TST: add test for Leap_Second.dat

tests/__init__.py

@@ -0,0 +1,25 @@
+from dataclasses import KW_ONLY, dataclass
+from datetime import datetime
+from typing import Generic, TypeVar
+
+T = TypeVar("T", int, float)
+
+
+@dataclass(kw_only=True, frozen=True, slots=True)
+class Bounds(Generic[T]):
+    lo: T | None = None
+    hi: T | None = None
+
+
+@dataclass(frozen=True, slots=True)
+class ColumnMetadata(Generic[T]):
+    name: str
+    _: KW_ONLY
+    type: type[T]
+    bounds: Bounds[T] = Bounds()
+
+
+JAN_1_1962_DT = datetime(1962, 1, 1)
+JAN_1_1962_MJD = 37665.0
+NOW_DT = datetime.now()
+NOW_MJD = JAN_1_1962_MJD + (NOW_DT - JAN_1_1962_DT).days

tests/test_Leap_Second_dat.py

@@ -0,0 +1,109 @@
+import csv
+import re
+from dataclasses import KW_ONLY, dataclass
+from datetime import datetime
+from itertools import takewhile
+from pathlib import Path
+from typing import Generic, Mapping, NewType, TypeVar
+
+import pytest
+
+from . import JAN_1_1962_MJD, NOW_DT, NOW_MJD, Bounds, ColumnMetadata
+
+DATA_DIR = Path(__file__).parents[1] / "astropy_iers_data" / "data"
+DATA_FILE = DATA_DIR / "Leap_Second.dat"
+
+T = TypeVar("T", int, float)
+
+EXPECTED_COLUMN_META = [
+    ColumnMetadata("MJD", type=float, bounds=Bounds(lo=JAN_1_1962_MJD, hi=NOW_MJD)),
+    ColumnMetadata("day", type=int, bounds=Bounds(lo=1, hi=12)),
+    ColumnMetadata("month", type=int, bounds=Bounds(lo=1, hi=31)),
+    ColumnMetadata("year", type=int, bounds=Bounds(lo=1972, hi=NOW_DT.year)),
+    ColumnMetadata(
+        "tai-utc",
+        type=int,
+        bounds=Bounds(
+            lo=0,
+            # there is no theoritical upper bound here,
+            # however a value much larger than 100 is definitely suspicious
+            # as of 2026, when the maximum (then current) value was 37
+            hi=100,
+        ),
+    ),
+]
+
+
+@dataclass(kw_only=True, frozen=True, slots=True)
+class CSV_Data:
+    header: list[str]
+    rows: list[list[str]]
+    names: list[str]
+    types: list[type]
+
+    def __post_init__(self):
+        ncols = len(self.names)
+        assert len(self.types) == ncols
+        assert all(len(r) == ncols for r in self.rows)
+
+    @property
+    def ncols(self) -> int:
+        return len(self.types)
+
+    @property
+    def nrows(self) -> int:
+        return len(self.rows)
+
+    def get_columns(
+        self, *, max_rows: int | None = None
+    ) -> list[list[float] | list[int]]:
+        # last nrows of each column, cast to infered Python types
+        nrows = min(max_rows or self.nrows, self.nrows)
+        tail_rows = self.rows[self.nrows - nrows :]
+        assert len(tail_rows) == nrows
+        tail_cols = [[r[i] for r in tail_rows] for i in range(self.ncols)]
+        return [
+            [tp(c[i]) for i in range(nrows)]
+            for c, tp in zip(
+                tail_cols,
+                self.types,
+                strict=True,
+            )
+        ]
+
+
+@pytest.fixture(scope="module")
+def csv_data():
+    with DATA_FILE.open() as fh:
+        header = [
+            L.removeprefix("#").strip()
+            for L in takewhile(lambda L: L.startswith("#"), fh)
+        ]
+        rows = list(csv.reader(fh, delimiter=" ", skipinitialspace=True))
+
+    return CSV_Data(
+        header=header,
+        rows=rows,
+        names=["MJD", "day", "month", "year", "TAI-UTC"],
+        types=[float, int, int, int, int],
+    )
+
+
+def test_header(csv_data):
+    assert "File expires on" in "\n".join(csv_data.header)
+
+
+def test_data(csv_data, subtests):
+    cols = csv_data.get_columns()
+    for c, meta in zip(cols, EXPECTED_COLUMN_META, strict=True):
+        tp = meta.type
+        with subtests.test(f"{meta.name} (types)"):
+            assert all(type(e) == tp for e in c)
+
+        if (lo := meta.bounds.lo) is not None:
+            with subtests.test(f"{meta.name} (min)"):
+                assert min(c) >= lo
+
+        if (hi := meta.bounds.hi) is not None:
+            with subtests.test(f"{meta.name} (max)"):
+                assert max(c) <= hi

tests/test_eopc04_1962-now.py

@@ -8,32 +8,11 @@
 
 import pytest
 
+from . import JAN_1_1962_MJD, NOW_DT, NOW_MJD, Bounds, ColumnMetadata
+
 DATA_DIR = Path(__file__).parents[1] / "astropy_iers_data" / "data"
 DATA_FILE = DATA_DIR / "eopc04.1962-now"
 
-T = TypeVar("T", int, float)
-
-
-@dataclass(kw_only=True, frozen=True, slots=True)
-class Bounds(Generic[T]):
-    lo: T | None = None
-    hi: T | None = None
-
-
-@dataclass(frozen=True, slots=True)
-class ColumnMetadata(Generic[T]):
-    name: str
-    _: KW_ONLY
-    type: type[T]
-    bounds: Bounds[T] = Bounds()
-
-
-JAN_1_1962_DT = datetime(1962, 1, 1)
-JAN_1_1962_MJD = 37665.0
-NOW_DT = datetime.now()
-NOW_MJD = JAN_1_1962_MJD + (NOW_DT - JAN_1_1962_DT).days
-
-
 EXPECTED_COLUMN_META = [
     ColumnMetadata("YR", type=int, bounds=Bounds(lo=1962, hi=NOW_DT.year)),
     ColumnMetadata("MM", type=int, bounds=Bounds(lo=1, hi=12)),
@@ -116,7 +95,9 @@ def compatible_py_types(self) -> list[type]:
     def names(self) -> list[str]:
         return re.split(r"\s\s+", self.header[5].removeprefix("#").strip())
 
-    def get_columns(self, *, max_rows: int | None = None) -> list[list[float] | list[int]]:
+    def get_columns(
+        self, *, max_rows: int | None = None
+    ) -> list[list[float] | list[int]]:
         # last nrows of each column, cast to infered Python types
         nrows = min(max_rows or self.nrows, self.nrows)
         tail_rows = self.rows[self.nrows - nrows :]