🧪 [test] add comprehensive tests for pd_to_datetime

tests/test_cons.py

@@ -0,0 +1,53 @@
+import pandas as pd
+import pytest
+from xalpha.cons import pd_to_datetime
+
+
+def test_pd_to_datetime_single_string():
+    res = pd_to_datetime("2023-01-01")
+    assert isinstance(res, pd.Timestamp)
+    assert res == pd.Timestamp("2023-01-01")
+
+
+def test_pd_to_datetime_list():
+    res = pd_to_datetime(["2023-01-01", "2023-01-02"])
+    assert isinstance(res, pd.DatetimeIndex)
+    assert res[0] == pd.Timestamp("2023-01-01")
+
+
+def test_pd_to_datetime_series():
+    series = pd.Series(["2023-01-01", "2023-01-02"])
+    res = pd_to_datetime(series)
+    # pd.to_datetime(pd.Series) returns a pd.Series
+    assert isinstance(res, pd.Series)
+    assert res.iloc[0] == pd.Timestamp("2023-01-01")
+
+
+def test_pd_to_datetime_with_format():
+    res = pd_to_datetime("20230101", format="%Y%m%d")
+    assert res == pd.Timestamp("2023-01-01")
+
+
+def test_pd_to_datetime_already_datetime():
+    dt = pd.Timestamp("2023-01-01")
+    res = pd_to_datetime(dt)
+    assert res == dt
+
+
+def test_pd_to_datetime_mixed_format():
+    # This test verifies behavior for mixed formats, which pd_to_datetime handles
+    res = pd_to_datetime(["2023-01-01", "2023/01/02"])
+    assert res[0] == pd.Timestamp("2023-01-01")
+    assert res[1] == pd.Timestamp("2023-01-02")
+
+
+def test_pd_to_datetime_invalid_string():
+    # Verify that invalid strings raise a ValueError by default (standard pandas behavior)
+    with pytest.raises(ValueError):
+        pd_to_datetime("not-a-date")
+
+
+def test_pd_to_datetime_errors_coerce():
+    # Verify that errors='coerce' works as expected
+    res = pd_to_datetime("not-a-date", errors="coerce")
+    assert pd.isna(res)