fix(data): anchor seeded data window to the current date (#181)

app/features/demo/pipeline.py

@@ -29,7 +29,7 @@
 import time
 from collections.abc import AsyncIterator, Awaitable, Callable
 from dataclasses import dataclass, field
-from datetime import date, timedelta
+from datetime import UTC, date, datetime, timedelta
 from pathlib import Path
 from typing import Any
 
@@ -55,8 +55,10 @@
 DEMO_SCENARIO = "demo_minimal"
 DEMO_SEED_STORES = 3
 DEMO_SEED_PRODUCTS = 10
-DEMO_SEED_START = date(2024, 10, 1)
-DEMO_SEED_END = date(2024, 12, 31)
+# Seed window is anchored to *today* so the showcase always demos
+# current-looking data; it runs DEMO_SEED_SPAN_DAYS back from today (92 days
+# inclusive). Must stay >= 72 for a non-NaN backtest WAPE (see step_backtest).
+DEMO_SEED_SPAN_DAYS = 91
 
 DEMO_MODEL_TYPES: tuple[str, ...] = ("naive", "seasonal_naive", "moving_average")
 
@@ -272,6 +274,8 @@ async def step_seed(ctx: DemoContext, client: _Client) -> StepResult:
     """Seed the ``demo_minimal`` scenario (skipped when ``skip_seed`` is set)."""
     if ctx.skip_seed:
         return ("skip", "skip_seed=true (assuming a seeded database)", {})
+    seed_end = datetime.now(UTC).date()
+    seed_start = seed_end - timedelta(days=DEMO_SEED_SPAN_DAYS)
     body = await client.request(
         "seed",
         "POST",
@@ -281,8 +285,8 @@ async def step_seed(ctx: DemoContext, client: _Client) -> StepResult:
             "seed": ctx.seed,
             "stores": DEMO_SEED_STORES,
             "products": DEMO_SEED_PRODUCTS,
-            "start_date": DEMO_SEED_START.isoformat(),
-            "end_date": DEMO_SEED_END.isoformat(),
+            "start_date": seed_start.isoformat(),
+            "end_date": seed_end.isoformat(),
             "sparsity": 0.0,
             "dry_run": False,
         },

app/features/seeder/schemas.py

@@ -6,6 +6,8 @@
 
 from pydantic import BaseModel, Field, field_validator, model_validator
 
+from app.shared.seeder.config import default_seed_end_date, default_seed_start_date
+
 VALID_CHANNELS: frozenset[str] = frozenset({"in_store", "online", "click_collect", "wholesale"})
 """Allow-list for ``sales_daily.channel`` — mirrors the SQL CHECK."""
 
@@ -98,12 +100,12 @@ class GenerateParams(BaseModel):
         description="Number of products to generate",
     )
     start_date: date = Field(
-        default_factory=lambda: date(2024, 1, 1),
-        description="Start of date range",
+        default_factory=default_seed_start_date,
+        description="Start of date range (defaults to one year before today)",
     )
     end_date: date = Field(
-        default_factory=lambda: date(2024, 12, 31),
-        description="End of date range",
+        default_factory=default_seed_end_date,
+        description="End of date range (defaults to today)",
     )
     sparsity: float = Field(
         default=0.0,

app/features/seeder/service.py

@@ -4,7 +4,7 @@
 
 import time
 from dataclasses import replace
-from datetime import date, datetime
+from datetime import date, datetime, timedelta
 
 from sqlalchemy import func, select
 from sqlalchemy.ext.asyncio import AsyncSession
@@ -26,6 +26,7 @@
 from app.features.seeder import schemas
 from app.shared.seeder import DataSeeder, ScenarioPreset, SeederConfig
 from app.shared.seeder.config import (
+    DEMO_MINIMAL_SPAN_DAYS,
     BundleConfig,
     ChangepointConfig,
     ChangepointEvent,
@@ -39,6 +40,8 @@
     ReturnsConfig,
     SparsityConfig,
     SubstitutionConfig,
+    default_seed_end_date,
+    default_seed_start_date,
 )
 
