Closed d70-t closed 2 years ago
The build fails because sonde_id
and platform_id
are being read as bytestring objects when opened via OPeNDAP, which would require .decode("utf-8")
. However, when the datasets are opened via zarr (as it was the case before), the string are read in as already decoded strings.
Doing the decoding manually creates a bit of a mess, especially when done conditionally. That's why I'm still hesitant to to the "simple fix".
Due to the merge of eurec4a/eurec4a-intake#80, the updated dropsonde code now does work. The checks should pass after eurec4a/eurec4a-intake#81 is merged.
Building the book now fails because of eurec4a/eurec4a-intake#81 not yet being resolved. Otherwise, we should now be back at a working state 😄
The JOANNE dataset received an update which changed the sonde dimension from
sounding
tosonde_id
and removed thePW
variable. This made several changes throughout the book necessary.