Scrublet (#129)

* copy over scrublet from scanpy

* update normalize and log1p

* update scrublet

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* updated log1p

* updates verbose for filters

* first working implementation for scrublet

* add tests

* update to working

* use cusparse for dense

* added docs

* update docstring

* added release notes

* updates docs

* fixes docstring

* fixes docs

* updates releasenotes

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>

docs/api/scanpy_gpu.md

@@ -36,6 +36,16 @@ Any transformation of the data matrix that is not a tool. Other than `tools`, pr
    pp.harmony_integrate
 ```
 
+### Doublet detection
+```{eval-rst}
+.. autosummary::
+   :toctree: generated/
+
+   pp.scrublet
+   pp.scrublet_simulate_doublets
+```
+
+
 ### Neighbors
 ```{eval-rst}
 .. autosummary::

docs/release-notes/0.10.0.md

@@ -6,6 +6,11 @@
 * added `get` module {pr}`100` {smaller}`S Dicks`
 * switch `utils` functions to get {pr}`100` {smaller}`S Dicks`
 * added `get.aggregated` to create condensed `anndata` objects {pr}`100` {smaller}`S Dicks`
+* added `pp.scrublet` and `pp.scrublet_simulate_doublets` {pr}`129` {smaller}`S Dicks`
+
+```{rubric} Bug fixes
+```
+* Fixes an issue where `pp.normalize` and `pp.log1p` now use `copy` and `inplace` corretly {pr}`129` {smaller}`S Dicks`
 
 ```{rubric} Removals
 ```

src/rapids_singlecell/preprocessing/__init__.py

@@ -5,6 +5,7 @@
 from ._pca import pca
 from ._regress_out import regress_out
 from ._scale import scale
+from ._scrublet import scrublet, scrublet_simulate_doublets
 from ._simple import (
     calculate_qc_metrics,
     filter_cells,

src/rapids_singlecell/preprocessing/_normalize.py

@@ -12,10 +12,12 @@
 
 def normalize_total(
     adata: AnnData,
+    *,
     target_sum: Optional[int] = None,
     layer: Optional[str] = None,
     inplace: bool = True,
-) -> Optional[Union[sparse.csr_matrix, cp.ndarray]]:
+    copy: bool = False,
+) -> Optional[Union[AnnData, sparse.csr_matrix, cp.ndarray]]:
     """
     Normalizes rows in matrix so they sum to `target_sum`
 
@@ -31,15 +33,21 @@ def normalize_total(
             Layer to normalize instead of `X`. If `None`, `X` is normalized.
 
         inplace
-            Whether to update `adata` or return the normalized matrix.
+            Whether to update `adata` or return the matrix.
 
+        copy
+            Whether to return a copy or update `adata`. Not compatible with inplace=False.
 
     Returns
     -------
     Returns a normalized copy or  updates `adata` with a normalized version of \
     the original `adata.X` and `adata.layers['layer']`, depending on `inplace`.
 
     """
+    if copy:
+        if not inplace:
+            raise ValueError("`copy=True` cannot be used with `inplace=False`.")
+        adata = adata.copy()
     X = _get_obs_rep(adata, layer=layer)
 
     _check_gpu_X(X)
@@ -88,13 +96,21 @@ def normalize_total(
 
     if inplace:
         _set_obs_rep(adata, X, layer=layer)
-    else:
+
+    if copy:
+        return adata
+    elif not inplace:
         return X
 
 
 def log1p(
-    adata: AnnData, layer: Optional[str] = None, copy: bool = False
-) -> Optional[Union[sparse.csr_matrix, cp.ndarray]]:
+    adata: AnnData,
+    *,
+    layer: Optional[str] = None,
+    obsm: Optional[str] = None,
+    inplace: bool = True,
+    copy: bool = False,
+) -> Optional[Union[AnnData, sparse.csr_matrix, cp.ndarray]]:
     """
     Calculated the natural logarithm of one plus the sparse matrix.
 
@@ -106,8 +122,14 @@ def log1p(
         layer
             Layer to normalize instead of `X`. If `None`, `X` is normalized.
 
+        obsm
+            Entry of obsm to transform.
+
+        inplace
+            Whether to update `adata` or return the matrix.
+
         copy
-            Whether to return a copy or update `adata`.
+            Whether to return a copy or update `adata`. Not compatible with inplace=False.
 
     Returns
     -------
@@ -116,15 +138,26 @@ def log1p(
             in-place and returns None.
 
     """
-    X = _get_obs_rep(adata, layer=layer)
+    if copy:
+        if not inplace:
+            raise ValueError("`copy=True` cannot be used with `inplace=False`.")
+        adata = adata.copy()
+    X = _get_obs_rep(adata, layer=layer, obsm=obsm)
 
     _check_gpu_X(X)
 
-    X = X.log1p()
-    adata.uns["log1p"] = {"base": None}
-    if not copy:
-        _set_obs_rep(adata, X, layer=layer)
+    if isinstance(X, cp.ndarray):
+        X = cp.log1p(X)
     else:
+        X = X.log1p()
+
+    adata.uns["log1p"] = {"base": None}
+    if inplace:
+        _set_obs_rep(adata, X, layer=layer, obsm=obsm)
+
+    if copy:
+        return adata
+    elif not inplace:
         return X