fix: documentation generation

README.md

@@ -23,7 +23,7 @@ $ uv sync
 ## Build documentation
 
 ```console
-$ sphinx-build -M html docs docs/_build
+$ rm -rf docs/{generated,_build} && sphinx-build -M html docs docs/_build
 ```
 
 ## Citation

docs/api.md

@@ -1,45 +1,52 @@
 # API
 
-## scanvi_explainer
+```
+import scanvi_explainer
+```
+
+```{eval-rst}
+.. currentmodule:: scanvi_explainer
+
+```
 
-### Utilities: `utils`
+## SCANVIDeep
 
 ```{eval-rst}
 .. currentmodule:: scanvi_explainer
 ```
 
 ```{eval-rst}
 .. autosummary::
-    :toctree: generated
+    :toctree: generated/
     :nosignatures:
 
-    utils.train_test_group_split
+    scanvi_deep.SCANVIDeep
 ```
 
-### Plots: `plots`
+## Utilities
 
 ```{eval-rst}
 .. currentmodule:: scanvi_explainer
 ```
 
 ```{eval-rst}
 .. autosummary::
-    :toctree: generated
+    :toctree: generated/
     :nosignatures:
 
-    plots.feature_plot
+    utils.train_test_group_split
 ```
 
-## SCANVIDeep
+## Plots
 
 ```{eval-rst}
 .. currentmodule:: scanvi_explainer
 ```
 
 ```{eval-rst}
 .. autosummary::
-    :toctree: generated
+    :toctree: generated/
     :nosignatures:
 
-    scanvi_deep
-```
+    plots.feature_plot
+```

docs/conf.py

@@ -10,6 +10,9 @@
 from importlib.metadata import metadata
 from pathlib import Path
 
+HERE = Path(__file__).parent
+sys.path[:0] = [str(HERE.parent)]
+
 # -- Project information -----------------------------------------------------
 
 # NOTE: If you installed your project in editable mode, this might be stale.
@@ -120,8 +123,8 @@
 
 pygments_style = "default"
 
-nitpick_ignore = [
-    # If building the documentation fails because of a missing link that is outside your control,
-    # you can add an exception to this list.
-    ("optional"),
-]
+# nitpick_ignore = [
+#     # If building the documentation fails because of a missing link that is outside your control,
+#     # you can add an exception to this list.
+#     ("py:class", "py:obj", "optional"),
+# ]

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "scanvi-explainer"
-version = "0.2.0"
+version = "0.2.1"
 description = "Exapliner of scANVI using SHAP"
 authors = [
     {name = "Martin Proks", email = "[email protected]"},

src/scanvi_explainer/__init__.py

@@ -1,5 +1,6 @@
 import importlib.util
 
+from . import plots, utils
 from .scanvi_deep import SCANVIDeep
 
 if not importlib.util.find_spec("torch"):
@@ -8,4 +9,4 @@
 if not importlib.util.find_spec("scvi"):
     raise ImportError("Missing torch package! Run pip install torch")
 
-__all__ = ["SCANVIDeep", "utils", "plots"]
+__all__ = ["plots", "SCANVIDeep", "utils"]

src/scanvi_explainer/plots.py

@@ -1,27 +1,27 @@
 import matplotlib.pyplot as plt
-import numpy.typing as npt
+import numpy.typing as ntp
 import pandas as pd
 import seaborn as sns
 from scvi import REGISTRY_KEYS
 
-from .scanvi_deep import SCANVIDeep
+from scanvi_explainer.scanvi_deep import SCANVIDeep
 
 
 def feature_plot(
     explainer: SCANVIDeep,
-    shap_values: npt.NDArray,
+    shap_values: ntp.ArrayLike,
     subset: bool = False,
     top_n: int = 10,
 ) -> None:
     """Prints feature contribution (absolute mean SHAP value) for each cell type (top 10).
 
     Parameters
     ----------
-    explainer : SCANVIDeep
-        SCANVIDeep explainer
-    shap_values : npt.NDArray
-        Expected SHAP values
-    subset : bool, optional
+    explainer
+        :class:`~scanvi_explainer.scanvi_deep.SCANVIDeep` explainer
+    shap_values
+        :class:`numpy.typing.ArrayLike` Expected SHAP values
+    subset : :obj:`bool`, optional
         When set to true, calculate contribution by subsetting for test cells which belong to that
         particual classifier.
         When set to false, be generic and return contributing features even when testing set has

src/scanvi_explainer/utils.py

@@ -4,7 +4,7 @@
 import numpy as np
 import torch
 from scvi import REGISTRY_KEYS
-from scvi.model._scanvi import SCANVI
+from scvi.model import SCANVI
 from sklearn.model_selection import train_test_split
 
 
@@ -60,9 +60,9 @@ def train_test_group_split(
         Annotated dataset
     groupby : str
         Column in metadata by which the dataset should be split by
-    train_size : float, optional
+    train_size : :obj:`float`, optional
         Training size (background), by default 0.8
-    batch_size : int, optional
+    batch_size : :obj:`int`, optional
         Number of cells used from each group, by default 128
 
     Returns