r.regression.multi: Add JSON support

[NishantBansal2003]

Fixes: #6010

This PR adds JSON support to the r.regression.multi module. The JSON output looks like:

{
 "n": 5250,
 "Rsq": 0.11506060292031361,
 "Rsqadj": 0.11455453769133173,
 "RMSE": 0.026742231300608307,
 "MAE": 0.020033893876888382,
 "F": 227.36318626709763,
 "b0": 0.05114332327652484,
 "AIC": -38017.8682981307,
 "AICc": -38017.8682981307,
 "BIC": -37991.60436470836,
 "predictor1": "elevation",
 "b1": 0.0015163347280906883,
 "F1": 637.1296966058939,
 "AIC1": -37418.09754711971,
 "AICc1": -37418.09754711971,
 "BIC1": -37406.96558040854,
 "predictor2": "aspect",
 "b2": 2.2652561928494878e-05,
 "F2": 51.058508308209774,
 "AIC2": -37969.0179184159,
 "AICc2": -37969.0179184159,
 "BIC2": -37957.885951704724,
 "predictor3": "slope",
 "b3": 0.002806567566624831,
 "F3": 115.70987418133342,
 "AIC3": -37905.32879012771,
 "AICc3": -37905.32879012771,
 "BIC3": -37894.19682341653
}

This PR includes the following changes:

1. Adds a format option with shell, and json modes for output formatting.
2. Adds format = shell option; the -g flag is now deprecated.
3. Adds tests covering each of the new formats.
4. Adds a Python example to the documentation for parsing JSON output.

NOTE: For multiple mapx values, I kept the JSON structure flat, just like the shell output. Let me know if a nested (array) structure would make more sense.