Different size and not correct numbers in fcoef variable in cal_<WMO>.mat

[kamwal]

I found that in Python output, in cal_.mat the "fcoef" variable has size 1x4. All of these four values in the row have the same number. That is in disagreement with output from Matlab, where the size of "fcoef" in 1x2 and these two values have different number.

WMO: 1901847 Reference data: CTD_for_DMQC_2019V01 WMO boxes: wmo_boxes_ctd.mat set_calseries:

` # Config calseries parameters

    breaks = []
    max_breaks = 4  # 0 for linear trend, -1 for offset
    calseries = np.ones((1, no_profiles)).flatten()
    # example for splitting time series at profile 33
    # calseries = np.concatenate((np.ones((1, 33)), 2 * np.ones(1,no_profiles - 33)))
    calib_profile_no = profile_no
    use_percent_gt = 0.5
    use_theta_lt = 0
    use_theta_gt = 0
    use_pres_lt = 0
    use_pres_gt = 0`

Canfiguration: ` # Objective Mapping Parameters #

    # max number of historical casts used in objective mapping
    'CONFIG_MAX_CASTS': 300,

    # 1=use PV constraint, 0=don't use PV constraint, in objective mapping
    'MAP_USE_PV': 0,

    # 1=use SAF separation criteria, 0=don't use SAF separation criteria, in objective mapping
    'MAP_USE_SAF': 0,

    # spatial decorrelation scales, in degrees
    'MAPSCALE_LONGITUDE_LARGE': 8,
    'MAPSCALE_LONGITUDE_SMALL': 4,
    'MAPSCALE_LATITUDE_LARGE': 4,
    'MAPSCALE_LATITUDE_SMALL': 2,

    # cross-isobath scales, dimensionless, see BS(2005)
    'MAPSCALE_PHI_LARGE': 0.1,
    'MAPSCALE_PHI_SMALL': 0.02,

    # temporal decorrelation scale, in years
    'MAPSCALE_AGE_LARGE': 20,
    'MAPSCALE_AGE_SMALL': 5,

    # exclude the top xxx dbar of the water column
    'MAP_P_EXCLUDE': 100,

    # only use historical data that are within +/- yyy dbar from float data
    'MAP_P_DELTA': 200
}`

[edsmall-bodc]

@kamwal Can you also just attach the outputs that disagree? Just in a comment, so I have the actual numbers too. Thanks.

[kamwal]

Matlab output from right (fcoef), Python output from left (fcoef_py)

[edsmall-bodc]

Just a note, the above images are the wrong way round: python output on the right, matlab output on the left.

[kamwal]

In the zip file I included source and mapped file

1901847.zip

[edsmall-bodc]

Problem with calculation:

A[0] = A[0] * y_scale * y_0 should be A[0] = A[0] * y_scale + y_0

[edsmall-bodc]

Fixed in https://github.com/euroargodev/argodmqc_owc/pull/40

@kamwal the branch is set up so that you can just run the analysis with the float that caused the issue. When it is confirmed as fix, I will remove the float from the branch before merging