Merge pull request #78 from JuBiotech/type-fixes

Fix various type hints

Author: lhelleckes

calibr8/core.py

@@ -9,18 +9,17 @@
 import json
 import logging
 import os
-import typing
 import warnings
 from pathlib import Path
-from typing import Callable, Optional, Sequence, Tuple, Union
+from typing import Callable, DefaultDict, List, Optional, Sequence, Tuple, Union
 
 import numpy
 import scipy
 
 from . import utils
 from .utils import DistributionType, pm
 
-__version__ = "7.1.1"
+__version__ = "7.1.2"
 _log = logging.getLogger("calibr8")
 
 
@@ -170,7 +169,7 @@ def _interval_prob(x_cdf: numpy.ndarray, cdf: numpy.ndarray, a: float, b: float)
     return cdf[ib] - cdf[ia]
 
 
-def _get_eti(x_cdf: numpy.ndarray, cdf: numpy.ndarray, ci_prob: float) -> typing.Tuple[float, float]:
+def _get_eti(x_cdf: numpy.ndarray, cdf: numpy.ndarray, ci_prob: float) -> Tuple[float, float]:
     """Find the equal tailed interval (ETI) corresponding to a certain credible interval probability level.
 
     Parameters
@@ -203,8 +202,8 @@ def _get_hdi(
     guess_lower: float,
     guess_upper: float,
     *,
-    history: typing.Optional[typing.DefaultDict[str, typing.List]] = None,
-) -> typing.Tuple[float, float]:
+    history: Optional[DefaultDict[str, List]] = None,
+) -> Tuple[float, float]:
     """Find the highest density interval (HDI) corresponding to a certain credible interval probability level.
 
     Parameters
@@ -600,7 +599,7 @@ def likelihood(self, *, y, x, theta=None, scan_x: bool = False):
                 return numpy.exp([self.loglikelihood(y=y, x=xi, theta=theta) for xi in x])
         return numpy.exp(self.loglikelihood(y=y, x=x, theta=theta))
 
-    def objective(self, independent, dependent, minimize=True) -> typing.Callable:
+    def objective(self, independent, dependent, minimize=True) -> Callable:
         """Creates an objective function for fitting to data.
 
         Parameters
@@ -628,7 +627,7 @@ def objective(x):
 
         return objective
 
-    def save(self, filepath: Union[Path, os.PathLike]):
+    def save(self, filepath: Union[str, Path, os.PathLike]):
         """Save key properties of the calibration model to a JSON file.
 
         Parameters
@@ -654,7 +653,7 @@ def save(self, filepath: Union[Path, os.PathLike]):
         return
 
     @classmethod
-    def load(cls, filepath: Union[Path, os.PathLike]):
+    def load(cls, filepath: Union[str, Path, os.PathLike]):
         """Instantiates a model from a JSON file of key properties.
 
         Parameters

calibr8/optimization.py

@@ -3,8 +3,7 @@
 likelihood estimation of calibration model parameters.
 """
 import logging
-import typing
-from typing import Any, Mapping, Optional, Sequence, Tuple, Union
+from typing import Any, Literal, Mapping, Optional, Sequence, Tuple, Union
 
 import numpy
 import scipy.optimize
@@ -14,7 +13,11 @@
 _log = logging.getLogger("calibr8.optimization")
 
 
-def _mask_and_warn_inf_or_nan(x: numpy.ndarray, y: numpy.ndarray, on: typing.Optional[str] = None):
+def _mask_and_warn_inf_or_nan(
+    x: Union[Sequence[float], numpy.ndarray],
+    y: Union[Sequence[float], numpy.ndarray],
+    on: Optional[Literal["x", "y"]] = None,
+) -> Tuple[numpy.ndarray, numpy.ndarray]:
     """Filters `x` and `y` such that only finite elements remain.
 
