Raise clear error when uprating year is outside table range (#352)

* Raise clear error when uprating year is outside table range

* Apply ruff format

Author: MaxGhenis

changelog.d/fix-uprating-end-year-guard.fixed.md

@@ -0,0 +1 @@
+Raise `UpratingYearOutOfRangeError` with a clear message when `uprate_values` or `uprate_dataset` is called with a year outside the `[START_YEAR, END_YEAR]` range of the uprating factor table, instead of surfacing a pandas `KeyError` or silently returning wrong values.

policyengine_uk_data/tests/test_uprating_range.py

@@ -0,0 +1,88 @@
+"""Tests for year-range validation in `policyengine_uk_data.utils.uprating`.
+
+The uprating factor table covers ``[START_YEAR, END_YEAR]`` (2020–2034 on
+main). Callers that request a year outside this range previously got a
+raw pandas ``KeyError`` or a silently wrong value. After the fix in
+finding U5 they must get ``UpratingYearOutOfRangeError`` with a clear
+message.
+"""
+
+from __future__ import annotations
+
+import importlib.util
+
+import pandas as pd
+import pytest
+
+if importlib.util.find_spec("policyengine_uk") is None:
+    pytest.skip(
+        "policyengine_uk not available in test environment",
+        allow_module_level=True,
+    )
+
+
+def _seed_uprating_table(tmp_path, start=2020, end=2034):
+    """Write a minimal uprating_factors.csv covering a known year range."""
+
+    storage = tmp_path / "storage"
+    storage.mkdir()
+    years = list(range(start, end + 1))
+    table = pd.DataFrame(
+        {str(y): [1.0 + 0.02 * (y - start)] for y in years},
+        index=pd.Index(["employment_income"], name="Variable"),
+    )
+    table.to_csv(storage / "uprating_factors.csv")
+    return storage
+
+
+def test_uprate_dataset_rejects_year_above_end(tmp_path, monkeypatch):
+    from policyengine_uk_data.utils import uprating as uprating_module
+    from policyengine_uk_data.utils.uprating import (
+        UpratingYearOutOfRangeError,
+        uprate_dataset,
+    )
+
+    storage = _seed_uprating_table(tmp_path)
+    monkeypatch.setattr(uprating_module, "STORAGE_FOLDER", storage)
+
+    # target_year=2035 is above END_YEAR=2034 → must raise with clear message.
+    class _Dataset:
+        time_period = 2023
+        tables = []
+
+        def copy(self):
+            return self
+
+    with pytest.raises(UpratingYearOutOfRangeError, match="target_year=2035"):
+        uprate_dataset(_Dataset(), target_year=2035)
+
+
+def test_uprate_values_rejects_years_outside_range(tmp_path, monkeypatch):
+    from policyengine_uk_data.utils import uprating as uprating_module
+    from policyengine_uk_data.utils.uprating import (
+        UpratingYearOutOfRangeError,
+        uprate_values,
+    )
+
+    storage = _seed_uprating_table(tmp_path)
+    monkeypatch.setattr(uprating_module, "STORAGE_FOLDER", storage)
+
+    with pytest.raises(UpratingYearOutOfRangeError, match="end_year=2099"):
+        uprate_values(100.0, "employment_income", start_year=2020, end_year=2099)
+
+    with pytest.raises(UpratingYearOutOfRangeError, match="start_year=1999"):
+        uprate_values(100.0, "employment_income", start_year=1999, end_year=2034)
+
+
+def test_uprate_values_accepts_supported_range(tmp_path, monkeypatch):
+    from policyengine_uk_data.utils import uprating as uprating_module
+    from policyengine_uk_data.utils.uprating import uprate_values
+
+    storage = _seed_uprating_table(tmp_path)
+    monkeypatch.setattr(uprating_module, "STORAGE_FOLDER", storage)
+
+    # A supported year range still works and returns a sensible factor.
+    result = uprate_values(100.0, "employment_income", start_year=2020, end_year=2034)
+    # Seed table has linear growth 0.02 per year; 2020→2034 = 14 years ×
+    # 0.02 = 1.28 multiplier off a 1.0 base.
+    assert result == pytest.approx(128.0)

policyengine_uk_data/utils/uprating.py

@@ -6,6 +6,28 @@
 END_YEAR = 2034
 
 
+class UpratingYearOutOfRangeError(ValueError):
+    """Raised when a caller asks for an uprating factor outside the table range.
+
+    The uprating factor table is written by `create_policyengine_uprating_factors_table()`
+    for years in ``[START_YEAR, END_YEAR]`` and does not include columns for years
+    beyond ``END_YEAR``. Silently returning a KeyError (or a stale last-year factor)
+    would produce wrong values; raising a specific, actionable error instead tells
+    the caller to either regenerate the table with a later ``END_YEAR`` or pick a
+    supported target year.
+    """
+
+
+def _check_year_in_range(year: int, *, kind: str) -> None:
+    if year < START_YEAR or year > END_YEAR:
+        raise UpratingYearOutOfRangeError(
+            f"{kind}={year} is outside the uprating table range "
+            f"[{START_YEAR}, {END_YEAR}]. Regenerate the table via "
+            f"`create_policyengine_uprating_factors_table()` with a later "
+            f"END_YEAR, or pick a supported year."
+        )
+
+
 def create_policyengine_uprating_factors_table():
     from policyengine_uk.system import system
 
@@ -45,7 +67,10 @@ def create_policyengine_uprating_factors_table():
     return df
 
 
-def uprate_values(values, variable_name, start_year=2020, end_year=2034):
+def uprate_values(values, variable_name, start_year=START_YEAR, end_year=END_YEAR):
+    _check_year_in_range(start_year, kind="start_year")
+    _check_year_in_range(end_year, kind="end_year")
+
     uprating_factors = pd.read_csv(STORAGE_FOLDER / "uprating_factors.csv")
     uprating_factors = uprating_factors.set_index("Variable")
     uprating_factors = uprating_factors.loc[variable_name]
@@ -57,11 +82,14 @@ def uprate_values(values, variable_name, start_year=2020, end_year=2034):
     return values * relative_change
 
 
-def uprate_dataset(dataset: UKSingleYearDataset, target_year=2034):
+def uprate_dataset(dataset: UKSingleYearDataset, target_year: int = END_YEAR):
+    _check_year_in_range(target_year, kind="target_year")
+
     dataset = dataset.copy()
     uprating_factors = pd.read_csv(STORAGE_FOLDER / "uprating_factors.csv")
     uprating_factors = uprating_factors.set_index("Variable")
     start_year = dataset.time_period
+    _check_year_in_range(int(start_year), kind="dataset.time_period")
 
     for table in dataset.tables:
         for variable in table.columns: