{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "title": "Landvex Urban Intelligence Training Dataset",
  "description": "Schema for training data used in the six-layer urban taxonomy model",
  "version": "1.0",
  "last_updated": "2026-06-28",
  
  "definitions": {
    "geo_point": {
      "type": "object",
      "required": ["lat", "lng"],
      "properties": {
        "lat": {"type": "number", "minimum": -90, "maximum": 90},
        "lng": {"type": "number", "minimum": -180, "maximum": 180},
        "accuracy": {"type": "number", "description": "GPS accuracy in meters"}
      }
    },
    
    "timestamp": {
      "type": "string",
      "format": "date-time",
      "description": "ISO 8601 timestamp"
    },
    
    "confidence_score": {
      "type": "number",
      "minimum": 0,
      "maximum": 100,
      "description": "Confidence in this observation (0-100)"
    }
  },
  
  "type": "object",
  "required": ["observation_id", "location", "timestamp", "contributor", "layers"],
  
  "properties": {
    "observation_id": {
      "type": "string",
      "pattern": "^[A-Z]{2}-[0-9]{4}-[0-9]{6}-[0-9]{4}$",
      "description": "Unique ID: CC-YYYY-MMDD-NNNN"
    },
    
    "location": {
      "type": "object",
      "required": ["coordinates", "district", "city", "country"],
      "properties": {
        "coordinates": {"$ref": "#/definitions/geo_point"},
        "district": {"type": "string"},
        "city": {"type": "string"},
        "country": {"type": "string", "pattern": "^[A-Z]{2}$"},
        "address": {"type": "string"},
        "landmark": {"type": "string"}
      }
    },
    
    "timestamp": {"$ref": "#/definitions/timestamp"},
    
    "contributor": {
      "type": "object",
      "required": ["contributor_id", "verification_level"],
      "properties": {
        "contributor_id": {"type": "string"},
        "verification_level": {
          "type": "string",
          "enum": ["unverified", "basic", "verified", "expert"]
        },
        "training_completed": {"type": "boolean"},
        "observation_count": {"type": "integer"}
      }
    },
    
    "layers": {
      "type": "object",
      "required": ["physical", "operational"],
      "properties": {
        "physical": {
          "type": "object",
          "properties": {
            "building_type": {
              "type": "string",
              "enum": ["residential", "commercial", "industrial", "mixed", "institutional", "vacant_land"]
            },
            "building_height": {"type": "integer", "description": "Approximate height in meters"},
            "building_condition": {
              "type": "string",
              "enum": ["excellent", "good", "fair", "poor", "derelict"]
            },
            "building_age_estimate": {
              "type": "string",
              "enum": ["new", "recent", "mature", "old", "historic"]
            },
            "road_type": {
              "type": "string",
              "enum": ["highway", "arterial", "collector", "local", "alley", "pedestrian"]
            },
            "road_condition": {
              "type": "string",
              "enum": ["excellent", "good", "fair", "poor", "unpassable"]
            },
            "sidewalk_present": {"type": "boolean"},
            "sidewalk_condition": {
              "type": "string",
              "enum": ["excellent", "good", "fair", "poor", "none"]
            },
            "street_lighting": {
              "type": "string",
              "enum": ["excellent", "adequate", "poor", "none"]
            },
            "greenery": {
              "type": "string",
              "enum": ["abundant", "moderate", "sparse", "none"]
            },
            "pedestrian_count": {"type": "integer"},
            "vehicle_count": {"type": "integer"},
            "vehicle_types": {
              "type": "array",
              "items": {
                "type": "string",
                "enum": ["cars", "motorcycles", "bicycles", "buses", "trucks", "tuk_tuks", "other"]
              }
            },
            "images": {
              "type": "array",
              "items": {
                "type": "object",
                "properties": {
                  "image_id": {"type": "string"},
                  "url": {"type": "string", "format": "uri"},
                  "angle": {"type": "string", "enum": ["front", "side", "aerial", "interior"]},
                  "timestamp": {"$ref": "#/definitions/timestamp"}
                }
              }
            },
            "confidence": {"$ref": "#/definitions/confidence_score"}
          }
        },
        
        "operational": {
          "type": "object",
          "properties": {
            "businesses_observed": {"type": "integer"},
            "businesses_open": {"type": "integer"},
            "business_types": {
              "type": "array",
              "items": {
                "type": "string",
                "enum": ["retail", "food_service", "services", "entertainment", "healthcare", "education", "manufacturing", "logistics", "other"]
              }
            },
            "activity_level": {
              "type": "string",
              "enum": ["very_high", "high", "moderate", "low", "very_low"]
            },
            "customer_queue_length": {"type": "integer"},
            "delivery_vehicles_present": {"type": "integer"},
            "construction_active": {"type": "boolean"},
            "construction_type": {
              "type": "string",
              "enum": ["new_building", "renovation", "infrastructure", "demolition", "none"]
            },
            "event_present": {"type": "boolean"},
            "event_type": {"type": "string"},
            "street_vendors_count": {"type": "integer"},
            "parking_availability": {
              "type": "string",
              "enum": ["abundant", "adequate", "limited", "none"]
            },
            "noise_level": {
              "type": "string",
              "enum": ["very_loud", "loud", "moderate", "quiet", "silent"]
            },
            "cleanliness": {
              "type": "string",
              "enum": ["excellent", "good", "fair", "poor", "very_poor"]
            },
            "confidence": {"$ref": "#/definitions/confidence_score"}
          }
        },
        
