Throw exception if not enough data is passed to cmor.write in Python

[mauzey1]

Resolves #485

This change checks the amount of data being passed to cmor.write in Python if it is the expected amount for the amount of time passed (either by setting ntimes_passed or passing time values). It will throw an exception if the amount is lower than expected (or if ntimes_passed is negative).

I also added a test to see if the exception is being thrown.

As for Fortran, it looks like it already handles this issue.

cmor/Src/cmor_fortran_interface.f90

Lines 1032 to 1037 in 8a48023

	if (tshape(i).gt.dshape(i)) then
	write(msg, *) 'In FORTRAN: you passed a 2d array that'&
	//' is not big enough to hold your var, your data shape is:',&
	dshape,'expected shape was:',tshape,char(0)
	error_code = CMOR_CRITICAL
	call cmor_handle_error(msg,error_code)

[doutriaux1]

@mauzey1 could we have a FORTRAN test? Also why is the test called "CV" it's not prepare/cv related.

[mauzey1]

@doutriaux1 It is currently named test_python_CMIP6_CV_not_enough_data.py since the CMIP6_CV tests are the only Python tests currently being used in CI. #493 will add the other Python tests back in.

I will get started on a Fortran test.

[mauzey1]

@doutriaux1 Is there a way to catch an error in Fortran? If cmor_write has an error it will return CMOR_CRITICAL, but I want to catch the "not enough data" message.

[doutriaux1]

@mauzey1 I was dealing with these sort of things back here: https://github.com/PCMDI/cmor/pull/387/files

Let me know if it does not help you.

[mauzey1]

@doutriaux1 I am not sure what to take from that.

Another issue with the Fortran test is that the error uses CMOR_CRITICAL, which exits the program at cmor_handle_error before cmor_write can return the error flag.

[mauzey1]

@doutriaux1 I have moved the not-enough-data test from the CMIP6 CV tests to the Python tests in the Makefile.