Merge pull request #31 from infocusp/scaLR-updates

saurabhinfocusp · web-flow · commit b5f72ce8f9bd · 2026-04-03T12:09:14.000+05:30
Anndata_python_version_other updates
diff --git a/.github/workflows/docs.yml b/.github/workflows/docs.yml
@@ -11,6 +11,8 @@ jobs:
     steps:
       - uses: actions/checkout@v4
       - uses: actions/setup-python@v5
+        with:
+          python-version: "3.10.20"
       - name: Install scalr requirements
         run: |
           pip install -r requirements.txt
diff --git a/.github/workflows/publish-to-pypi.yml b/.github/workflows/publish-to-pypi.yml
@@ -15,7 +15,7 @@ jobs:
     - name: Set up Python
       uses: actions/setup-python@v5
       with:
-        python-version: "3.9"
+        python-version: "3.10.20"
     - name: Install pypa/build
       run: >-
         python3 -m
diff --git a/.github/workflows/run_isort.yml b/.github/workflows/run_isort.yml
@@ -10,10 +10,10 @@ jobs:
 
     steps:
     - uses: actions/checkout@v2
-    - name: Set up Python 3.9
+    - name: Set up Python 3.10.20
       uses: actions/setup-python@v2
       with:
-        python-version: 3.9
+        python-version: "3.10.20"
     - name: Install isort
       run: pip install isort
     - name: Run isort
diff --git a/.github/workflows/run_pytest.yml b/.github/workflows/run_pytest.yml
@@ -11,7 +11,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: ["3.9"]
+        python-version: ["3.10.20"]
 
     steps:
       - uses: actions/checkout@v3
diff --git a/README.md b/README.md
@@ -27,10 +27,10 @@
 ## Pre-requisites and installation scaLR
 
 
-- ScaLR can be installed using git or pip. It is tested in Python 3.10 and it is recommended to use that environment.
+- ScaLR can be installed using git or pip. It is tested in Python 3.10.20 and it is recommended to use that environment.
 
 ```
-conda create -n scaLR_env python=3.10
+conda create -n scaLR_env python=3.10.20
 
 conda activate scaLR_env
 ```
@@ -374,5 +374,4 @@ Performs evaluation of best model trained on user-defined metrics on the test se
   
 ## Citation
 
-Jogani Saiyam, Anand Santosh Pol, Mayur Prajapati, Amit Samal, Kriti Bhatia, Jayendra Parmar, Urvik Patel, Falak Shah, Nisarg Vyas, and Saurabh Gupta. "scaLR: a low-resource deep neural network-based platform for single cell analysis and biomarker discovery." bioRxiv (2024): 2024-09.
-
+Jogani, S., Pol, A. S., Prajapati, M., Samal, A., Bhatia, K., Parmar, J., ... & Gupta, S. (2025). scaLR: a low-resource deep neural network-based platform for single cell analysis and biomarker discovery. Briefings in Bioinformatics, 26(3), bbaf243.
diff --git a/config/config.yaml b/config/config.yaml
@@ -1,7 +1,7 @@
 # Config file for pipeline run.
 
 # DEVICE SETUP.
-device: 'cuda'
+device: 'cpu'
 
 # EXPERIMENT.
 experiment:
@@ -16,8 +16,7 @@ data:
     num_workers: 1
 
     train_val_test:
-        full_datapath: '/path/to/anndata.h5ad'
-
+        full_datapath: 'path/to/adata.h5ad'
         splitter_config:
             name: GroupSplitter
             params:
@@ -35,7 +34,7 @@ data:
     #       params: 
     #             **args
 
-    target: Cell_Type    
+    target: cell_type    
 
 
 # FEATURE SELECTION.
diff --git a/pyproject.toml b/pyproject.toml
@@ -6,14 +6,14 @@ build-backend = "hatchling.build"
 
 name = "pyscaLR"
 version = "1.1.0"
-requires-python = ">=3.10"
+requires-python = ">=3.10.20"
 authors = [
   { name="Infocusp", email="saurabh@infocusp.com" },
 ]
 description = "scaLR: Single cell analysis using low resource."
 readme = "README.md"
 classifiers = [
-    "Programming Language :: Python :: 3.9",
+    "Programming Language :: Python :: 3.10",
     "Operating System :: OS Independent",
     "Intended Audience :: Science/Research"
 ]
diff --git a/requirements.txt b/requirements.txt
@@ -1,4 +1,4 @@
-anndata==0.10.9
+anndata>=0.11.2,<0.12
 isort==5.13.2
 loky==3.4.1
 memory-profiler==0.61.0
@@ -15,4 +15,4 @@ tensorboard==2.17.0
 toml==0.10.2
 torch==2.4.1 --index-url https://download.pytorch.org/whl/cu118
 tqdm==4.66.5
-yapf==0.40.2
+yapf==0.40.2
diff --git a/scalr/analysis/dge_lmem.py b/scalr/analysis/dge_lmem.py
@@ -11,7 +11,6 @@
 
 from anndata import AnnData
 from anndata import ImplicitModificationWarning
-import anndata as ad
 from anndata.experimental import AnnCollection
 from joblib import delayed
 from joblib import Parallel
diff --git a/scalr/analysis/dge_pseudobulk.py b/scalr/analysis/dge_pseudobulk.py
@@ -4,8 +4,7 @@
 from os import path
 from typing import Optional, Tuple, Union
 
