chore(typing): add type hints

src/ruptures/base.py

@@ -5,6 +5,7 @@
 """
 
 import abc
+from typing_extensions import Self
 from ruptures.utils import pairwise
 
 
@@ -18,17 +19,17 @@ class BaseEstimator(metaclass=abc.ABCMeta):
     """
 
     @abc.abstractmethod
-    def fit(self, *args, **kwargs):
+    def fit(self, *args, **kwargs) -> Self:
         """To call the segmentation algorithm."""
         pass
 
     @abc.abstractmethod
-    def predict(self, *args, **kwargs):
+    def predict(self, *args, **kwargs) -> list[int]:
         """To call the segmentation algorithm."""
         pass
 
     @abc.abstractmethod
-    def fit_predict(self, *args, **kwargs):
+    def fit_predict(self, *args, **kwargs) -> list[int]:
         """To call the segmentation algorithm."""
         pass
 
@@ -43,17 +44,17 @@ class BaseCost(object, metaclass=abc.ABCMeta):
     """
 
     @abc.abstractmethod
-    def fit(self, *args, **kwargs):
+    def fit(self, *args, **kwargs) -> Self:
         """Set the parameters of the cost function, for instance the Gram
         matrix, etc."""
         pass
 
     @abc.abstractmethod
-    def error(self, start, end):
+    def error(self, start: int, end: int) -> float:
         """Returns the cost on segment [start:end]."""
         pass
 
-    def sum_of_costs(self, bkps):
+    def sum_of_costs(self, bkps: list[int]) -> float:
         """Returns the sum of segments cost for the given segmentation.
 
         Args:
@@ -67,5 +68,5 @@ def sum_of_costs(self, bkps):
 
     @property
     @abc.abstractmethod
-    def model(self):
+    def model(self) -> str:
         pass

src/ruptures/costs/costautoregressive.py

@@ -1,7 +1,9 @@
 import numpy as np
 from numpy.lib.stride_tricks import as_strided
 from numpy.linalg import lstsq
+from numpy.typing import NDArray
 from copy import deepcopy
+from typing_extensions import Self
 
 from ruptures.base import BaseCost
 from ruptures.costs import NotEnoughPoints
@@ -12,7 +14,7 @@ class CostAR(BaseCost):
 
     model = "ar"
 
-    def __init__(self, order=4):
+    def __init__(self, order: int = 4) -> None:
         """Initialize the object.
 
         Args:
@@ -23,7 +25,7 @@ def __init__(self, order=4):
         self.min_size = max(5, order + 1)
         self.order = order
 
-    def fit(self, signal):
+    def fit(self, signal: NDArray[np.number]) -> Self:
         """Set parameters of the instance. The signal must be 1D.
 
         Args:
@@ -49,7 +51,7 @@ def fit(self, signal):
         self.signal[: self.order] = self.signal[self.order]
         return self
 
-    def error(self, start, end):
+    def error(self, start: int, end: int) -> float:
         """Return the approximation cost on the segment [start:end].
 
         Args:

src/ruptures/costs/costclinear.py

@@ -1,6 +1,8 @@
 r"""Continuous linear change."""
 
 import numpy as np
+from numpy.typing import NDArray
+from typing_extensions import Self
 
 from ruptures.base import BaseCost
 from ruptures.costs import NotEnoughPoints
@@ -11,12 +13,12 @@ class CostCLinear(BaseCost):
 
     model = "clinear"
 
-    def __init__(self):
+    def __init__(self) -> None:
         """Initialize the object."""
         self.signal = None
         self.min_size = 3
 
-    def fit(self, signal) -> "CostCLinear":
+    def fit(self, signal: NDArray[np.number]) -> Self:
         """Set parameters of the instance.
 
         Args:
@@ -32,7 +34,7 @@ def fit(self, signal) -> "CostCLinear":
 
         return self
 
-    def error(self, start, end) -> float:
+    def error(self, start: int, end: int) -> float:
         """Return the approximation cost on the segment [start:end].
 