 logger = get_logger(__name__)
@@ -315,14 +318,21 @@ def list_scenarios() -> list[schemas.ScenarioInfo]:
     Returns:
         List of ScenarioInfo with preset details.
     """
+    # Date ranges are anchored to *today* so the picker reflects the windows the
+    # seeder will actually produce. holiday_rush is the one exception: it is
+    # deliberately calendar-pinned to a 2024 Q4 holiday window.
+    today = default_seed_end_date()
+    year_ago = default_seed_start_date()
+    demo_start = today - timedelta(days=DEMO_MINIMAL_SPAN_DAYS)
+
     scenarios = [
         schemas.ScenarioInfo(
             name="retail_standard",
             description="Normal retail patterns with mild seasonality and linear trend",
             stores=10,
             products=50,
-            start_date=date(2024, 1, 1),
-            end_date=date(2024, 12, 31),
+            start_date=year_ago,
+            end_date=today,
         ),
         schemas.ScenarioInfo(
             name="holiday_rush",
@@ -337,40 +347,40 @@ def list_scenarios() -> list[schemas.ScenarioInfo]:
             description="Noisy, unpredictable data with frequent anomalies for robustness testing",
             stores=10,
             products=50,
-            start_date=date(2024, 1, 1),
-            end_date=date(2024, 12, 31),
+            start_date=year_ago,
+            end_date=today,
         ),
         schemas.ScenarioInfo(
             name="stockout_heavy",
             description="Frequent stockouts (25% probability) for inventory modeling",
             stores=10,
             products=50,
-            start_date=date(2024, 1, 1),
-            end_date=date(2024, 12, 31),
+            start_date=year_ago,
+            end_date=today,
         ),
         schemas.ScenarioInfo(
             name="new_launches",
             description="100 products with gradual launch ramp patterns",
             stores=10,
             products=100,
-            start_date=date(2024, 1, 1),
-            end_date=date(2024, 12, 31),
+            start_date=year_ago,
+            end_date=today,
         ),
         schemas.ScenarioInfo(
             name="sparse",
             description="50% missing combinations and random date gaps for gap handling",
             stores=10,
             products=50,
-            start_date=date(2024, 1, 1),
-            end_date=date(2024, 12, 31),
+            start_date=year_ago,
+            end_date=today,
         ),
         schemas.ScenarioInfo(
             name="demo_minimal",
             description="Tiny preset for the make demo target (3 stores x 10 products x 92 days)",
             stores=3,
             products=10,
-            start_date=date(2024, 10, 1),
-            end_date=date(2024, 12, 31),
+            start_date=demo_start,
+            end_date=today,
         ),
     ]
 

app/features/seeder/tests/test_service.py

@@ -1,11 +1,12 @@
 """Unit tests for seeder service layer."""
 
-from datetime import date
+from datetime import date, timedelta
 from unittest.mock import AsyncMock, MagicMock, patch
 
 import pytest
 
 from app.features.seeder import schemas, service
+from app.shared.seeder.config import DEMO_MINIMAL_SPAN_DAYS, default_seed_end_date
 
 
 class TestListScenarios:
@@ -27,14 +28,18 @@ def test_returns_all_scenarios(self):
         assert "demo_minimal" in names
 
     def test_demo_minimal_dimensions(self):
-        """Test that demo_minimal is the small preset for the make demo target."""
+        """Test that demo_minimal is the small preset for the make demo target.
+
+        The window is anchored to *today* so the scenario picker reflects what
+        the seeder will actually produce.
+        """
         scenarios = service.list_scenarios()
 
         demo = next(s for s in scenarios if s.name == "demo_minimal")
         assert demo.stores == 3
         assert demo.products == 10
-        assert demo.start_date == date(2024, 10, 1)
-        assert demo.end_date == date(2024, 12, 31)
+        assert demo.end_date == default_seed_end_date()
+        assert demo.start_date == default_seed_end_date() - timedelta(days=DEMO_MINIMAL_SPAN_DAYS)
 
     def test_scenario_info_structure(self):
         """Test that scenarios have required fields."""

