Hostaway MCP Server

"""Field projection utility for response summarization. Extracts essential fields from objects to reduce token usage. Based on research.md R003: Field projection (primary) + extractive (secondary). """ from typing import Any def project_fields(obj: dict[str, Any], fields: list[str]) -> dict[str, Any]: """Project specified fields from an object. Args: obj: Source object (dictionary) fields: List of field names to extract (supports nested paths with dot notation) Returns: New dict containing only the specified fields Example: >>> booking = { ... "id": "BK12345", ... "status": "confirmed", ... "guestName": "John Doe", ... "guestEmail": "john@example.com", ... "guestAddress": {"city": "NYC", "state": "NY"}, ... "priceBreakdown": {...} # 15 fields ... } >>> project_fields(booking, ["id", "status", "guestName", "guestAddress.city"]) {'id': 'BK12345', 'status': 'confirmed', 'guestName': 'John Doe', 'guestAddress': {'city': 'NYC'}} """ result: dict[str, Any] = {} for field in fields: # Handle nested field paths (e.g., "guestAddress.city") if "." in field: parts = field.split(".") value = obj for part in parts[:-1]: value = value.get(part, {}) if not isinstance(value, dict): break else: # Successfully traversed to parent final_key = parts[-1] if isinstance(value, dict) and final_key in value: # Build nested structure in result current = result for part in parts[:-1]: if part not in current: current[part] = {} current = current[part] current[final_key] = value[final_key] # Simple field elif field in obj: result[field] = obj[field] return result def get_essential_fields(obj_type: str) -> list[str]: """Get essential field list for common object types. Args: obj_type: Type identifier (e.g., "booking", "listing", "financial_transaction") Returns: List of essential field names for summarization Example: >>> get_essential_fields("booking") ['id', 'status', 'guestName', 'checkInDate', 'checkOutDate', 'totalPrice', 'propertyId'] """ # Essential fields per object type (based on data-model.md examples) field_map: dict[str, list[str]] = { "booking": [ "id", "status", "guestName", "checkInDate", "checkOutDate", "totalPrice", "currency", "propertyId", ], "listing": [ "id", "name", "status", "address", "maxGuests", "bedroomCount", "bathroomCount", "basePrice", ], "financial_transaction": [ "id", "type", "amount", "currency", "status", "bookingId", "createdAt", ], "conversation": [ "id", "bookingId", "guestName", "lastMessageAt", "unreadCount", "status", ], "review": [ "id", "bookingId", "rating", "comment", "createdAt", "guestName", ], } return field_map.get(obj_type, ["id"]) # Fallback to just ID def estimate_field_count(obj: dict[str, Any]) -> int: """Recursively count total fields in an object. Args: obj: Object to analyze Returns: Total number of fields (including nested) Example: >>> booking = {"id": "BK12345", "guestAddress": {"city": "NYC", "state": "NY"}} >>> estimate_field_count(booking) 4 # id, guestAddress, guestAddress.city, guestAddress.state """ count = 0 for value in obj.values(): count += 1 if isinstance(value, dict): count += estimate_field_count(value) elif isinstance(value, list) and value and isinstance(value[0], dict): # For lists, count fields in first item (assume homogeneous) count += estimate_field_count(value[0]) return count def calculate_projection_ratio( original: dict[str, Any], projected: dict[str, Any], ) -> float: """Calculate field reduction ratio after projection. Args: original: Original object projected: Projected object Returns: Ratio of fields retained (0.0 to 1.0) Example: >>> original = {"id": "BK12345", "field1": 1, "field2": 2, "field3": 3} >>> projected = {"id": "BK12345", "field1": 1} >>> calculate_projection_ratio(original, projected) 0.5 # 2 out of 4 fields retained """ original_count = estimate_field_count(original) projected_count = estimate_field_count(projected) if original_count == 0: return 1.0 return projected_count / original_count