         Args:

src/ruptures/costs/costcosine.py

@@ -1,6 +1,8 @@
 r"""CostCosine (kernel change point detection with the cosine similarity)"""
 
+from typing_extensions import Self
 import numpy as np
+from numpy.typing import NDArray
 from ruptures.base import BaseCost
 from ruptures.costs import NotEnoughPoints
 from scipy.spatial.distance import pdist, squareform
@@ -11,14 +13,14 @@ class CostCosine(BaseCost):
 
     model = "cosine"
 
-    def __init__(self):
+    def __init__(self) -> None:
         """Initialize the object."""
         self.signal = None
         self.min_size = 1
         self._gram = None
 
     @property
-    def gram(self):
+    def gram(self) -> NDArray[np.number]:
         """Generate the gram matrix (lazy loading).
 
         Only access this function after a `.fit()` (otherwise
@@ -28,7 +30,7 @@ def gram(self):
             self._gram = squareform(1 - pdist(self.signal, metric="cosine"))
         return self._gram
 
-    def fit(self, signal) -> "CostCosine":
+    def fit(self, signal: NDArray[np.number]) -> Self:
         """Set parameters of the instance.
 
         Args:
@@ -43,7 +45,7 @@ def fit(self, signal) -> "CostCosine":
             self.signal = signal
         return self
 
-    def error(self, start, end) -> float:
+    def error(self, start: int, end: int) -> float:
         """Return the approximation cost on the segment [start:end].
 
         Args:

src/ruptures/costs/costl1.py

@@ -1,6 +1,9 @@
 r"""CostL1 (least absolute deviation)"""
 
+from typing_extensions import Self
+
 import numpy as np
+from numpy.typing import NDArray
 
 from ruptures.base import BaseCost
 from ruptures.costs import NotEnoughPoints
@@ -16,7 +19,7 @@ def __init__(self) -> None:
         self.signal = None
         self.min_size = 2
 
-    def fit(self, signal) -> "CostL1":
+    def fit(self, signal: NDArray[np.number]) -> Self:
         """Set parameters of the instance.
 
         Args:
@@ -32,7 +35,7 @@ def fit(self, signal) -> "CostL1":
 
         return self
 
-    def error(self, start, end) -> float:
+    def error(self, start: int, end: int) -> float:
         """Return the approximation cost on the segment [start:end].
 
         Args:

src/ruptures/costs/costl2.py

@@ -1,5 +1,10 @@
 r"""CostL2 (least squared deviation)"""
 
+from numpy.typing import NDArray
+import numpy as np
+from typing_extensions import Self
+
+
 from ruptures.costs import NotEnoughPoints
 
 from ruptures.base import BaseCost
@@ -10,12 +15,12 @@ class CostL2(BaseCost):
 
     model = "l2"
 
-    def __init__(self):
+    def __init__(self) -> None:
         """Initialize the object."""
         self.signal = None
         self.min_size = 1
 
-    def fit(self, signal) -> "CostL2":
+    def fit(self, signal: NDArray[np.number]) -> Self:
         """Set parameters of the instance.
 
         Args:
@@ -31,7 +36,7 @@ def fit(self, signal) -> "CostL2":
 
         return self
 
-    def error(self, start, end) -> float:
+    def error(self, start: int, end: int) -> float:
         """Return the approximation cost on the segment [start:end].
 
         Args:

src/ruptures/costs/costlinear.py

@@ -1,6 +1,10 @@
 r"""Linear model change."""
 
+from typing_extensions import Self
+
+import numpy as np
 from numpy.linalg import lstsq
+from numpy.typing import NDArray
 
 from ruptures.base import BaseCost
 from ruptures.costs import NotEnoughPoints
@@ -11,13 +15,13 @@ class CostLinear(BaseCost):
 
     model = "linear"
 
-    def __init__(self):
+    def __init__(self) -> None:
         """Initialize the object."""
         self.signal = None
         self.covar = None
         self.min_size = 2
 
-    def fit(self, signal) -> "CostLinear":
+    def fit(self, signal: NDArray[np.number]) -> Self:
         """Set parameters of the instance. The first column contains the
         observed variable. The other columns contains the covariates.
 
