Skip to content

✨ signal: improved decimate, detrend, unique_roots, and vectorstrength #1086

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
jorenham merged 4 commits into from
Jan 6, 2026
Merged

✨ signal: improved decimate, detrend, unique_roots, and vectorstrength #1086

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
eb6042f
✨ `signal`: improve `detrend`
jorenham Jan 6, 2026
1b41f51
✨ `signal`: improve `unique_roots`
jorenham Jan 6, 2026
8156f55
✨ `signal`: improve `decimate`
jorenham Jan 6, 2026
87d80b6
✨ `signal`: improve `vectorstrength`
jorenham Jan 6, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
108 changes: 78 additions & 30 deletions scipy-stubs/signal/_signaltools.pyi
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -79,7 +79,6 @@ _FiltFiltMethod: TypeAlias = L["pad", "gust"]
_ResidualKind: TypeAlias = L["lowpass", "all"]
_FilterType: TypeAlias = L["iir", "fir"] | dlti

_F16_64: TypeAlias = np.float64 | np.float32 | np.float16
_C64_128: TypeAlias = np.complex128 | np.complex64

_ToResampleWindow: TypeAlias = Callable[[onp.Array1D[_InexactT]], onp.ToFloat1D] | onp.ToFloat1D | _ToWindow
Expand Down Expand Up @@ -1009,29 +1008,69 @@ def hilbert2(

###

# TODO(jorenham): improve
@overload
#
@overload # f64
def detrend(
data: onp.ToFloatND, axis: int = -1, type: _TrendType = "linear", bp: int | onp.ToJustIntND = 0, overwrite_data: bool = False
) -> onp.ArrayND[_F16_64]: ...
@overload
data: onp.ToIntND | onp.ToJustFloat64_ND,
axis: int = -1,
type: _TrendType = "linear",
bp: int | onp.ToJustIntND = 0,
overwrite_data: bool = False,
) -> onp.ArrayND[np.float64]: ...
@overload # f16
def detrend(
data: onp.ToJustFloat16_ND,
axis: int = -1,
type: _TrendType = "linear",
bp: int | onp.ToJustIntND = 0,
overwrite_data: bool = False,
) -> onp.ArrayND[np.float16]: ...
@overload # f32
def detrend(
data: onp.ToJustFloat32_ND,
axis: int = -1,
type: _TrendType = "linear",
bp: int | onp.ToJustIntND = 0,
overwrite_data: bool = False,
) -> onp.ArrayND[np.float32]: ...
@overload # c64
def detrend(
data: onp.ToJustComplex64_ND,
axis: int = -1,
type: _TrendType = "linear",
bp: int | onp.ToJustIntND = 0,
overwrite_data: bool = False,
) -> onp.ArrayND[np.complex128]: ...
@overload # c128
def detrend(
data: onp.ToJustComplex128_ND,
axis: int = -1,
type: _TrendType = "linear",
bp: int | onp.ToJustIntND = 0,
overwrite_data: bool = False,
) -> onp.ArrayND[np.complex128]: ...
@overload # fallback
def detrend(
data: onp.ToComplexND,
axis: int = -1,
type: _TrendType = "linear",
bp: int | onp.ToJustIntND = 0,
overwrite_data: bool = False,
) -> onp.ArrayND[_C64_128 | _F16_64]: ...
) -> onp.ArrayND[Any]: ...

# TODO(jorenham): improve
@overload
#
@overload # f64
def unique_roots(
p: onp.ToFloat1D, tol: float = 0.001, rtype: _ResidueType = "min"
) -> tuple[onp.Array1D[npc.floating], onp.Array1D[np.int_]]: ...
@overload
p: onp.ToInt1D | onp.ToJustFloat64_1D, tol: float = 0.001, rtype: _ResidueType = "min"
) -> tuple[onp.Array1D[np.float64], onp.Array1D[np.int_]]: ...
@overload # c128
def unique_roots(
p: onp.ToJustComplex128_1D, tol: float = 0.001, rtype: _ResidueType = "min"
) -> tuple[onp.Array1D[np.complex128], onp.Array1D[np.int_]]: ...
@overload # T: inexact
def unique_roots(
p: onp.ToComplex1D, tol: float = 0.001, rtype: _ResidueType = "min"
) -> tuple[onp.Array1D[npc.inexact], onp.Array1D[np.int_]]: ...
p: onp.ToArray1D[_InexactT, _InexactT], tol: float = 0.001, rtype: _ResidueType = "min"
) -> tuple[onp.Array1D[_InexactT], onp.Array1D[np.int_]]: ...

#
def residue(
Expand Down Expand Up @@ -1242,25 +1281,34 @@ def resample_poly(
cval: float | None = None,
) -> onp.ArrayND[Any, _WorkaroundForPyright]: ...

# TODO(jorenham): improve
@overload
# `float16` is upcast to `float64`, while `float32` (and `float64`) retain their dtype
@overload # f64
def decimate(
x: onp.ToFloatND, q: int, n: int | None = None, ftype: _FilterType = "iir", axis: int = -1, zero_phase: bool = True
) -> onp.ArrayND[npc.floating]: ...
@overload
x: onp.ToArrayND[float, np.float64 | np.float16 | npc.integer | np.bool_],
q: int,
n: int | None = None,
ftype: _FilterType = "iir",
axis: int = -1,
zero_phase: bool = True,
) -> onp.ArrayND[np.float64]: ...
@overload # f32
def decimate(
x: onp.ToJustFloat32_ND, q: int, n: int | None = None, ftype: _FilterType = "iir", axis: int = -1, zero_phase: bool = True
) -> onp.ArrayND[np.float32]: ...
@overload # c128
def decimate(
x: onp.ToJustComplex128_ND, q: int, n: int | None = None, ftype: _FilterType = "iir", axis: int = -1, zero_phase: bool = True
) -> onp.ArrayND[np.complex128]: ...
@overload # c64
def decimate(
x: onp.ToComplexND, q: int, n: int | None = None, ftype: _FilterType = "iir", axis: int = -1, zero_phase: bool = True
) -> onp.ArrayND[npc.inexact]: ...
x: onp.ToJustComplex64_ND, q: int, n: int | None = None, ftype: _FilterType = "iir", axis: int = -1, zero_phase: bool = True
) -> onp.ArrayND[np.complex64]: ...

# TODO(jorenham): improve
@overload
def vectorstrength(events: onp.ToFloat1D, period: onp.ToFloat) -> _Tuple2[npc.floating]: ...
@overload
def vectorstrength(events: onp.ToFloat1D, period: onp.ToFloat1D) -> _Tuple2[onp.Array1D[npc.floating]]: ...
@overload
def vectorstrength(events: onp.ToComplex1D, period: onp.ToComplex) -> _Tuple2[npc.inexact]: ...
@overload
def vectorstrength(events: onp.ToComplex1D, period: onp.ToComplex1D) -> _Tuple2[onp.Array1D[npc.inexact]]: ...
# complex periods are always cast to float64, and therefore not supported here
@overload # 0d
def vectorstrength(events: float | onp.ToFloat1D, period: float) -> _Tuple2[np.float64]: ...
@overload # 1d
def vectorstrength(events: float | onp.ToFloat1D, period: onp.ToFloat1D) -> _Tuple2[onp.Array1D[np.float64]]: ...

#
@overload
Expand Down