Writing a data processor means implementing the DataProcessor interface defined at dataprocessor.go and adding it to the NewDataProcessor factory function.

type DataProcessor interface {
	Init(params map[string]string, identifiers map[string]string, measurements map[string]string, categories map[string]string, tags []string) error
	OnData(data []byte) ([]byte, error)
	GetRecord() (arrow.Record, error)
}

Data Processors are consumed in the Spice.ai pod manifest in the data section. E.g.

data:
  processor:
    name: flux-csv

The data processor name is self-declared by the component, but must be unique across all components.