethod

Attributes:

Name Type Description
SystemType Environment

Requires PYSPARK

Source code in src/sdk/python/rtdip_sdk/pipelines/transformers/spark/edgex_opcua_json_to_pcdm.py
69
70
71
72
73
74
75
@staticmethod
def system_type():
    """
    Attributes:
        SystemType (Environment): Requires PYSPARK
    """
    return SystemType.PYSPARK

transform()

Returns:

Name Type Description
DataFrame DataFrame

A dataframe with the specified column converted to PCDM

Source code in src/sdk/python/rtdip_sdk/pipelines/transformers/spark/edgex_opcua_json_to_pcdm.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
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
def transform(self) -> DataFrame:
    """
    Returns:
        DataFrame: A dataframe with the specified column converted to PCDM
    """
    df = (
        self.data.withColumn(
            self.source_column_name,
            from_json(self.source_column_name, EDGEX_SCHEMA),
        )
        .select("*", explode("{}.readings".format(self.source_column_name)))
        .selectExpr(
            "explode({}.readings.{}) as TagName".format(
                self.source_column_name, self.tagname_field
            ),
            "to_utc_timestamp(to_timestamp((col.origin / 1000000000)), current_timezone()) as EventTime",
            "col.value as Value",
            "col.valueType as ValueType",
        )
        .withColumn("Status", lit(self.status_null_value))
        .withColumn("ChangeType", lit(self.change_type_value))
        .withColumn(
            "ValueType",
            (
                when(col("ValueType") == "Int8", "integer")
                .when(col("ValueType") == "Int16", "integer")
                .when(col("ValueType") == "Int32", "integer")
                .when(col("ValueType") == "Int64", "integer")
                .when(col("ValueType") == "Uint8", "integer")
                .when(col("ValueType") == "Uint16", "integer")
                .when(col("ValueType") == "Uint32", "integer")
                .when(col("ValueType") == "Uint64", "integer")
                .when(col("ValueType") == "Float32", "float")
                .when(col("ValueType") == "Float64", "float")
                .when(col("ValueType") == "Bool", "bool")
                .otherwise("string")
            ),
        )
    )

    return df.select(
        "TagName", "EventTime", "Status", "Value", "ValueType", "ChangeType"
    )