I originally started just excising the metadata writing part but soon enough realised that you really want a slightly different code path -- diverging before the csv component is written but after the metadata is written. I've added a metadata_only flag to write_csvy to achieve this.
I still split out write_metadata because it's more modular and can be tested independently, and can be used to write a comparable file from a list generated independent of this package.
Added 14 tests, 100% coverage of write_metadata.
Check was like that when I found it... it's not more broken.
Addresses #17.
I originally started just excising the metadata writing part but soon enough realised that you really want a slightly different code path -- diverging before the csv component is written but after the metadata is written. I've added a
metadata_only
flag towrite_csvy
to achieve this.I still split out
write_metadata
because it's more modular and can be tested independently, and can be used to write a comparable file from a list generated independent of this package.Added 14 tests, 100% coverage of
write_metadata
. Check was like that when I found it... it's not more broken.I recommend squash on merge.