dascore/proc/filter.py (1)

449-518: Clean up remaining docstring typos for slope_filter.
Multiple minor typos remain; propose a quick sweep.

Apply:

@@
-    Most commonly this used as an F-K (frequency wavenumber)
+    Most commonly this is used as an F-K (frequency–wavenumber)
@@
-        the filter selects the apparent velocites between 'vb' and 'vc'
+        the filter selects the apparent velocities between 'vb' and 'vc'
@@
-        This can be used for up-down/left-right separation, assuming a
-        near-linear fiber layout.
+        This can be used for up/down or left/right separation, assuming a
+        near-linear fiber layout.
@@
-        specified by the inner `filt` parameterse are attenuated rather
+        specified by the inner `filt` parameters are attenuated rather
@@
-    The [fk recipe](`docs/recipes/fk.qmd`) provides addtional examples.
+    The [FK recipe](`docs/recipes/fk.qmd`) provides additional examples.

dascore/transform/fourier.py (1)

374-381: Type hint and docstring: accept overlap=None explicitly.

The implementation supports None, but the signature says Quantity | int. Update the hint and docstring to avoid confusion.

Example signature tweak:

def stft(
    patch: PatchType,
    taper_window: str | ndarray | tuple[str, Any, ...] = "hann",
    overlap: Quantity | int | None = 50 * percent,
    samples: bool = False,
    detrend: bool = False,
    **kwargs,
):
    """
    ...
    overlap
        The overlap between windows. If None, use 0 (no overlap).
    ...

Also add:

from dascore.exceptions import PatchError, ParameterError

at the import section.

tests/test_integrations/test_misc_integrations.py (5)

14-19: Make train_patch fixture module-scoped and resilient to missing data.

Avoid repeated I/O per test and skip cleanly if the dataset isn’t available (useful for offline CI).

Apply this diff to remove the class-scoped fixture:

-    @pytest.fixture
-    def train_patch(self):
-        """Get an example patch with trains in it."""
-        out = dc.spool(fetch("UoU_lf_urban.hdf5"))[0]
-        return out

Add this module-level fixture outside the class:

# top-level (outside the class)
@pytest.fixture(scope="module")
def train_patch():
    """Get an example patch with trains in it (once per module)."""
    try:
        return dc.spool(fetch("UoU_lf_urban.hdf5"))[0]
    except Exception as exc:
        pytest.skip(f"dataset UoU_lf_urban.hdf5 unavailable: {exc}")

---

20-26: Avoid shadowing and strengthen assertions.

Rename the local variable to avoid shadowing the stft method name, and assert dims are preserved post-filter.

-        stft = patch.stft(time=0.1, overlap=None)
-        out = stft.pass_filter(time=(0.01, ...))
+        stft_patch = patch.stft(time=0.1, overlap=None)
+        out = stft_patch.pass_filter(time=(0.01, ...))
         assert isinstance(out, dc.Patch)
+        assert out.dims == stft_patch.dims
+        assert "ft_time" in stft_patch.dims

---

27-35: Rename local variable for clarity.

Minor readability: avoid reusing the name “stft.”

-        stft = train_patch.stft(
+        stft_patch = train_patch.stft(
             time=3, overlap=0.5, taper_window=("tukey", 0.1), detrend=True
         )
-        # Then sum power over 1 to 2 Hz
-        banded_patch = (stft.abs() ** 2).select(ft_time=(1, 2)).sum("ft_time").squeeze()
+        # Then sum power over 1 to 2 Hz
+        banded_patch = (
+            stft_patch.abs() ** 2
+        ).select(ft_time=(1, 2)).sum("ft_time").squeeze()

---

36-38: Prefer canonical unit spelling to avoid pint parser differences.

“mi/hour” is reliably parsed across pint versions; “miles/hour” can be version/definition dependent.

-        mph = dc.get_quantity("miles/hour")
+        mph = dc.get_quantity("mi/hour")

If you prefer miles/hour textually, confirm your supported environments all parse "miles/hour" equivalently.

---

27-39: Add simple invariants to catch regressions.

Validate that slope_filter preserves shape/dims and that history is appended.

         sf = banded_patch.slope_filter(filt)
-        assert "slope_filter" in str(sf.attrs.history)
+        assert sf.data.shape == banded_patch.data.shape
+        assert sf.dims == banded_patch.dims
+        assert "slope_filter" in str(sf.attrs.history)

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

💡 Knowledge Base configuration:

• MCP integration is disabled by default for public repositories
• Jira integration is disabled by default for public repositories
• Linear integration is disabled by default for public repositories

You can enable these sources in your CodeRabbit configuration.

Learnt from: d-chambers
PR: DASDAE/dascore#523
File: dascore/transform/fourier.py:433-434
Timestamp: 2025-09-02T16:22:22.829Z
Learning: In dascore STFT implementation, use `require_evenly_sampled=True` on `get_coord` calls instead of manual validation for coord.step to ensure sampling rate calculations are safe.

dascore/core/attrs.py (1)

252-254: Confirm dims behavior when clearing coords.

When falsy coords are passed, coords are cleared but dims remain unchanged. If the intended invariant is “dims reflect current coords,” consider also updating dims (e.g., to tuple()) or deriving from new_coords.

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

💡 Knowledge Base configuration:

• MCP integration is disabled by default for public repositories
• Jira integration is disabled by default for public repositories
• Linear integration is disabled by default for public repositories

You can enable these sources in your CodeRabbit configuration.