app/shared/seeder/config.py

@@ -3,10 +3,30 @@
 from __future__ import annotations
 
 from dataclasses import dataclass, field
-from datetime import date
+from datetime import UTC, date, datetime, timedelta
 from enum import Enum
 from typing import Literal
 
+DEFAULT_SEED_SPAN_DAYS = 365
+"""Span of the default seeded window. Generated data ends *today* and runs
+this many days backwards, so datasets, forecasts, and the demo showcase
+always look current instead of being frozen in a hard-coded calendar year."""
+
+DEMO_MINIMAL_SPAN_DAYS = 91
+"""Span of the ``demo_minimal`` window (92 calendar days inclusive). Must stay
+>= 72 so an expanding backtest with n_splits=3 + horizon=14 +
+min_train_size=30 produces a non-NaN WAPE."""
+
+
+def default_seed_end_date() -> date:
+    """End of the default seeded window — anchored to the current (UTC) date."""
+    return datetime.now(UTC).date()
+
+
+def default_seed_start_date() -> date:
+    """Start of the default seeded window — ``DEFAULT_SEED_SPAN_DAYS`` before today."""
+    return datetime.now(UTC).date() - timedelta(days=DEFAULT_SEED_SPAN_DAYS)
+
 
 class ScenarioPreset(str, Enum):
     """Pre-built scenario presets for common testing needs."""
@@ -472,8 +492,8 @@ class SeederConfig:
     """
 
     seed: int = 42
-    start_date: date = field(default_factory=lambda: date(2024, 1, 1))
-    end_date: date = field(default_factory=lambda: date(2024, 12, 31))
+    start_date: date = field(default_factory=default_seed_start_date)
+    end_date: date = field(default_factory=default_seed_end_date)
     dimensions: DimensionConfig = field(default_factory=DimensionConfig)
     time_series: TimeSeriesConfig = field(default_factory=TimeSeriesConfig)
     retail: RetailPatternConfig = field(default_factory=RetailPatternConfig)
@@ -519,6 +539,10 @@ def from_scenario(cls, scenario: ScenarioPreset, seed: int = 42) -> SeederConfig
             )
 
         if scenario == ScenarioPreset.HOLIDAY_RUSH:
+            # Deliberately calendar-pinned: the holiday dates and Q4 monthly
+            # seasonality below model a specific 2024 Black Friday / Christmas
+            # window, so this scenario is NOT re-anchored to today. Pass an
+            # explicit start_date/end_date to shift it.
             return cls(
                 seed=seed,
                 start_date=date(2024, 10, 1),
@@ -606,14 +630,17 @@ def from_scenario(cls, scenario: ScenarioPreset, seed: int = 42) -> SeederConfig
             )
 
         if scenario == ScenarioPreset.DEMO_MINIMAL:
-            # Tiny preset for the `make demo` target. Keeps wall-clock comfortable
-            # on a developer laptop while still producing a non-NaN backtest WAPE
-            # with strategy=expanding, n_splits=3, horizon=14, min_train_size=30
-            # (needs >= 30 + 3*14 = 72 days; 92 days here leaves margin).
+            # Tiny preset for the `make demo` target. Anchored to *today* so the
+            # showcase always demos current-looking data; the window runs
+            # DEMO_MINIMAL_SPAN_DAYS back from today (92 days inclusive). Keeps
+            # wall-clock comfortable on a developer laptop while still producing
+            # a non-NaN backtest WAPE with strategy=expanding, n_splits=3,
+            # horizon=14, min_train_size=30 (needs >= 30 + 3*14 = 72 days).
+            demo_end = default_seed_end_date()
             return cls(
                 seed=seed,
-                start_date=date(2024, 10, 1),
-                end_date=date(2024, 12, 31),
+                start_date=demo_end - timedelta(days=DEMO_MINIMAL_SPAN_DAYS),
+                end_date=demo_end,
                 dimensions=DimensionConfig(stores=3, products=10),
                 time_series=TimeSeriesConfig(
                     base_demand=100,

app/shared/seeder/tests/test_config.py

@@ -1,11 +1,14 @@
 """Tests for seeder configuration."""
 