        "economic": {
          "type": "object",
          "properties": {
            "ownership_type": {
              "type": "string",
              "enum": ["family_owned", "sole_proprietor", "partnership", "corporate", "institutional", "government", "cooperative", "unknown"]
            },
            "business_size": {
              "type": "string",
              "enum": ["micro", "small", "medium", "large", "enterprise"]
            },
            "price_level": {
              "type": "string",
              "enum": ["budget", "mid_range", "premium", "luxury"]
            },
            "payment_methods": {
              "type": "array",
              "items": {
                "type": "string",
                "enum": ["cash", "card", "mobile", "crypto", "barter"]
              }
            },
            "informal_economy_signs": {
              "type": "array",
              "items": {
                "type": "string",
                "enum": ["no_receipt", "cash_only", "unmarked_prices", "negotiable_prices", "unlicensed_vendors", "home_based"]
              }
            },
            "revenue_estimate": {
              "type": "string",
              "enum": ["very_high", "high", "moderate", "low", "very_low", "unknown"]
            },
            "profitability_indicator": {
              "type": "string",
              "enum": ["strongly_profitable", "profitable", "break_even", "loss_making", "unknown"]
            },
            "confidence": {"$ref": "#/definitions/confidence_score"}
          }
        },
        
        "institutional": {
          "type": "object",
          "properties": {
            "zoning": {
              "type": "string",
              "enum": ["residential", "commercial", "industrial", "mixed_use", "agricultural", "special", "unknown"]
            },
            "zoning_compliance": {
              "type": "string",
              "enum": ["fully_compliant", "minor_violations", "major_violations", "non_compliant", "unknown"]
            },
            "permit_displayed": {"type": "boolean"},
            "permit_type": {
              "type": "string",
              "enum": ["business_license", "health_permit", "building_permit", "none_visible", "unknown"]
            },
            "tax_id_visible": {"type": "boolean"},
            "regulatory_signs": {
              "type": "array",
              "items": {"type": "string"}
            },
            "enforcement_presence": {
              "type": "string",
              "enum": ["visible", "occasional", "rare", "none", "unknown"]
            },
            "confidence": {"$ref": "#/definitions/confidence_score"}
          }
        },
        
        "social": {
          "type": "object",
          "properties": {
            "family_business": {"type": "boolean"},
            "family_members_present": {"type": "integer"},
            "migrant_worker_signs": {
              "type": "array",
              "items": {
                "type": "string",
                "enum": ["foreign_language_signage", "remittance_services", "migrant_goods", "documentation_services", "none"]
              }
            },
            "tourist_ratio_estimate": {
              "type": "string",
              "enum": ["mostly_tourist", "mixed", "mostly_local", "unknown"]
            },
            "community_facilities": {
              "type": "array",
              "items": {
                "type": "string",
                "enum": ["temple", "church", "mosque", "community_center", "school", "park", "market", "none"]
              }
            },
            "social_interaction_level": {
              "type": "string",
              "enum": ["very_high", "high", "moderate", "low", "very_low"]
            },
            "safety_perception": {
              "type": "string",
              "enum": ["very_safe", "safe", "neutral", "unsafe", "very_unsafe"]
            },
            "confidence": {"$ref": "#/definitions/confidence_score"}
          }
        },
        
        "temporal": {
          "type": "object",
          "properties": {
            "observation_time": {"type": "string", "format": "time"},
            "day_of_week": {
              "type": "string",
              "enum": ["monday", "tuesday", "wednesday", "thursday", "friday", "saturday", "sunday"]
            },
            "season": {
              "type": "string",
              "enum": ["spring", "summer", "autumn", "winter"]
            },
            "weather": {
              "type": "string",
              "enum": ["sunny", "cloudy", "rainy", "stormy", "hot", "cold", "mild"]
            },
            "time_pattern": {
              "type": "string",
              "enum": ["rush_hour_morning", "midday", "rush_hour_evening", "night", "weekend", "holiday"]
            },
            "seasonal_indicator": {
              "type": "string",
              "enum": ["peak_season", "shoulder_season", "low_season", "normal"]
            },
            "economic_cycle_indicator": {
              "type": "string",
              "enum": ["expansion", "peak", "contraction", "trough", "unknown"]
            },
            "confidence": {"$ref": "#/definitions/confidence_score"}
          }
        }
      }
    },
    
    "ai_analysis": {
      "type": "object",
      "properties": {
        "image_classification": {
          "type": "array",
          "items": {
            "type": "object",
            "properties": {
              "label": {"type": "string"},
              "confidence": {"type": "number", "minimum": 0, "maximum": 1}
            }
          }
        },
        "object_detection": {
          "type": "array",
          "items": {
            "type": "object",
            "properties": {
              "object": {"type": "string"},
              "count": {"type": "integer"},
              "confidence": {"type": "number", "minimum": 0, "maximum": 1}
            }
          }
        },
        "scene_understanding": {"type": "string"},
        "anomaly_flags": {
          "type": "array",
          "items": {"type": "string"}
        },
        "quality_score": {"type": "number", "minimum": 0, "maximum": 100}
      }
    },
    
    "metadata": {
      "type": "object",
      "properties": {
        "collection_device": {"type": "string"},
        "app_version": {"type": "string"},
        "sync_timestamp": {"$ref": "#/definitions/timestamp"},
        "review_status": {
          "type": "string",
          "enum": ["pending", "ai_reviewed", "human_reviewed", "approved", "rejected"]
        },
        "reviewer_id": {"type": "string"},
        "rejection_reason": {"type": "string"}
      }
    }
  }
}