feat: add Pydantic schemas, SQLAlchemy models, and parser structure

Sprint 3 implementation:
- Pydantic schemas for TRREB, CMHC, and dimension data validation
- SQLAlchemy models with PostGIS geometry for fact and dimension tables
- Parser structure (stubs) for TRREB PDF and CMHC CSV processing

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

portfolio_app/toronto/models/__init__.py

@@ -1 +1,28 @@
 """SQLAlchemy models for Toronto housing data."""
+
+from .base import Base, create_tables, get_engine, get_session_factory
+from .dimensions import (
+    DimCMHCZone,
+    DimNeighbourhood,
+    DimPolicyEvent,
+    DimTime,
+    DimTRREBDistrict,
+)
+from .facts import FactPurchases, FactRentals
+
+__all__ = [
+    # Base
+    "Base",
+    "get_engine",
+    "get_session_factory",
+    "create_tables",
+    # Dimensions
+    "DimTime",
+    "DimTRREBDistrict",
+    "DimCMHCZone",
+    "DimNeighbourhood",
+    "DimPolicyEvent",
+    # Facts
+    "FactPurchases",
+    "FactRentals",
+]

portfolio_app/toronto/models/base.py

@@ -0,0 +1,30 @@
+"""SQLAlchemy base configuration and engine setup."""
+
+from sqlalchemy import Engine, create_engine
+from sqlalchemy.orm import DeclarativeBase, Session, sessionmaker
+
+from portfolio_app.config import get_settings
+
+
+class Base(DeclarativeBase):  # type: ignore[misc]
+    """Base class for all SQLAlchemy models."""
+
+    pass
+
+
+def get_engine() -> Engine:
+    """Create database engine from settings."""
+    settings = get_settings()
+    return create_engine(settings.database_url, echo=False)
+
+
+def get_session_factory() -> sessionmaker[Session]:
+    """Create session factory."""
+    engine = get_engine()
+    return sessionmaker(bind=engine)
+
+
+def create_tables() -> None:
+    """Create all tables in database."""
+    engine = get_engine()
+    Base.metadata.create_all(engine)

