Allow specifying Spark resources per Python-Py Transformer

[yruslan]

Background

Currently, Python transformations run with Spark resource settings configured per environment.
This is good if the cluster has dynamic resource allocation enabled. But for clusters that don't have dynamic resource allocation enabled it is not possible to fine-tune resources for each transformer.

Solution

Pramen-Py already supports specifying resources per transformer #30. Need to add the capability to Pramen.

Example

{
    name = "Python transformer"
    type = "python_transformation"
    python.class = "SomeTransformer"
    schedule.type = "daily"

    info.date.expr = "@runDate"

    output.table = "transformed_python"

    dependencies = [
      {
        tables = [ table1 ]
        date.from = "@infoDate"
      }
    ]

    spark.config {
      spark.executor.instances = 4
      spark.executor.cores = 1
      spark.executor.memory = "4g"
    }
}

Generated YAML:

run_transformers:
- info_date: 2022-02-18
  output_table: transformed_python
  name: SomeTransformer
  spark_config:
    spark.executor.cores: 1
    spark.executor.instances: 4
    spark.executor.memory: 4g

Docs on Spark config: https://spark.apache.org/docs/latest/configuration.html