As of #22, we have designed the schema to allow for usage types using cli (expecting to write a file to disk) and python (expecting to either return an object or write a file to disk). For Extractors where the output is a file, we provide the {{ output_path }} templating string. However, there is currently no way of specifying what the default is. This might be important for Extractors that do not allow customisation of the output path.
An option would be to add additional attribute, usage:output:, which would let the writer of the Extractor to describe this behaviour. For instance, if yadg is invoked without providing the optional outfile argument:
yadg extract {{ input_type }} {{ input_path }}
the output file will be the root of the {{ input_path }} with a .nc suffix. Currently, there is no way to describe this behaviour.
ml-evs
commented
10 months ago
As of #22, we have designed the schema to allow for
usagetypes usingcli(expecting to write a file to disk) andpython(expecting to either return an object or write a file to disk). ForExtractorswhere the output is a file, we provide the{{ output_path }}templating string. However, there is currently no way of specifying what the default is. This might be important forExtractorsthat do not allow customisation of the output path.An option would be to add additional attribute,
usage:output:, which would let the writer of theExtractorto describe this behaviour. For instance, ifyadgis invoked without providing the optional outfile argument:the output file will be the root of the
{{ input_path }}with a.ncsuffix. Currently, there is no way to describe this behaviour.