portfolio_app/toronto/models/dimensions.py

@@ -0,0 +1,104 @@
+"""SQLAlchemy models for dimension tables."""
+
+from datetime import date
+
+from geoalchemy2 import Geometry
+from sqlalchemy import Boolean, Date, Integer, Numeric, String, Text
+from sqlalchemy.orm import Mapped, mapped_column
+
+from .base import Base
+
+
+class DimTime(Base):
+    """Time dimension table."""
+
+    __tablename__ = "dim_time"
+
+    date_key: Mapped[int] = mapped_column(Integer, primary_key=True)
+    full_date: Mapped[date] = mapped_column(Date, nullable=False, unique=True)
+    year: Mapped[int] = mapped_column(Integer, nullable=False)
+    month: Mapped[int] = mapped_column(Integer, nullable=False)
+    quarter: Mapped[int] = mapped_column(Integer, nullable=False)
+    month_name: Mapped[str] = mapped_column(String(20), nullable=False)
+    is_month_start: Mapped[bool] = mapped_column(Boolean, default=True)
+
+
+class DimTRREBDistrict(Base):
+    """TRREB district dimension table with PostGIS geometry."""
+
+    __tablename__ = "dim_trreb_district"
+
+    district_key: Mapped[int] = mapped_column(
+        Integer, primary_key=True, autoincrement=True
+    )
+    district_code: Mapped[str] = mapped_column(String(3), nullable=False, unique=True)
+    district_name: Mapped[str] = mapped_column(String(100), nullable=False)
+    area_type: Mapped[str] = mapped_column(String(10), nullable=False)
+    geometry = mapped_column(Geometry("POLYGON", srid=4326), nullable=True)
+
+
+class DimCMHCZone(Base):
+    """CMHC zone dimension table with PostGIS geometry."""
+
+    __tablename__ = "dim_cmhc_zone"
+
+    zone_key: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
+    zone_code: Mapped[str] = mapped_column(String(10), nullable=False, unique=True)
+    zone_name: Mapped[str] = mapped_column(String(100), nullable=False)
+    geometry = mapped_column(Geometry("POLYGON", srid=4326), nullable=True)
+
+
+class DimNeighbourhood(Base):
+    """City of Toronto neighbourhood dimension.
+
+    Note: No FK to fact tables in V1 - reference overlay only.
+    """
+
+    __tablename__ = "dim_neighbourhood"
+
+    neighbourhood_id: Mapped[int] = mapped_column(Integer, primary_key=True)
+    name: Mapped[str] = mapped_column(String(100), nullable=False)
+    geometry = mapped_column(Geometry("POLYGON", srid=4326), nullable=True)
+    population: Mapped[int | None] = mapped_column(Integer, nullable=True)
+    land_area_sqkm: Mapped[float | None] = mapped_column(Numeric(10, 4), nullable=True)
+    pop_density_per_sqkm: Mapped[float | None] = mapped_column(
+        Numeric(10, 2), nullable=True
+    )
+    pct_bachelors_or_higher: Mapped[float | None] = mapped_column(
+        Numeric(5, 2), nullable=True
+    )
+    median_household_income: Mapped[float | None] = mapped_column(
+        Numeric(12, 2), nullable=True
+    )
+    pct_owner_occupied: Mapped[float | None] = mapped_column(
+        Numeric(5, 2), nullable=True
+    )
+    pct_renter_occupied: Mapped[float | None] = mapped_column(
+        Numeric(5, 2), nullable=True
+    )
+    census_year: Mapped[int] = mapped_column(Integer, default=2021)
+
+
+class DimPolicyEvent(Base):
+    """Policy event dimension for time-series annotation."""
+
+    __tablename__ = "dim_policy_event"
+
+    event_id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
+    event_date: Mapped[date] = mapped_column(Date, nullable=False)
+    effective_date: Mapped[date | None] = mapped_column(Date, nullable=True)
+    level: Mapped[str] = mapped_column(
+        String(20), nullable=False
+    )  # federal/provincial/municipal
+    category: Mapped[str] = mapped_column(
+        String(20), nullable=False
+    )  # monetary/tax/regulatory/supply/economic
+    title: Mapped[str] = mapped_column(String(200), nullable=False)
+    description: Mapped[str | None] = mapped_column(Text, nullable=True)
+    expected_direction: Mapped[str] = mapped_column(
+        String(10), nullable=False
+    )  # bearish/bullish/neutral
+    source_url: Mapped[str | None] = mapped_column(String(500), nullable=True)
+    confidence: Mapped[str] = mapped_column(
+        String(10), default="medium"
+    )  # high/medium/low

