format docstring, use auto doc typehints

dance/data/base.py

@@ -33,43 +33,43 @@ def _check_types_and_sizes(types, sizes):
 
 
 class BaseData(ABC):
-    """Base data object."""
+    """Base data object.
 
-    _FEATURE_CONFIGS: List[str] = ["feature_mod", "feature_channel", "feature_channel_type"]
-    _LABEL_CONFIGS: List[str] = ["label_mod", "label_channel", "label_channel_type"]
-    _DATA_CHANNELS: List[str] = ["obs", "var", "obsm", "varm", "obsp", "varp", "layers", "uns"]
+    The ``dance`` data object is a wrapper of the :class:`~anndata.AnnData` object, with several utility methods to
+    help retrieving data in specific splits in specific format (see :meth:`~BaseData.get_split_idx` and
+    :meth:`~BaseData.get_feature`). The :class:`~anndata.AnnData` objcet is saved in the attribute ``data`` and can be
+    accessed directly.
 
-    def __init__(self, data: Union[AnnData, MuData], train_size: Optional[int] = None, val_size: int = 0,
-                 test_size: int = -1):
-        """Initialize data object.
+    Warning
+    -------
+    Since the underlying data object is a reference to the input :class:`~anndata.AnnData` object, please be extra
+    cautious ***NOT*** initializing two different dance ``data`` object using the same :class:`~anndata.AnnData`
+    object! If you are unsure, we recommend always initialize the dance ``data`` object using a ``copy`` of the input
+    :class:`~anndata.AnnData` object, e.g.,
 
-        The `dance` data object is a wrapper of the :class:`~anndata.AnnData` object, with several utility methods
-        to help retrieving data in specific splits in specific format (see :meth:`~BaseData.get_split_idx` and
-        :meth:`~BaseData.get_feature`). The `AnnData` objcet is saved in the attribute :attr:`data` and can be
-        accessed directly.
+        >>> adata = anndata.AnnData(...)
+        >>> ddata = dance.data.Data(adata.copy())
 
-        Warning
-        -------
-        Since the underlying data object is a reference to the input :class:`~anndata.AnnData` object, please be
-        extra cautious ***NOT*** initializing two different dance `data` object using the same `AnnData` object!
-        If you are unsure, we recommend always initialize the dance `data` object using a `copy` of the input
-        `AnnData` object, e.g.,
+    Parameters
+    ----------
+    data
+        Cell data.
+    train_size
+        Number of cells to be used for training. If not specified, not splits will be generated.
+    val_size
+        Number of cells to be used for validation. If set to -1, use what's left from training and testing.
+    test_size
+        Number of cells to be used for testing. If set to -1, used what's left from training and validation.
 
-            >>> adata = anndata.AnnData(...)
-            >>> ddata = dance.data.Data(adata.copy())
 
-        Parameters
-        ----------
-        data
-            Cell data.
-        train_size
-            Number of cells to be used for training. If not specified, not splits will be generated.
-        val_size
-            Number of cells to be used for validation. If set to -1, use what's left from training and testing.
-        test_size
-            Number of cells to be used for testing. If set to -1, used what's left from training and validation.
+    """
 
-        """
+    _FEATURE_CONFIGS: List[str] = ["feature_mod", "feature_channel", "feature_channel_type"]
+    _LABEL_CONFIGS: List[str] = ["label_mod", "label_channel", "label_channel_type"]
+    _DATA_CHANNELS: List[str] = ["obs", "var", "obsm", "varm", "obsp", "varp", "layers", "uns"]
+
+    def __init__(self, data: Union[AnnData, MuData], train_size: Optional[int] = None, val_size: int = 0,
+                 test_size: int = -1):
         super().__init__()
 
         self._data = data
@@ -143,8 +143,8 @@ def config(self) -> Dict[str, Any]:
 
         Notes
         -----
-        The configuration dictionary is saved in the :attr:`data` attribute, which is an :class:`~anndata.AnnData`
-        object. Inparticular, the config will be saved in the `.uns` attribute with the key `"dance_config"`.
+        The configuration dictionary is saved in the ``data`` attribute, which is an :class:`~anndata.AnnData`
+        object. Inparticular, the config will be saved in the ``.uns`` attribute with the key ``"dance_config"``.
 
         """
         return self._data.uns["dance_config"]
@@ -166,8 +166,8 @@ def set_config_from_dict(self, config_dict: Dict[str, Any], *, overwrite: bool =
             Configuration dictionary.
         overwrite
             Used to determine the behaviour of resolving config conflicts. In the case of a conflict, where the config
-            dict passed contains a key with value that differs from an existing setting, if `overwrite` is set to
-            `False`, then raise a `KeyError`. Otherwise, overwrite the configuration with the new values.
+            dict passed contains a key with value that differs from an existing setting, if ``overwrite`` is set to
+            ``False``, then raise a ``KeyError``. Otherwise, overwrite the configuration with the new values.
 
