Solarsally #181
Solarsally #181
Conversation
| ds = xarray.open_dataset(io.BytesIO(ncfile_bytes)) | ||
|
|
||
| # Initialize Google Cloud Storage client | ||
| storage_client = storage.Client() |
There was a problem hiding this comment.
Revert the storage client code once debugging is complete.
| # Set the time using numpy.datetime64 for the expected format | ||
| times[0] = numpy.datetime64("2010-02-02T18:00:00") | ||
| # Close the NetCDF file and retrieve its contents as bytes | ||
| memfile = ncfile.close() |
There was a problem hiding this comment.
based on my understanding we retrieve the contents of ncfile using memfile.tobytes(). I'm concerned about memfile = ncfile.close(), because the close() method doesn’t return any value, leading to a potential error.
just be careful and consider not closing the NetCDF file until after assigned the necessary data for xarray
There was a problem hiding this comment.
This case is with in-memory NetCDF files using netCDF4.Dataset, the behavior is a bit different than in typical file handling when the memory parameter is set to 1 (indicating in-memory storage). By specifying memory=1, we are instructing netCDF4 to create a file in memory instead of on disk, which changes the behavior of how the file is handled. now when we use "memfile = ncfile.close()" , instead of simply closing the file, the in-memory buffer (which contains the NetCDF file's contents) is returned. This buffer is similar to a file-like object, and we can then access the contents in bytes.
I have added two options for the testing file, one for GCP and another for fake data. I have also tested both of them locally, it works with no error. Now the same code handles both fake and real data. @Katsutoshii