portfolio_app/toronto/models/facts.py

@@ -0,0 +1,69 @@
+"""SQLAlchemy models for fact tables."""
+
+from sqlalchemy import ForeignKey, Integer, Numeric, String
+from sqlalchemy.orm import Mapped, mapped_column, relationship
+
+from .base import Base
+
+
+class FactPurchases(Base):
+    """Fact table for TRREB purchase/sales data.
+
+    Grain: One row per district per month.
+    """
+
+    __tablename__ = "fact_purchases"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
+    date_key: Mapped[int] = mapped_column(
+        Integer, ForeignKey("dim_time.date_key"), nullable=False
+    )
+    district_key: Mapped[int] = mapped_column(
+        Integer, ForeignKey("dim_trreb_district.district_key"), nullable=False
+    )
+    sales_count: Mapped[int] = mapped_column(Integer, nullable=False)
+    dollar_volume: Mapped[float] = mapped_column(Numeric(15, 2), nullable=False)
+    avg_price: Mapped[float] = mapped_column(Numeric(12, 2), nullable=False)
+    median_price: Mapped[float] = mapped_column(Numeric(12, 2), nullable=False)
+    new_listings: Mapped[int] = mapped_column(Integer, nullable=False)
+    active_listings: Mapped[int] = mapped_column(Integer, nullable=False)
+    avg_dom: Mapped[int] = mapped_column(Integer, nullable=False)  # Days on market
+    avg_sp_lp: Mapped[float] = mapped_column(
+        Numeric(5, 2), nullable=False
+    )  # Sale/List ratio
+
+    # Relationships
+    time = relationship("DimTime", backref="purchases")
+    district = relationship("DimTRREBDistrict", backref="purchases")
+
+
+class FactRentals(Base):
+    """Fact table for CMHC rental market data.
+
+    Grain: One row per zone per bedroom type per survey year.
+    """
+
+    __tablename__ = "fact_rentals"
+
+    id: Mapped[int] = mapped_column(Integer, primary_key=True, autoincrement=True)
+    date_key: Mapped[int] = mapped_column(
+        Integer, ForeignKey("dim_time.date_key"), nullable=False
+    )
+    zone_key: Mapped[int] = mapped_column(
+        Integer, ForeignKey("dim_cmhc_zone.zone_key"), nullable=False
+    )
+    bedroom_type: Mapped[str] = mapped_column(String(20), nullable=False)
+    universe: Mapped[int | None] = mapped_column(Integer, nullable=True)
+    avg_rent: Mapped[float | None] = mapped_column(Numeric(10, 2), nullable=True)
+    median_rent: Mapped[float | None] = mapped_column(Numeric(10, 2), nullable=True)
+    vacancy_rate: Mapped[float | None] = mapped_column(Numeric(5, 2), nullable=True)
+    availability_rate: Mapped[float | None] = mapped_column(
+        Numeric(5, 2), nullable=True
+    )
+    turnover_rate: Mapped[float | None] = mapped_column(Numeric(5, 2), nullable=True)
+    rent_change_pct: Mapped[float | None] = mapped_column(Numeric(5, 2), nullable=True)
+    reliability_code: Mapped[str | None] = mapped_column(String(2), nullable=True)
+
+    # Relationships
+    time = relationship("DimTime", backref="rentals")
+    zone = relationship("DimCMHCZone", backref="rentals")

portfolio_app/toronto/parsers/__init__.py

@@ -1 +1,9 @@
-"""Data parsers for Toronto housing data sources."""
+"""Parsers for Toronto housing data sources."""
+
+from .cmhc import CMHCParser
+from .trreb import TRREBParser
+
+__all__ = [
+    "TRREBParser",
+    "CMHCParser",
+]

