Implemented IGH Oceanic View #2226

jkrasting · 2020-05-15T15:18:05Z

The current implementation of the Interrupted Goode Homolosine
projection emphasizes land area. This is a compliment projection
that emphasizes ocean area.
A value of lon0=-160 produces a reasonable real-world map.
Tested update using cartopy:

- The current implementation of the Interrupted Goode Homolosine projection emphasizes land area. This is a compliment projection that emphasizes ocean area. - A value of lon0=-150 produces a reasonable real-world map.

kbevers

Thanks for adding this. Very cool to get this projection in PROJ!

There's a few things missing from this PR:

igh_o.png does not seem to be commited
Tests - can be added in test\gie\, either as a standalone file or in e.g. builtins.gie.
Less important, but still relevant: More descriptive docs. It would be nice to have a short introduction that explains what this projection is.

src/projections/igh_o.cpp

jkrasting

Thanks for your feedback - just submitted a PR with a test case and image.

jkrasting · 2020-05-15T17:17:17Z

FYI -- see complement pull request with cartopy

src/projections/igh_o.cpp

rouault · 2020-05-15T19:51:45Z

/Users/travis/build/OSGeo/PROJ/docs/source/operations/projections/igh_o.rst:document isn't included in any toctree

==> It should be referenced in docs/source/operations/projections/index.rst

- new name is more descriptive - replaced #define with const declaration

- In igh_o.cpp: bool ok = false

- Applies to igh and igh_o projections

src/projections/igh.cpp

- changed in igh*.cpp

rouault · 2020-05-15T23:57:03Z

I actually see this projection was a feature request of #1442 . But in #1442 (comment) , the interruptions values suggested seem different to the one you propose, but the projection maps look similar. I'm not sure who is right or who is wrong or maybe if both proposals are identical, but wanted to raise this and if we could know if there's interoperability with the equivalent ArcGIS projection (https://desktop.arcgis.com/en/arcmap/10.3/guide-books/map-projections/goodes-homolosine.htm)

jkrasting · 2020-05-16T01:55:48Z

Yes I see some differences. Thanks for pointing this out. The original reference from 1925 is here: https://www.tandfonline.com/doi/abs/10.1080/00045602509356949

Some adjustments are needed, but this version is pretty close

kbevers

Unfortunately the original paper doesn't mention how the zones are divided in the oceanic view, so there's no help there. What's implemented now is likely good enough and I'm sure we'll made aware if it isn't.

I think this is looking good now. Thanks for putting in the extra effort improving the code, very much appreaciated.

jkrasting · 2020-05-16T13:11:33Z

No, but the map is instructive. It looks like lon_0 should be -160 and some minor adjustments to the true longitudes and split lines are needed. Will update and push a new commit today. Thats @kbevers and @rouault for pointing this out.

rouault · 2020-05-16T13:34:03Z

Another source for inspiration / comparison is this implementation based on d3.js: https://bl.ocks.org/Fil/d90a94d7333bfef753e92a4abb7611b4 / https://observablehq.com/@d3/interrupted-homolosine-oceans. The convention for the lobes is explained at https://github.com/d3/d3-geo-projection/#interrupted-projections . Looking at the values for the standard IGH, they seem to match with PROJ igh (at list the limits of zones)

- Updates based on Goode, JP, 1925. (DOI: 10.1080/00045602509356949) - High-latitudes "lobes" added for projections #1 and OSGeo#3 - Central longitude changed to -160 - Added fill reference to docstrings for igh.cpp and igh_o.cpp

jkrasting · 2020-05-16T23:17:59Z

The lobes are present in the original reference. Zones 1 and 3 have them for the ocean projection. I added them into the latest commit.

docs/plot/plotdefs.json

- igh --> igh_o

test/gie/builtins.gie

- Each of the 12 sub-projections is tested independently.

rouault · 2020-05-18T14:35:31Z

test/gie/builtins.gie

@@ -2183,10 +2209,35 @@ expect  -0.001790497 0.000895247
 accept  -200 -100
 expect  -0.001790496 -0.000895247

