Make more parameters optional

sgkit/stats/pca.py

@@ -20,24 +20,28 @@ def pca_est(
     n_components: int = 10,
     *,
     ploidy: Optional[int] = None,
-    scaler: Union[BaseEstimator, str] = "patterson",
-    algorithm: Literal["tsqr", "randomized"] = "tsqr",
+    scaler: Optional[Union[BaseEstimator, str]] = None,
+    algorithm: Optional[Literal["tsqr", "randomized"]] = None,
     n_iter: int = 0,
     random_state: RandomStateType = 0,
     variable: str = "call_alternate_allele_count",
 ) -> BaseEstimator:
     """ Create PCA estimator """
-    if not ploidy:
+    if ploidy is None:
         if "ploidy" not in ds.dims:
             raise ValueError(
                 "`ploidy` must be specified explicitly when not present in dataset dimensions"
             )
         ploidy = ds.dims["ploidy"]
+    if scaler is None:
+        scaler = "patterson"
     if isinstance(scaler, str):
         if scaler != "patterson":
             raise ValueError(
                 f"Only 'patterson' scaler currently supported (not '{scaler}')"
             )
+    if algorithm is None:
+        algorithm = "tsqr"
     if algorithm not in ["tsqr", "randomized"]:
         raise ValueError(
             f"`algorithm` must be one of ['tsqr', 'randomized'] (not '{algorithm}')"
@@ -133,8 +137,8 @@ def pca(
     n_components: int = 10,
     *,
     ploidy: Optional[int] = None,
-    scaler: Union[BaseEstimator, str] = "patterson",
-    algorithm: Literal["tsqr", "randomized"] = "tsqr",
+    scaler: Optional[Union[BaseEstimator, str]] = None,
+    algorithm: Optional[Literal["tsqr", "randomized"]] = None,
     n_iter: int = 0,
     random_state: RandomStateType = 0,
     variable: str = "call_alternate_allele_count",
@@ -154,7 +158,7 @@ def pca(
         dimension in provided dataset if not passed explicitly.
     scaler
         Scaler implementation used to normalize alternate allele counts, by default 'patterson'.
-        This is currently the only supported scaler, but a custom estimator may also be passed.
+        This is currently the only supported scaler but a custom estimator may also be passed.
         See ``sgkit.stats.preprocessing.PattersonScaler`` for more details.
     algorithm
         Name of SVD algorithm to use, by default "tsqr". Must be in ["tsqr", "randomized"].