portfolio_app/toronto/parsers/cmhc.py

@@ -0,0 +1,147 @@
+"""CMHC CSV processor for rental market survey data.
+
+This module provides the structure for processing CMHC (Canada Mortgage and Housing
+Corporation) rental market survey data from CSV exports.
+"""
+
+from pathlib import Path
+from typing import Any, cast
+
+import pandas as pd
+
+from portfolio_app.toronto.schemas import CMHCAnnualSurvey, CMHCRentalRecord
+
+
+class CMHCParser:
+    """Parser for CMHC Rental Market Survey CSV data.
+
+    CMHC conducts annual rental market surveys and publishes data including:
+    - Average and median rents by zone and bedroom type
+    - Vacancy rates
+    - Universe (total rental units)
+    - Year-over-year rent changes
+
+    Data is available via the Housing Market Information Portal as CSV exports.
+    """
+
+    # Expected columns in CMHC CSV exports
+    REQUIRED_COLUMNS = {
+        "zone_code",
+        "zone_name",
+        "bedroom_type",
+        "survey_year",
+    }
+
+    # Column name mappings from CMHC export format
+    COLUMN_MAPPINGS = {
+        "Zone Code": "zone_code",
+        "Zone Name": "zone_name",
+        "Bedroom Type": "bedroom_type",
+        "Survey Year": "survey_year",
+        "Universe": "universe",
+        "Average Rent ($)": "avg_rent",
+        "Median Rent ($)": "median_rent",
+        "Vacancy Rate (%)": "vacancy_rate",
+        "Availability Rate (%)": "availability_rate",
+        "Turnover Rate (%)": "turnover_rate",
+        "% Change in Rent": "rent_change_pct",
+        "Reliability Code": "reliability_code",
+    }
+
+    def __init__(self, csv_path: Path) -> None:
+        """Initialize parser with path to CSV file.
+
+        Args:
+            csv_path: Path to the CMHC CSV export file.
+        """
+        self.csv_path = csv_path
+        self._validate_path()
+
+    def _validate_path(self) -> None:
+        """Validate that the CSV path exists and is readable."""
+        if not self.csv_path.exists():
+            raise FileNotFoundError(f"CSV not found: {self.csv_path}")
+        if not self.csv_path.suffix.lower() == ".csv":
+            raise ValueError(f"Expected CSV file, got: {self.csv_path.suffix}")
+
+    def parse(self) -> CMHCAnnualSurvey:
+        """Parse the CSV and return structured data.
+
+        Returns:
+            CMHCAnnualSurvey containing all extracted records.
+
+        Raises:
+            ValueError: If required columns are missing.
+        """
+        df = self._load_csv()
+        df = self._normalize_columns(df)
+        self._validate_columns(df)
+        records = self._convert_to_records(df)
+        survey_year = self._infer_survey_year(df)
+
+        return CMHCAnnualSurvey(survey_year=survey_year, records=records)
+
+    def _load_csv(self) -> pd.DataFrame:
+        """Load CSV file into DataFrame.
+
+        Returns:
+            Raw DataFrame from CSV.
+        """
+        return pd.read_csv(self.csv_path)
+
+    def _normalize_columns(self, df: pd.DataFrame) -> pd.DataFrame:
+        """Normalize column names to standard format.
+
+        Args:
+            df: DataFrame with original column names.
+
+        Returns:
+            DataFrame with normalized column names.
+        """
+        rename_map = {k: v for k, v in self.COLUMN_MAPPINGS.items() if k in df.columns}
+        return df.rename(columns=rename_map)
+
+    def _validate_columns(self, df: pd.DataFrame) -> None:
+        """Validate that all required columns are present.
+
+        Args:
+            df: DataFrame to validate.
+
+        Raises:
+            ValueError: If required columns are missing.
+        """
+        missing = self.REQUIRED_COLUMNS - set(df.columns)
+        if missing:
+            raise ValueError(f"Missing required columns: {missing}")
+
+    def _convert_to_records(self, df: pd.DataFrame) -> list[CMHCRentalRecord]:
+        """Convert DataFrame rows to validated schema records.
+
+        Args:
+            df: Normalized DataFrame.
+
+        Returns:
+            List of validated CMHCRentalRecord objects.
+        """
+        records = []
+        for _, row in df.iterrows():
+            record_data = row.to_dict()
+            # Handle NaN values
+            record_data = {
+                k: (None if pd.isna(v) else v) for k, v in record_data.items()
+            }
+            records.append(CMHCRentalRecord(**cast(dict[str, Any], record_data)))
+        return records
+
+    def _infer_survey_year(self, df: pd.DataFrame) -> int:
+        """Infer survey year from data.
+
+        Args:
+            df: DataFrame with survey_year column.
+
+        Returns:
+            Survey year as integer.
+        """
+        if "survey_year" in df.columns:
+            return int(df["survey_year"].iloc[0])
+        raise ValueError("Cannot infer survey year from data.")