-from datetime import date
+from datetime import date, timedelta
 
 from app.shared.seeder.config import (
+    DEMO_MINIMAL_SPAN_DAYS,
     ScenarioPreset,
     SeederConfig,
     TimeSeriesConfig,
+    default_seed_end_date,
+    default_seed_start_date,
 )
 
 
@@ -39,12 +42,16 @@ class TestSeederConfig:
     """Tests for SeederConfig."""
 
     def test_default_values(self):
-        """Test default configuration values."""
+        """Test default configuration values.
+
+        The default date window is anchored to *today* and runs
+        DEFAULT_SEED_SPAN_DAYS backwards, so seeded data always looks current.
+        """
         config = SeederConfig()
 
         assert config.seed == 42
-        assert config.start_date == date(2024, 1, 1)
-        assert config.end_date == date(2024, 12, 31)
+        assert config.end_date == default_seed_end_date()
+        assert config.start_date == default_seed_start_date()
         assert config.dimensions.stores == 10
         assert config.dimensions.products == 50
         assert config.batch_size == 1000
@@ -93,15 +100,16 @@ def test_from_scenario_sparse(self):
     def test_from_scenario_demo_minimal(self):
         """Test demo_minimal scenario preset.
 
-        This preset powers the `make demo` target; the date range MUST cover at
-        least 72 days so an expanding backtest with n_splits=3 + horizon=14 +
-        min_train_size=30 produces non-NaN WAPE.
+        This preset powers the `make demo` target; the window is anchored to
+        *today* and MUST cover at least 72 days so an expanding backtest with
+        n_splits=3 + horizon=14 + min_train_size=30 produces non-NaN WAPE.
         """
         config = SeederConfig.from_scenario(ScenarioPreset.DEMO_MINIMAL, seed=42)
 
         assert config.seed == 42
-        assert config.start_date == date(2024, 10, 1)
-        assert config.end_date == date(2024, 12, 31)
+        assert config.end_date == default_seed_end_date()
+        assert config.start_date == default_seed_end_date() - timedelta(days=DEMO_MINIMAL_SPAN_DAYS)
+        assert (config.end_date - config.start_date).days >= 72
         assert config.dimensions.stores == 3
         assert config.dimensions.products == 10
         assert config.time_series.trend == "linear"

app/shared/seeder/tests/test_core.py

@@ -5,7 +5,12 @@
 
 import pytest
 
-from app.shared.seeder.config import SeederConfig, SparsityConfig
+from app.shared.seeder.config import (
+    SeederConfig,
+    SparsityConfig,
+    default_seed_end_date,
+    default_seed_start_date,
+)
 from app.shared.seeder.core import DataSeeder, SeederResult
 
 
@@ -402,12 +407,12 @@ def test_default_dimensions(self):
         assert seeder.config.dimensions.products == 50
 
     def test_default_date_range(self):
-        """Test default date range is full year 2024."""
+        """Test the default date range is anchored to today and runs backwards."""
         config = SeederConfig()
         seeder = DataSeeder(config)
 
-        assert seeder.config.start_date == date(2024, 1, 1)
-        assert seeder.config.end_date == date(2024, 12, 31)
+        assert seeder.config.start_date == default_seed_start_date()
+        assert seeder.config.end_date == default_seed_end_date()
 
     def test_custom_sparsity(self):
         """Test custom sparsity configuration."""