         """
         # Check config key validity
@@ -305,14 +305,14 @@ def get_feature(self, *, split_name: Optional[str] = None, return_type: FeatType
             How should the features be returned. **numpy**: return as a numpy array; **torch**: return as a torch
             tensor; **anndata**: return as an anndata object.
         channel
-            Return a particular channel as features. If `channel_type` is `obs`, then return the column named by
-            `channel`, similarly for `var`. If `channel_type` is `obsm`, `obsp`, `varm`, `varp`, `layers`, or `uns`,
-            then return the value correspond to the `channel` in the dictionary. If `channel` is not specified
-            (default), then return the default layer (`.X`).
+            Return a particular channel as features. If ``channel_type`` is ``obs``, then return the column named by
+            ``channel``, similarly for ``var``. If ``channel_type`` is ``obsm``, ``obsp``, ``varm``, ``varp``,
+            ``layers``, or ``uns``, then return the value correspond to the ``channel`` in the dictionary. If
+            ``channel`` is not specified (default), then return the default layer (``.X``).
         channel_type
-            Channel type to use, default to `obsm`.
+            Channel type to use, default to ``obsm``.
         mod
-            Modality to use, default to `None`. Options other than `None` are only available when the underlying
+            Modality to use, default to ``None``. Options other than ``None`` are only available when the underlying
             data object is :class:`~mudata.Mudata`.
 
         """

dance/datasets/base.py

@@ -13,6 +13,16 @@
 
 
 class BaseDataset(ABC):
+    """BaseDataset abstract object.
+
+    Parameters
+    ----------
+    root
+        Root directory of the dataset.
+    full_download
+        If set to ``True``, then attempt to download all raw files of the dataset.
+
+    """
 
     _DISPLAY_ATTRS: Tuple[str] = ()
 

dance/modules/multi_modality/joint_embedding/jae.py

@@ -4,18 +4,17 @@
 
 Reference
 ---------
-Liu Q, Chen S, Jiang R, et al. Simultaneous deep generative modelling and clustering of single-cell genomic data[J]. Nature machine intelligence, 2021, 3(6): 536-544.
+Liu Q, Chen S, Jiang R, et al. Simultaneous deep generative modelling and clustering of single-cell genomic data[J].
+Nature machine intelligence, 2021, 3(6): 536-544.
 
 """
 
-import math
 import os
 
 import numpy as np
 import torch
 import torch.nn as nn
 import torch.nn.functional as F
-import torch.optim as optim
 from torch.utils.data import DataLoader
 
 from dance.utils import SimpleIndexDataset
@@ -34,7 +33,8 @@ class JAEWrapper:
     Parameters
     ----------
     args : argparse.Namespace
-        A Namespace object that contains arguments of JAE. For details of parameters in parser args, please refer to link (parser help document).
+        A Namespace object that contains arguments of JAE. For details of parameters in parser args, please refer to
+        link (parser help document).
     dataset : dance.datasets.multimodality.JointEmbeddingNIPSDataset
         Joint embedding dataset.
 

dance/modules/multi_modality/match_modality/scmogcn.py

@@ -2,7 +2,8 @@
 
 Reference
 ---------
-Wen, Hongzhi, et al. "Graph Neural Networks for Multimodal Single-Cell Data Integration." arXiv preprint arXiv:2203.01884 (2022).
+Wen, Hongzhi, et al. "Graph Neural Networks for Multimodal Single-Cell Data Integration." arXiv preprint
+arXiv:2203.01884 (2022).
 
 """
 import math
@@ -39,7 +40,8 @@ class ScMoGCNWrapper:
     Parameters
     ----------
     args : argparse.Namespace
-        A Namespace object that contains arguments of ScMoGCN. For details of parameters in parser args, please refer to link (parser help document).
+        A Namespace object that contains arguments of ScMoGCN. For details of parameters in parser args, please refer
+        to link (parser help document).
     layers : list[int]
         Specification of dimensions of hidden layers.
     temp : int optional

dance/modules/single_modality/cell_type_annotation/actinn.py

@@ -20,7 +20,18 @@
 
 
 class ACTINN(nn.Module):
-    """The ACTINN cell-type classification model."""
+    """The ACTINN cell-type classification model.
+
+    Parameters
+    ----------
+    hidden_dims : :obj:`tuple` of int
+        Hidden layer dimensions.
+    lambd : float
+        Regularization parameter
+    device : str
+        Training device
+
+    """
 
     def __init__(
             self,
@@ -29,18 +40,6 @@ def __init__(
             lambd: float = 0.01,
             device: str = "cpu",
     ):
-        """Initialize the ACTINN model.
-
-        Parameters
-        ----------
-        hidden_dims : :obj:`tuple` of int
-            Hidden layer dimensions.
-        lambd : float
-            Regularization parameter
-        device : str
-            Training device
-
-        """
         super().__init__()
 
         # Save attributes