portfolio_app/toronto/parsers/trreb.py

@@ -0,0 +1,82 @@
+"""TRREB PDF parser for monthly market watch reports.
+
+This module provides the structure for parsing TRREB (Toronto Regional Real Estate Board)
+monthly Market Watch PDF reports into structured data.
+"""
+
+from pathlib import Path
+from typing import Any
+
+from portfolio_app.toronto.schemas import TRREBMonthlyRecord, TRREBMonthlyReport
+
+
+class TRREBParser:
+    """Parser for TRREB Market Watch PDF reports.
+
+    TRREB publishes monthly Market Watch reports as PDFs containing:
+    - Summary statistics by area (416, 905, Total)
+    - District-level breakdowns
+    - Year-over-year comparisons
+
+    The parser extracts tabular data from these PDFs and validates
+    against the TRREBMonthlyRecord schema.
+    """
+
+    def __init__(self, pdf_path: Path) -> None:
+        """Initialize parser with path to PDF file.
+
+        Args:
+            pdf_path: Path to the TRREB Market Watch PDF file.
+        """
+        self.pdf_path = pdf_path
+        self._validate_path()
+
+    def _validate_path(self) -> None:
+        """Validate that the PDF path exists and is readable."""
+        if not self.pdf_path.exists():
+            raise FileNotFoundError(f"PDF not found: {self.pdf_path}")
+        if not self.pdf_path.suffix.lower() == ".pdf":
+            raise ValueError(f"Expected PDF file, got: {self.pdf_path.suffix}")
+
+    def parse(self) -> TRREBMonthlyReport:
+        """Parse the PDF and return structured data.
+
+        Returns:
+            TRREBMonthlyReport containing all extracted records.
+
+        Raises:
+            NotImplementedError: PDF parsing not yet implemented.
+        """
+        raise NotImplementedError(
+            "PDF parsing requires pdfplumber/tabula-py. "
+            "Implementation pending Sprint 4 data ingestion."
+        )
+
+    def _extract_tables(self) -> list[dict[str, Any]]:
+        """Extract raw tables from PDF pages.
+
+        Returns:
+            List of dictionaries representing table data.
+        """
+        raise NotImplementedError("Table extraction not yet implemented.")
+
+    def _parse_district_table(
+        self, table_data: list[dict[str, Any]]
+    ) -> list[TRREBMonthlyRecord]:
+        """Parse district-level statistics table.
+
+        Args:
+            table_data: Raw table data extracted from PDF.
+
+        Returns:
+            List of validated TRREBMonthlyRecord objects.
+        """
+        raise NotImplementedError("District table parsing not yet implemented.")
+
+    def _infer_report_date(self) -> tuple[int, int]:
+        """Infer report year and month from PDF filename or content.
+
+        Returns:
+            Tuple of (year, month).
+        """
+        raise NotImplementedError("Date inference not yet implemented.")

portfolio_app/toronto/schemas/__init__.py

@@ -1 +1,39 @@
 """Pydantic schemas for Toronto housing data validation."""
+
+from .cmhc import BedroomType, CMHCAnnualSurvey, CMHCRentalRecord, ReliabilityCode
+from .dimensions import (
+    AreaType,
+    CMHCZone,
+    Confidence,
+    ExpectedDirection,
+    Neighbourhood,
+    PolicyCategory,
+    PolicyEvent,
+    PolicyLevel,
+    TimeDimension,
+    TRREBDistrict,
+)
+from .trreb import TRREBMonthlyRecord, TRREBMonthlyReport
+
+__all__ = [
+    # TRREB
+    "TRREBMonthlyRecord",
+    "TRREBMonthlyReport",
+    # CMHC
+    "CMHCRentalRecord",
+    "CMHCAnnualSurvey",
+    "BedroomType",
+    "ReliabilityCode",
+    # Dimensions
+    "TimeDimension",
+    "TRREBDistrict",
+    "CMHCZone",
+    "Neighbourhood",
+    "PolicyEvent",
+    # Enums
+    "AreaType",
+    "PolicyLevel",
+    "PolicyCategory",
+    "ExpectedDirection",
+    "Confidence",
+]