+accept -100.0 22.0


those points in the inverse direction are all close to lat=lon=0 and not super relevant. I would just remove them completely, but as suggested previously, after each expect statement above add a "roundtrip 1" line, that will check that when using the inverse path the values of expect are back projected to the ones of accept

I added in the roundtrip option for all zones in my latest commit, but I am seeing failures with infs in zones 11 and 12. I am stumped. Do you see anything obvious with how I implemented the tests?

test/gie/builtins.gie

rouault

tests can be simplified & improved. See 2 comments

- Tests fail for zones 11 & 12 with inf results

- Applies to igh and igh_o

This reverts commit 93bee75.

- Zones 11 & 12 were reversed; error caused roundtrip test to fail

jkrasting · 2020-05-18T23:15:47Z

0 was fine. z is a int. assigning it a boolean value is confusing (even if false will ultimately evaluate to 0 once cast to int)

I reverted the commit.

rouault · 2020-05-18T23:32:46Z

did you regenerate the image after your latest adjustments ?

jkrasting · 2020-05-18T23:33:45Z

did you regenerate the image after your latest adjustments ?

It shouldn't change the image but I'll double check to confirm.

rouault · 2020-05-18T23:44:56Z

I would expect bf6a8dc to have impacts on the forward path, no ?

jkrasting · 2020-05-18T23:50:44Z

The plot is identical. The change in bf6a8dc is aesthetic and was done to be consistent to match igh.cpp. That block of code identifies the delta value for latitude which did not change when comparing those two zones.

Implemented IGH Oceanic View

a6b972e

- The current implementation of the Interrupted Goode Homolosine projection emphasizes land area. This is a compliment projection that emphasizes ocean area. - A value of lon0=-150 produces a reasonable real-world map.

jkrasting mentioned this pull request May 15, 2020

Implemented IGH Oceanic View SciTools/cartopy#1561

Merged

Fixed table formatting error in igh_o.rst

d782832

kbevers requested changes May 15, 2020

View changes

src/projections/igh_o.cpp Outdated Show resolved Hide resolved

src/projections/igh_o.cpp Outdated Show resolved Hide resolved

src/projections/igh_o.cpp Outdated Show resolved Hide resolved

src/projections/igh_o.cpp Outdated Show resolved Hide resolved

Added test and image for igh_o projection

b46179a

jkrasting commented May 15, 2020

View changes

jkrasting added 2 commits May 15, 2020 14:43

Updated igh_o testing values

f661d28

Added docstrings igh_o.cpp

9f1d467

jkrasting force-pushed the dev/igh_o branch from e839502 to 9f1d467 Compare May 15, 2020 18:43

jkrasting added 2 commits May 15, 2020 15:08

Edits to igh_o test values

1aaa388

Updated sign error on some igh_o test values

73f5d6f

rouault reviewed May 15, 2020

View changes

src/projections/igh_o.cpp Show resolved Hide resolved

Referenced igh_o rst doc file in index

4f1fced

jkrasting added 4 commits May 15, 2020 15:52

Replaced d4044118 with phi_boundary

d34748b

- new name is more descriptive - replaced #define with const declaration

Changed int=0 to bool type

bc96737

- In igh_o.cpp: bool ok = false

Extended doc updates to igh.cpp

cf4f889

Replaced SETUP macro with setup_zone func

28acad1

- Applies to igh and igh_o projections

rouault reviewed May 15, 2020

View changes

src/projections/igh.cpp Outdated Show resolved Hide resolved

jkrasting added 2 commits May 15, 2020 18:18

Removed trailing backslashes from setup_zone

c32875b

- changed in igh*.cpp

Simplified 'struct pj_opaque *Q' call in igh*.cpp

87761d4

jkrasting requested a review from kbevers May 15, 2020 23:43

kbevers approved these changes May 16, 2020

View changes

Updated sub-projection boundaries for igh_o

68e164a

- Updates based on Goode, JP, 1925. (DOI: 10.1080/00045602509356949) - High-latitudes "lobes" added for projections #1 and OSGeo#3 - Central longitude changed to -160 - Added fill reference to docstrings for igh.cpp and igh_o.cpp

