[saket0187]

This PR addresses #6970

I have tried to write this in a cleaner way covering all edge cases. I have also added one test with nested loops to check each key-value pair (with floating-point handling). However, if it looks like too much, we can just use assertEqual. I have removed a few keys like timestamp and generator from expected output as they might cause problems.

Let me know if any changes are needed from my side or if tests need to be modified (or if expected values need to be simplified).

[petrasovaa]

When I run the new test without the change, I get:

======================================================================
FAIL: test_json_profile (__main__.TestClipling.test_json_profile)
Test JSON profile output structure and key values
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/akratoc/dev/grass/grass_main/raster/r.geomorphon/testsuite/test_r_geom.py", line 336, in test_json_profile
    self.assertAlmostEqual(
AssertionError: 133.6702117919922 != 133.67 within 4 places (0.00021179199220000555 difference)

----------------------------------------------------------------------
Ran 3 tests in 14.386s

FAILED (failures=1)

So the PR changes the way the floating point values are printed? It's not a big deal, but it would be good to clarify.

[saket0187]

When I run the new test without the change, I get:

======================================================================
FAIL: test_json_profile (__main__.TestClipling.test_json_profile)
Test JSON profile output structure and key values
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/akratoc/dev/grass/grass_main/raster/r.geomorphon/testsuite/test_r_geom.py", line 336, in test_json_profile
    self.assertAlmostEqual(
AssertionError: 133.6702117919922 != 133.67 within 4 places (0.00021179199220000555 difference)

----------------------------------------------------------------------
Ran 3 tests in 14.386s

FAILED (failures=1)

So the PR changes the way the floating point values are printed? It's not a big deal, but it would be good to clarify.

Yes, The old path went through format_token_common(), which used %.2f for T_MTR, so the number was rounded before it reached JSON.

[petrasovaa]

Thanks!

[infrastation]

How does the new JSON test for r.geomorphon manage to pass using an expected output that does not define the top-level keys format_version_major, format_version_minor, timestamp and generator? r.geomorphon is supposed to produce all four at the end of the JSON map, and though for testing purposes it is fine — and even advised — to ignore the volatile values of the latter two, the former two values are essential encoding data.

As far as it looks on the diff, the test does not verify the presence of the four keys and does no verify the two values. Could someone verify this? I do not have GRASS rigged up now.