@@ -33,7 +37,7 @@ def fit(self, signal) -> "CostLinear":
         self.covar = signal[:, 1:]
         return self
 
-    def error(self, start, end) -> float:
+    def error(self, start: int, end: int) -> float:
         """Return the approximation cost on the segment [start:end].
 
         Args:

src/ruptures/costs/costml.py

@@ -1,7 +1,11 @@
 r"""Change detection with a Mahalanobis-type metric."""
 
 import numpy as np
+from numpy.typing import NDArray
+
 from numpy.linalg import inv
+from typing import Optional
+from typing_extensions import Self
 
 from ruptures.base import BaseCost
 from ruptures.exceptions import NotEnoughPoints
@@ -12,7 +16,7 @@ class CostMl(BaseCost):
 
     model = "mahalanobis"
 
-    def __init__(self, metric=None):
+    def __init__(self, metric: Optional[NDArray[np.number]] = None) -> None:
         """Create a new instance.
 
         Args:
@@ -25,7 +29,7 @@ def __init__(self, metric=None):
         self.gram = None
         self.min_size = 2
 
-    def fit(self, signal) -> "CostMl":
+    def fit(self, signal: NDArray[np.number]) -> Self:
         """Set parameters of the instance.
 
         Args:
@@ -47,7 +51,7 @@ def fit(self, signal) -> "CostMl":
 
         return self
 
-    def error(self, start, end):
+    def error(self, start: int, end: int) -> float:
         """Return the approximation cost on the segment [start:end].
 
         Args:

src/ruptures/costs/costnormal.py

@@ -1,5 +1,8 @@
 r"""Gaussian process changes (CostNormal)"""
 
+from typing import Optional
+from numpy.typing import NDArray
+from typing_extensions import Self
 import warnings
 
 import numpy as np
@@ -13,7 +16,7 @@ class CostNormal(BaseCost):
 
     model = "normal"
 
-    def __init__(self, add_small_diag=True):
+    def __init__(self, add_small_diag: Optional[bool] = True) -> None:
         """Initialize the object.
 
         Args:
@@ -32,7 +35,7 @@ def __init__(self, add_small_diag=True):
                 UserWarning,
             )
 
-    def fit(self, signal) -> "CostNormal":
+    def fit(self, signal: NDArray[np.number]) -> Self:
         """Set parameters of the instance.
 
         Args:
@@ -49,7 +52,7 @@ def fit(self, signal) -> "CostNormal":
         self.n_samples, self.n_dims = self.signal.shape
         return self
 
-    def error(self, start, end) -> float:
+    def error(self, start: int, end: int) -> float:
         """Return the approximation cost on the segment [start:end].
 
         Args:

src/ruptures/costs/costrank.py

@@ -1,7 +1,9 @@
 r"""Rank-based cost function (CostRank)"""
 
+from typing_extensions import Self
 import numpy as np
 from numpy.linalg import pinv, LinAlgError
+from numpy.typing import NDArray
 from scipy.stats.mstats import rankdata
 
 from ruptures.base import BaseCost
@@ -13,13 +15,13 @@ class CostRank(BaseCost):
 
     model = "rank"
 
-    def __init__(self):
+    def __init__(self) -> None:
         """Initialize the object."""
         self.inv_cov = None
         self.ranks = None
         self.min_size = 2
 
-    def fit(self, signal) -> "CostRank":
+    def fit(self, signal: NDArray[np.number]) -> Self:
         """Set parameters of the instance.
 
         Args:
@@ -55,7 +57,7 @@ def fit(self, signal) -> "CostRank":
 
         return self
 
-    def error(self, start, end):
+    def error(self, start: int, end: int) -> float:
         """Return the approximation cost on the segment [start:end].
 
         Args: