♻️ Standardize provider configuration structure

[bdamokos]

Standardize provider configuration structure

Issue Description

Currently, each provider (STIB, De Lijn, BKK) uses slightly different methods and structures for configuring monitored lines and stops. This inconsistency makes the system harder to maintain and configure.

Current State

1. STIB:

   • Uses STIB_STOPS with complex nested structure for lines and directions
   • Has both legacy and new configuration formats

2. De Lijn:

   • Uses simple STOP_IDS list
   • Separate MONITORED_LINES list

3. BKK:

   • Uses STOP_IDS list
   • Uses MONITORED_LINES list
   • Has additional provider-specific fields

Proposed Solution

Create a standardized configuration structure for all providers:

PROVIDER_CONFIG = {
    'provider_name': {
        'stops': [
            {
                'id': 'stop_id',
                'name': 'stop_name',  # Optional
                'lines': {  # Optional
                    'line_id': ['destination1', 'destination2']
                },
                'direction': 'direction'  # Optional
            }
        ],
        'monitored_lines': ['line_id1', 'line_id2'],  # Optional
        'provider_specific': {  # For any provider-specific settings
            'key': 'value'
        }
    }
}

Implementation Steps

1. Create migration scripts for each provider
2. Update configuration loading code
3. Update documentation
4. Add validation for the new structure
5. Deprecate old configuration formats

Success Criteria

• Single, consistent configuration format for all providers
• Clear documentation for users
• Backward compatibility or clear migration path
• Validation for configuration format

[bdamokos]

Progress Update: Configuration Standardization

We have implemented a standardized configuration structure that works across all providers. The new schema is:

PROVIDER_CONFIG = {
    'provider_name': {
        'provider_specific': {
            # Provider-specific settings
            'API_KEY': os.getenv('PROVIDER_API_KEY'),
            'CACHE_DIR': Path('cache/provider'),
            'GTFS_DIR': Path('cache/provider/gtfs'),
            'RATE_LIMIT_DELAY': 0.5,  # seconds
            'GTFS_CACHE_DURATION': 86400*30,  # seconds
            '_AVAILABLE_LANGUAGES': ['en', 'fr'],  # Optional
            'GTFS_USED_FILES': ['stops.txt', 'routes.txt', ...],  # Optional
            # ... other provider-specific settings
        },
        'stops': [
            {
                'id': 'stop_id',
                'name': 'stop_name',
                'lines': {
                    'line_id': [
                        {
                            'type': 'direction_name',  # or 'stop_name'
                            'value': 'Heen'  # or actual stop name
                        }
                    ]
                },
                'direction': 'City|Suburb'  # Legacy field for STIB
            }
        ],
        'monitored_lines': ['line1', 'line2']  # Optional global line monitoring
    }
}

Example from BKK configuration:

'bkk': {
    'provider_specific': {
        'PROVIDER_ID': 'mdb-990',
        'API_KEY': os.getenv('BKK_API_KEY'),
        'CACHE_DIR': Path('cache/bkk'),
        'GTFS_DIR': Path('cache/bkk/gtfs'),
        'RATE_LIMIT_DELAY': 0.5,
        'GTFS_CACHE_DURATION': 7 * 24 * 60 * 60
    },
    'stops': [
        {
            'id': 'F01111',
            'name': 'Wesselényi utca / Erzsébet körút',
            'lines': {
                '3060': [
                    {
                        'type': 'direction_name',
                        'value': 'Széll Kálmán tér M'
                    }
                ]
            }
        }
    ],
    'monitored_lines': ['3060']
}

Current Status:

• All providers (STIB, De Lijn, BKK) using new schema in default.py
• Backward compatibility layer working
• Each provider's specific settings preserved in provider_specific

Known Limitation: Direction handling varies by provider