portfolio_app/toronto/schemas/cmhc.py

@@ -0,0 +1,81 @@
+"""Pydantic schemas for CMHC rental market data."""
+
+from decimal import Decimal
+from enum import Enum
+
+from pydantic import BaseModel, Field
+
+
+class BedroomType(str, Enum):
+    """CMHC bedroom type categories."""
+
+    BACHELOR = "Bachelor"
+    ONE_BED = "1 Bedroom"
+    TWO_BED = "2 Bedroom"
+    THREE_BED_PLUS = "3 Bedroom+"
+    TOTAL = "Total"
+
+
+class ReliabilityCode(str, Enum):
+    """CMHC data reliability codes.
+
+    Based on coefficient of variation (CV).
+    """
+
+    EXCELLENT = "a"  # CV <= 2.5%
+    GOOD = "b"  # 2.5% < CV <= 5%
+    FAIR = "c"  # 5% < CV <= 10%
+    POOR = "d"  # CV > 10%
+    SUPPRESSED = "**"  # Sample too small
+
+
+class CMHCRentalRecord(BaseModel):
+    """Schema for a single CMHC rental survey record.
+
+    Represents rental data for one zone and bedroom type in one survey year.
+    """
+
+    survey_year: int = Field(ge=1990, description="Survey year (October snapshot)")
+    zone_code: str = Field(max_length=10, description="CMHC zone identifier")
+    zone_name: str = Field(max_length=100, description="Zone name")
+    bedroom_type: BedroomType = Field(description="Bedroom category")
+    universe: int | None = Field(
+        default=None, ge=0, description="Total rental units in zone"
+    )
+    vacancy_rate: Decimal | None = Field(
+        default=None, ge=0, le=100, description="Vacancy rate (%)"
+    )
+    vacancy_rate_reliability: ReliabilityCode | None = Field(default=None)
+    availability_rate: Decimal | None = Field(
+        default=None, ge=0, le=100, description="Availability rate (%)"
+    )
+    average_rent: Decimal | None = Field(
+        default=None, ge=0, description="Average monthly rent ($)"
+    )
+    average_rent_reliability: ReliabilityCode | None = Field(default=None)
+    median_rent: Decimal | None = Field(
+        default=None, ge=0, description="Median monthly rent ($)"
+    )
+    rent_change_pct: Decimal | None = Field(
+        default=None, description="YoY rent change (%)"
+    )
+    turnover_rate: Decimal | None = Field(
+        default=None, ge=0, le=100, description="Unit turnover rate (%)"
+    )
+
+    model_config = {"str_strip_whitespace": True}
+
+
+class CMHCAnnualSurvey(BaseModel):
+    """Schema for a complete CMHC annual survey for Toronto.
+
+    Contains all zone and bedroom type combinations for one survey year.
+    """
+
+    survey_year: int
+    records: list[CMHCRentalRecord]
+
+    @property
+    def zone_count(self) -> int:
+        """Number of unique zones in survey."""
+        return len({r.zone_code for r in self.records})