rouault reviewed May 16, 2020

View changes

docs/plot/plotdefs.json Outdated Show resolved Hide resolved

jkrasting added 2 commits May 16, 2020 19:45

Updated testing values for igh_o

59813f1

Fixed typo in plotdefs.json

d2b5171

- igh --> igh_o

jkrasting force-pushed the dev/igh_o branch from 11157d0 to d2b5171 Compare May 16, 2020 23:46

rouault reviewed May 17, 2020

View changes

test/gie/builtins.gie Show resolved Hide resolved

jkrasting added 2 commits May 18, 2020 07:19

Sub-projection tests for igh and igh_o

b785d2c

- Each of the 12 sub-projections is tested independently.

Update test/gie/builtins.gie

640d260

jkrasting force-pushed the dev/igh_o branch from 39b6429 to 640d260 Compare May 18, 2020 11:23

jkrasting requested review from kbevers and rouault May 18, 2020 14:27

rouault reviewed May 18, 2020

View changes

test/gie/builtins.gie Outdated Show resolved Hide resolved

rouault requested changes May 18, 2020

View changes

jkrasting added 5 commits May 18, 2020 11:33

Added roundtrip for igh_o subprojection tests

5abdf10

- Tests fail for zones 11 & 12 with inf results

Updated if tests for bool vs int variables

93bee75

- Applies to igh and igh_o

Changed boundary from zone 3 to zone 4 in igh_o

bf6a8dc

Revert "Updated if tests for bool vs int variables"

a7105e3

This reverts commit 93bee75.

Fixed typo in igh_o zone definitions

4dc21af

- Zones 11 & 12 were reversed; error caused roundtrip test to fail

jkrasting requested a review from rouault May 18, 2020 23:15

rouault merged commit 2e54703 into OSGeo:master May 19, 2020
2 checks passed

rouault added this to the 7.1.0 milestone May 19, 2020

Implemented IGH Oceanic View #2226

Implemented IGH Oceanic View #2226

jkrasting commented May 15, 2020 •

edited

kbevers left a comment

jkrasting left a comment

jkrasting commented May 15, 2020

rouault commented May 15, 2020 •

edited

rouault commented May 15, 2020 •

edited

jkrasting commented May 16, 2020

kbevers left a comment

jkrasting commented May 16, 2020

rouault commented May 16, 2020

jkrasting commented May 16, 2020 •

edited

rouault May 18, 2020 •

edited

jkrasting May 18, 2020

rouault left a comment

jkrasting commented May 18, 2020

rouault commented May 18, 2020

jkrasting commented May 18, 2020

rouault commented May 18, 2020

jkrasting commented May 18, 2020 •

edited

Implemented IGH Oceanic View #2226

Implemented IGH Oceanic View #2226

Conversation

jkrasting commented May 15, 2020 • edited

kbevers left a comment

Choose a reason for hiding this comment

jkrasting left a comment

Choose a reason for hiding this comment

jkrasting commented May 15, 2020

rouault commented May 15, 2020 • edited

rouault commented May 15, 2020 • edited

jkrasting commented May 16, 2020

kbevers left a comment

Choose a reason for hiding this comment

jkrasting commented May 16, 2020

rouault commented May 16, 2020

jkrasting commented May 16, 2020 • edited

rouault May 18, 2020 • edited

Choose a reason for hiding this comment

jkrasting May 18, 2020

Choose a reason for hiding this comment

rouault left a comment

Choose a reason for hiding this comment

jkrasting commented May 18, 2020

rouault commented May 18, 2020

jkrasting commented May 18, 2020

rouault commented May 18, 2020

jkrasting commented May 18, 2020 • edited

jkrasting commented May 15, 2020 •

edited

rouault commented May 15, 2020 •

edited

rouault commented May 15, 2020 •

edited

jkrasting commented May 16, 2020 •

edited

rouault May 18, 2020 •

edited

jkrasting commented May 18, 2020 •

edited