r.coin: Add JSON output support

[sumitchintanwar]

This PR adds JSON output support for r.coin module.

Comprehensive unit tests are added to verify:

• JSON structure and field types
• Cross-unit consistency and area conversions
• Percent totals and known reference values

This improves machine-readability and prevents regressions in JSON output behavior.

Below is an example of JSON Output for r.coin:

{
    "module": "r.coin",
    "map1": "landuse",
    "map2": "elevation",
    "unit": {
        "code": "c",
        "name": "cells"
    },
    "coincidence": [
        {
            "cat1": 11,
            "cat2": 1,
            "count": 1,
            "area": 900,
            "value": 1
        },
        {
            "cat1": 11,
            "cat2": 2,
            "count": 4,
            "area": 3600,
            "value": 4
        }
    ]
}


Results: All 12 pass with no issues on the `nc_spm_full_v2beta1` dataset

[petrasovaa]

Let's first discuss the JSON structure... Here is a suggestions that mimicks the plain ascii:

  {
      "project": "spearfish",
      "created": "2025-06-01T...",
      "region": {
          "north": 4928000.00,
          "south": 4914000.00,
          "east": 609000.00,
          "west": 590000.00,
          "ewres": 30,
          "nsres": 30
      },
      "mask": null,
      "maps": [
          {"name": "owner", "title": "Ownership", "type": "raster"},
          {"name": "roads", "title": "Roads", "type": "raster"}
      ],
      "unit": "square miles",
      "row_cats": [0, 1, 2, 3, 4, 5],
      "col_cats": [1, 2],
      "matrix": [
          [50.63, 37.49],
          [ 1.53,  0.68],
          [ 2.92,  0.72],
          [ 3.97,  2.57],
          [ 0.65,  1.36],
          [ 0.19,  0.00]
      ],
      "row_totals": [88.12, 2.21, 3.64, 6.54, 2.00, 0.19],
      "row_totals_without_zero": [88.12, 2.21, 3.64, 6.54, 2.00, 0.19],
      "col_totals": [59.90, 42.80],
      "col_totals_without_zero": [9.27, 5.32],
      "totals": {"value": 102.70, "value_without_zero": 14.58}
  }

Note there are very similar tools - r.kappa and r.report, r.report has nested schema while r.kappa provide a value matrix. It seems r.coin is more similar to r.kappa in this regard, so hence I suggest the JSON above.

[sumitchintanwar]

Noted on the JSON format. Will come up with a fix soon.