portfolio_app/toronto/schemas/dimensions.py

@@ -0,0 +1,121 @@
+"""Pydantic schemas for dimension tables."""
+
+from datetime import date
+from decimal import Decimal
+from enum import Enum
+
+from pydantic import BaseModel, Field, HttpUrl
+
+
+class PolicyLevel(str, Enum):
+    """Government level for policy events."""
+
+    FEDERAL = "federal"
+    PROVINCIAL = "provincial"
+    MUNICIPAL = "municipal"
+
+
+class PolicyCategory(str, Enum):
+    """Policy event category."""
+
+    MONETARY = "monetary"
+    TAX = "tax"
+    REGULATORY = "regulatory"
+    SUPPLY = "supply"
+    ECONOMIC = "economic"
+
+
+class ExpectedDirection(str, Enum):
+    """Expected price impact direction."""
+
+    BULLISH = "bullish"  # Expected to increase prices
+    BEARISH = "bearish"  # Expected to decrease prices
+    NEUTRAL = "neutral"  # Uncertain or mixed impact
+
+
+class Confidence(str, Enum):
+    """Confidence level in policy event data."""
+
+    HIGH = "high"
+    MEDIUM = "medium"
+    LOW = "low"
+
+
+class AreaType(str, Enum):
+    """TRREB area type."""
+
+    WEST = "West"
+    CENTRAL = "Central"
+    EAST = "East"
+    NORTH = "North"
+
+
+class TimeDimension(BaseModel):
+    """Schema for time dimension record."""
+
+    date_key: int = Field(description="Date key in YYYYMMDD format")
+    full_date: date
+    year: int = Field(ge=2000, le=2100)
+    month: int = Field(ge=1, le=12)
+    quarter: int = Field(ge=1, le=4)
+    month_name: str = Field(max_length=20)
+    is_month_start: bool = True
+
+
+class TRREBDistrict(BaseModel):
+    """Schema for TRREB district dimension."""
+
+    district_code: str = Field(max_length=3, description="W01, C01, E01, etc.")
+    district_name: str = Field(max_length=100)
+    area_type: AreaType
+    geometry_wkt: str | None = Field(default=None, description="WKT geometry string")
+
+
+class CMHCZone(BaseModel):
+    """Schema for CMHC zone dimension."""
+
+    zone_code: str = Field(max_length=10)
+    zone_name: str = Field(max_length=100)
+    geometry_wkt: str | None = Field(default=None, description="WKT geometry string")
+
+
+class Neighbourhood(BaseModel):
+    """Schema for City of Toronto neighbourhood dimension.
+
+    Note: No FK to fact tables in V1 - reference overlay only.
+    """
+
+    neighbourhood_id: int = Field(ge=1, le=200)
+    name: str = Field(max_length=100)
+    geometry_wkt: str | None = Field(default=None)
+    population: int | None = Field(default=None, ge=0)
+    land_area_sqkm: Decimal | None = Field(default=None, ge=0)
+    pop_density_per_sqkm: Decimal | None = Field(default=None, ge=0)
+    pct_bachelors_or_higher: Decimal | None = Field(default=None, ge=0, le=100)
+    median_household_income: Decimal | None = Field(default=None, ge=0)
+    pct_owner_occupied: Decimal | None = Field(default=None, ge=0, le=100)
+    pct_renter_occupied: Decimal | None = Field(default=None, ge=0, le=100)
+    census_year: int = Field(default=2021, description="Census year for SCD tracking")
+
+
+class PolicyEvent(BaseModel):
+    """Schema for policy event dimension.
+
+    Used for time-series annotation. No causation claims.
+    """
+
+    event_date: date = Field(description="Date event was announced/occurred")
+    effective_date: date | None = Field(
+        default=None, description="Date policy took effect"
+    )
+    level: PolicyLevel
+    category: PolicyCategory
+    title: str = Field(max_length=200, description="Short event title for display")
+    description: str | None = Field(
+        default=None, description="Longer description for tooltip"
+    )
+    expected_direction: ExpectedDirection
+    source_url: HttpUrl | None = Field(default=None)
+    confidence: Confidence = Field(default=Confidence.MEDIUM)
+
+    model_config = {"str_strip_whitespace": True}

portfolio_app/toronto/schemas/trreb.py

@@ -0,0 +1,52 @@
+"""Pydantic schemas for TRREB monthly market data."""
+
+from datetime import date
+from decimal import Decimal
+
+from pydantic import BaseModel, Field
+
+
+class TRREBMonthlyRecord(BaseModel):
+    """Schema for a single TRREB monthly summary record.
+
+    Represents aggregated sales data for one district in one month.
+    """
+
+    report_date: date = Field(description="First of month (YYYY-MM-01)")
+    area_code: str = Field(
+        max_length=3, description="District code (W01, C01, E01, etc.)"
+    )
+    area_name: str = Field(max_length=100, description="District name")
+    area_type: str = Field(max_length=10, description="West / Central / East / North")
+    sales: int = Field(ge=0, description="Number of transactions")
+    dollar_volume: Decimal = Field(ge=0, description="Total sales volume ($)")
+    avg_price: Decimal = Field(ge=0, description="Average sale price ($)")
+    median_price: Decimal = Field(ge=0, description="Median sale price ($)")
+    new_listings: int = Field(ge=0, description="New listings count")
+    active_listings: int = Field(ge=0, description="Active listings at month end")
+    avg_sp_lp: Decimal = Field(
+        ge=0, le=200, description="Avg sale price / list price ratio (%)"
+    )
+    avg_dom: int = Field(ge=0, description="Average days on market")
+
+    model_config = {"str_strip_whitespace": True}
+
+
+class TRREBMonthlyReport(BaseModel):
+    """Schema for a complete TRREB monthly report.
+
+    Contains all district records for a single month.
+    """
+
+    report_date: date
+    records: list[TRREBMonthlyRecord]
+
+    @property
+    def total_sales(self) -> int:
+        """Total sales across all districts."""
+        return sum(r.sales for r in self.records)
+
+    @property
+    def district_count(self) -> int:
+        """Number of districts in report."""
+        return len(self.records)