Merge pull request #10 from mvinyard/bug-fixes

Bug fixes

setup.py

@@ -6,7 +6,7 @@
 
 
 # -- specify package version: --------------------------------------------------
-__version__ = "0.0.20"
+__version__ = "0.0.21"
 
 
 # -- import packages: ----------------------------------------------------------
@@ -19,8 +19,8 @@
 # -- run setup: ----------------------------------------------------------------
 setuptools.setup(
     name="torch-adata",
-    version="0.0.20",
-    python_requires=">3.7.0",
+    version="0.0.21",
+    python_requires=">3.9.0",
     author="Michael E. Vinyard",
     author_email="mvinyard@g.harvard.edu",
     url=None,
@@ -29,10 +29,10 @@
     description="torch-adata",
     packages=setuptools.find_packages(),
     install_requires=[
-        "anndata>=0.8",
+        "anndata>=0.9.1",
         "licorice_font>=0.0.3",
-        "pytorch-lightning>=1.7.7",
-        "torch>=1.12",        
+        "lightning>=2.0.1",
+        "torch>=2.0",        
     ],
     classifiers=[
         "Development Status :: 2 - Pre-Alpha",

torch_adata/__init__.py

@@ -6,7 +6,7 @@
 
 
 # -- specify package version: --------------------------------------------------
-__version__ = "0.0.20"
+__version__ = "0.0.21"
 
 
 # -- import modules: -----------------------------------------------------------

torch_adata/_core/_core_ancilliary/_identity_msg.py

@@ -26,7 +26,14 @@ def _groupby_msg(dataset):
     if not dataset._grouped_by:
         return
     msg_g1 = lf.font_format("Grouped by", ["BOLD"])
-    msg_g2 = lf.font_format(dataset._grouped_by, ["RED"])
+    if isinstance(dataset._grouped_by, list):
+        sub_g1 = []
+        for gb in dataset._grouped_by:
+            sub_g1.append(lf.font_format(gb, ["RED"]))
+            msg_g2 = "-".join(sub_g1)
+    else:
+        msg_g2 = lf.font_format(dataset._grouped_by, ["RED"])
+
     return "{}: '{}' with attributes:".format(msg_g1, msg_g2)
 
 def annotate_attr_size(dataset, attr_name_set):

torch_adata/_core/_lightning/_lightning_anndata_module.py

@@ -10,7 +10,7 @@
 
 
 # -- import packages: --------------------------------------------------------------------
-from pytorch_lightning import LightningDataModule
+from lightning import LightningDataModule
 from licorice_font import font_format
 import pandas as pd
 import anndata
@@ -35,7 +35,8 @@ def __init__(
         train_val_split=[0.8, 0.2],
         n_predict=2000,
         use_key="X_pca",
-        groupby="Time point",  # TODO: make optional
+        obs_keys=[],
+        groupby=None,
         train_key="train",
         val_key="val",
         test_key="test",

torch_adata/_tools/__init__.py

@@ -14,4 +14,5 @@
 from ._idx import idx
 from ._dummy_batch import dummy_batch
 from ._base_lightning_data_module import BaseLightningDataModule
-from ._anndataset_split import AnnDatasetSplit
+from ._anndataset_split import AnnDatasetSplit
+from ._fetch_data import fetch

torch_adata/_tools/_base_lightning_data_module.py

@@ -11,7 +11,7 @@
 
 
 # -- import packages: ----------------------------------------------------------
-from pytorch_lightning import LightningDataModule
+from lightning import LightningDataModule
 from torch.utils.data import DataLoader
 import anndata
 

torch_adata/_tools/_fetch_data.py

@@ -0,0 +1,22 @@
+
+__module_name__ = "_fetch_data.py"
+__doc__ = """Data fetching module. Key module called within torch_adata.AnnDataset.__init__()."""
+__author__ = ", ".join(["Michael E. Vinyard"])
+__email__ = ", ".join(["vinyard@g.harvard.edu"])
+
+
+# -- import packages: --------------------------------------------------------------------
+import torch
+import anndata
+
+
+# -- import local dependencies: ----------------------------------------------------------
+from .._core._core_ancilliary._fetch_data import Fetch
+
+
+# -- fetch X from adata: -----------------------------------------------------------------
+def fetch(adata: anndata.AnnData, use_key: str)->torch.Tensor:
+
+    f = Fetch(adata)
+    return f.X(use_key)
+