v.profile: Add JSON output format

[saket0187]

Fixes #6232

Problem Description

The v.profile module only supported plain text output with pipe-delimited or custom separators. We were required to add JSON support as well.

Solution

Implemented comprehensive output format support with three options:

1. Plain format (default): Traditional CSV-style output with quoted string values. Maintains full backward compatibility with existing scripts.

2. CSV format: Unquoted delimiter-separated values optimized for spreadsheet applications. Uses the same separator parameter as plain format but omits quotes for cleaner parsing.

3. JSON format: Structured output with proper data type handling. Each feature is represented as a JSON object containing:

   • category: Feature category number (always present, even without database)
   • distance: Distance along profile line
   • attributes

Implementation details:

• Uses GRASS GJSON API (G_json_* functions) for JSON serialization with proper NULL value handling
• Added comprehensive error checking with G_warning() for database operation failures
• Works correctly with and without database connections (Fi != NULL check)
• Updated format option to include "json" in descriptions

Testing

Automated Tests (test_v_profile.py):
Added testJsonFormat() that validates:

• JSON syntax correctness (parses without errors)
• Array structure with expected item count
• Required fields present (category, distance, attributes)
• Data type validation (category as int, distance as float)
• Value accuracy against expected results
• Attribute field completeness

All 6 test suite tests pass, including the new JSON format test.

Manual Testing with NC Basic dataset:

# Plain format (backward compatible)
v.profile input=poi_names_wake profile_map=roadsmajor buffer=200 profile_where="cat=354"

# CSV format
v.profile input=poi_names_wake profile_map=roadsmajor buffer=200 profile_where="cat=354" \
  format=csv separator=comma

# JSON format
v.profile input=poi_names_wake profile_map=roadsmajor buffer=200 profile_where="cat=354" \
  format=json

JSON Output example :-

[
    {
        "category": 572,
        "distance": 19537.972502200733,
        "attributes": {
            "cat": 572,
            "feature_id": "986138",
            "featurenam": "Greshams Lake",
            "class": "Reservoir",
            ...
        }
    }
]

[petrasovaa]

Thank you for the contribution, please read our contributing guidelines, specifically pre-commit to get the formatting correct.

[saket0187]

@petrasovaa

I've addressed your feedback:

• All original comments have been added.
• Removed PLAIN format (was similar to CSV)
• Kept original pipe (|) delimiter as default.
• Added JSON format with proper type handling
• Updated documentation with format examples

Please review the updated main.c and v.profile.md files and let me know if any other changes are needed.

Thanks!

[petrasovaa]

@petrasovaa

I've addressed your feedback:

• All original comments have been added.
• Removed PLAIN format (was similar to CSV)
• Kept original pipe (|) delimiter as default.
• Added JSON format with proper type handling
• Updated documentation with format examples

Please review the updated main.c and v.profile.md files and let me know if any other changes are needed.

Thanks!

Please keep the plain mode there for backwards compatibility with | as separator and CSV needs to have comma. All I was trying to avoid is too much code duplication.

[saket0187]

@petrasovaa
I've implemented the changes as discussed:

1. Added plain format - This is the old CSV behavior (pipe separator) and is default for backward compatibility
2. Updated csv format - Now uses comma separator as expected for proper CSV files
3. Eliminated code duplication - Both plain and CSV share the same output code via switch fall-through

The separator logic respects user overrides via the separator= option while providing format-appropriate defaults (pipe for plain, comma for CSV).

Let me know if you'd like any adjustments!