BUG #1886: Polar projection does not change pole.

[danlipsa]

Polar projection should use:
South Pole for y1 < y2
North Pole for y1 > y2.

[danlipsa]

Baselines in
CDAT/uvcdat-testdata#120

[danlipsa]

@doutriaux1 @aashish24 Please review. See also the issues observed while debugging this:
#1901
#1903
#1905

[danlipsa]

The following tests still fail on master and on this branch:
43 - vcs_read_old_scr_2 (Failed)
423 - vcs_test_continents (Failed)
561 - diags_test_02 (Failed)
562 - diags_test_03 (Failed)
563 - diags_test_04 (Failed)
564 - diags_test_41 (Failed)
565 - diags_test_05 (Failed)
566 - diags_test_06 (Failed)
567 - diags_test_07 (Failed)
568 - diags_test_08 (Failed)
569 - diags_test_09 (Failed)
570 - diags_test_10 (Failed)
571 - diags_test_11 (Failed)
572 - diags_test_12 (Failed)
573 - diags_test_13 (Failed)
574 - diags_test_15 (Failed)

[aashish24]

@danlipsa why there is a hole in test_vcs_basic_isofill_-3_proj_gmflip_NH_via_gm.png?

[aashish24]

this sentence does not read well --> such that parameters such that central meridian are set correctly

@danlipsa could you please clean it up?

[danlipsa]

Sure, I'll cleanup the comment. The command for test_vcs_basic_isofill_-3_proj_gmflip_NH_via_gm.png is

"/home/danlipsa/src/uvcdat/testing/vcs/test_vcs_basic_gms.py" "--gm_type=isofill" "--projection=-3" "--lat1=90" "--lat2=0" "--range_via_gm" "--gm_flips_lat_range" "--source=/home/danlipsa/build/uvcdat/uvcdat-testdata/baselines/vcs/test_vcs_basic_isofill_-3_proj_gmflip_NH_via_gm.png"

so
(90, 0) flipped -> (0, 90).
which means view from the South.
The hole is (-90, 0). Note the NH which is appended to the file is wrong: #1903

[aashish24]

@danlipsa I have some questions but I see the pattern for meshfill which seems reasonable to me but I will have detailed look after I get the answers.

[doutriaux1]

@danlipsa meshfill will be fixed in a separate PR correct?

[danlipsa]

@doutriaux1 Yes, we can definitely work on fixing the meshfill. #1905

[doutriaux1]

Ok I'll test the updated PR and merge in then.

[danlipsa]

@doutriaux1 Note meshfill might be trickier to get working as zooming does not work correctly for it:
#1912

[danlipsa]

@doutriaux1 @aashish24 Do I merge this in or any of you want to merge it?