     Parameters
@@ -31,6 +34,8 @@ def _mask_and_warn_inf_or_nan(x: numpy.ndarray, y: numpy.ndarray, on: typing.Opt
     x : array
     y : array
     """
+    x = numpy.asarray(x)
+    y = numpy.asarray(y)
     xdims = numpy.ndim(x)
     if xdims == 1:
         mask_x = numpy.isfinite(x)
@@ -82,8 +87,8 @@ def _warn_hit_bounds(theta, bounds, theta_names) -> bool:
 def fit_scipy(
     model: core.CalibrationModel,
     *,
-    independent: numpy.ndarray,
-    dependent: numpy.ndarray,
+    independent: Union[Sequence[float], numpy.ndarray],
+    dependent: Union[Sequence[float], numpy.ndarray],
     theta_guess: Union[Sequence[float], numpy.ndarray],
     theta_bounds: Sequence[Tuple[float, float]],
     minimize_kwargs: Optional[Mapping[str, Any]] = None,
@@ -154,8 +159,8 @@ def fit_scipy(
 def fit_scipy_global(
     model: core.CalibrationModel,
     *,
-    independent: numpy.ndarray,
-    dependent: numpy.ndarray,
+    independent: Union[Sequence[float], numpy.ndarray],
+    dependent: Union[Sequence[float], numpy.ndarray],
     theta_bounds: list,
     method: Optional[str] = None,
     maxiter: int = 5000,

calibr8/utils.py

@@ -3,10 +3,9 @@
 imports, timestamp parsing and plotting.
 """
 import datetime
-import typing
 import warnings
 from collections.abc import Iterable
-from typing import Literal, Optional, Sequence, Tuple
+from typing import List, Literal, Optional, Sequence, Tuple
 
 import matplotlib
 import numpy
@@ -52,7 +51,7 @@ def __getattr__(self, attr):
     pm = ImportWarner("pymc")
 
 
-def parse_datetime(s: typing.Optional[str]) -> typing.Optional[datetime.datetime]:
+def parse_datetime(s: Optional[str]) -> Optional[datetime.datetime]:
     """Parses a timezone-aware datetime formatted like 2020-08-05T13:37:00Z.
 
     Returns
@@ -65,7 +64,7 @@ def parse_datetime(s: typing.Optional[str]) -> typing.Optional[datetime.datetime
     return datetime.datetime.strptime(s.replace("Z", "+0000"), "%Y-%m-%dT%H:%M:%S%z")
 
 
-def format_datetime(dt: typing.Optional[datetime.datetime]) -> typing.Optional[str]:
+def format_datetime(dt: Optional[datetime.datetime]) -> Optional[str]:
     """Formats a datetime like 2020-08-05T13:37:00Z.
 
     Returns
@@ -176,7 +175,9 @@ def plot_norm_band(ax, independent, mu, scale):
     return artists
 
 
-def plot_t_band(ax, independent, mu, scale, df, *, residual_type: typing.Optional[str] = None):
+def plot_t_band(
+    ax, independent, mu, scale, df, *, residual_type: Optional[Literal["absolute", "relative"]] = None
+):
     """Helper function for plotting the 68, 90 and 95 % likelihood-bands of a t-distribution.
 
     Parameters
@@ -241,7 +242,9 @@ def plot_t_band(ax, independent, mu, scale, df, *, residual_type: typing.Optiona
     return artists
 
 
-def plot_continuous_band(ax, independent, model, residual_type: typing.Optional[str] = None):
+def plot_continuous_band(
+    ax, independent, model, residual_type: Optional[Literal["absolute", "relative"]] = None
+):
     """Helper function for plotting the 68, 90 and 95 % likelihood-bands of a univariate distribution.
 
     Parameters
@@ -364,9 +367,9 @@ def plot_model(
     *,
     fig: Optional[matplotlib.figure.Figure] = None,
     axs: Optional[Sequence[matplotlib.axes.Axes]] = None,
-    residual_type="absolute",
+    residual_type: Literal["absolute", "relative"] = "absolute",
     band_xlim: Tuple[Optional[float], Optional[float]] = (None, None),
-):
+) -> Tuple[matplotlib.figure.Figure, List[matplotlib.axes.Axes]]:
     """Makes a plot of the model with its data.
 
     Parameters
@@ -416,6 +419,8 @@ def plot_model(
         axs.append(fig.add_subplot(gs1[0, 1], sharey=axs[0]))
         pyplot.setp(axs[1].get_yticklabels(), visible=False)
         axs.append(fig.add_subplot(gs2[0, 2]))
+    else:
+        axs = list(axs)
 
     # ======= Left =======
     # Untransformed, outer range