Better way to send resource fields information

[hannelita]

At this time they are sent via array in the DAG params (e.g. "schema_fields_array": "['field1', 'field2']") and everything is treated as text type. What are good alternatives? Questions:

• Modify this array to become a dictionary of names and types. What is a good way to pass it? I believe doing it in plain text can be tedious. What about adding another node on the DAG to fetch the header of the CSV and automatically create a dictionary of fields, hard-coding everything to text?
• Should we start considering different types?
• If yes, define how to treat errors

[hannelita]

Waiting for @rufuspollock analysis to ETL pipeline