-from anndata import AnnData
-import anndata as ad
+from anndata import AnnData, concat
 from anndata.experimental import AnnCollection
 import matplotlib.pyplot as plt
 import numpy as np
@@ -94,15 +93,15 @@ def _make_design_matrix(self, adata: AnnData, cell_type: str):
             for sum_sample in condition_subset.obs[self.sum_column].unique():
                 sum_subset = condition_subset[condition_subset.obs[
                     self.sum_column] == sum_sample]
-                subdata = ad.AnnData(
+                subdata = AnnData(
                     X=sum_subset[:].X.sum(axis=0).reshape(
                         1, len(sum_subset.var_names)),
                     var=DataFrame(index=sum_subset.var_names),
                     obs=DataFrame(index=[f'{sum_sample}_{condition}']))
                 subdata.obs[self.design_factor_no_undrscr] = [condition]
                 design_matrix_list.append(subdata)
 
-        design_matrix = ad.concat(design_matrix_list)
+        design_matrix = concat(design_matrix_list)
         return design_matrix
 
     def get_differential_expression_results(self, design_matrix: AnnData,
diff --git a/scalr/nn/dataloader/simple_metadataloader.py b/scalr/nn/dataloader/simple_metadataloader.py
@@ -71,7 +71,7 @@ def collate_fn(
             x = torch.cat(
                 (x,
                  torch.as_tensor(self.metadata_onehotencoder[col].transform(
-                     adata_batch.obs[col].values.reshape(-1, 1)).A,
+                     adata_batch.obs[col].values.reshape(-1, 1)).toarray(),
                                  dtype=torch.float32)),
                 dim=1)
         return x, y
diff --git a/scalr/nn/dataloader/test_simple_metadataloader.py b/scalr/nn/dataloader/test_simple_metadataloader.py
@@ -1,6 +1,5 @@
 '''This is a test file for simplemetadataloader.'''
 
-import anndata
 import numpy as np
 import pandas as pd
 
diff --git a/scalr/utils/data_utils.py b/scalr/utils/data_utils.py
@@ -46,7 +46,7 @@ def get_random_samples(
     random_background_data = data[random_indices].X
 
     if not isinstance(random_background_data, np.ndarray):
-        random_background_data = random_background_data.A
+        random_background_data = random_background_data.toarray()
 
     random_background_data = torch.as_tensor(random_background_data,
                                              dtype=torch.float32)
diff --git a/scalr/utils/file_utils.py b/scalr/utils/file_utils.py
@@ -7,8 +7,8 @@
 from typing import Union
 
 from anndata import AnnData
-import anndata as ad
 from anndata.experimental import AnnCollection
+from anndata.io import read_h5ad
 from joblib import delayed
 from joblib import Parallel
 import numpy as np
@@ -142,7 +142,7 @@ def transform_and_write_data(data: AnnData, chunk_number: int):
             if transform:
                 data = AnnData(data.X, obs=data.obs, var=data.var)
                 if not isinstance(data.X, np.ndarray):
-                    data.X = data.X.A
+                    data.X = data.X.toarray()
                 data.X = transform(data.X)
 
             write_data(data, path.join(dirpath, f'{chunk_number}.h5ad'))
@@ -262,7 +262,7 @@ def read_csv(filepath: str, index_col: int = 0) -> pd.DataFrame:
 
 def read_anndata(filepath: str, backed: str = 'r') -> AnnData:
     """This file returns the Anndata object from filepath."""
-    data = ad.read_h5ad(filepath, backed=backed)
+    data = read_h5ad(filepath, backed=backed)
     return data
 
 
diff --git a/tutorials/analysis/differential_gene_expression/dge_lmem_main.py b/tutorials/analysis/differential_gene_expression/dge_lmem_main.py
@@ -13,14 +13,13 @@
 
 from anndata import AnnData
 from anndata import ImplicitModificationWarning
-import anndata as ad
 from anndata.experimental import AnnCollection
+from anndata.io import read_h5ad
 from joblib import Parallel, delayed
 import matplotlib.pyplot as plt
 import numpy as np
 import pandas as pd
 from pandas import DataFrame
-import scanpy as sc
 from scipy.optimize import OptimizeWarning
 import statsmodels.api as sm
 import statsmodels.formula.api as smf
@@ -31,7 +30,7 @@
 
 
 def main(config):
-    test_data = sc.read_h5ad(config['full_datapath'], backed='r')
+    test_data = read_h5ad(config['full_datapath'], backed='r')
     dirpath = config['dirpath']
     dge_type = config['dge_type']
     assert (dge_type == 'DgeLMEM') and ('lmem_params' in config), (
diff --git a/tutorials/analysis/differential_gene_expression/dge_pseudobulk_main.py b/tutorials/analysis/differential_gene_expression/dge_pseudobulk_main.py
@@ -5,22 +5,21 @@
 from typing import Optional, Union, Tuple
 import yaml
 
-import anndata as ad
 from anndata import AnnData
 from anndata.experimental import AnnCollection
+from anndata.io import read_h5ad
 import matplotlib.pyplot as plt
 import numpy as np
 import pandas as pd
 from pandas import DataFrame
 from pydeseq2.dds import DeseqDataSet
 from pydeseq2.ds import DeseqStats
-import scanpy as sc
 
 from scalr.analysis import DgePseudoBulk
 
 
 def main(config):
-    test_data = sc.read_h5ad(config['full_datapath'], backed='r')
+    test_data = read_h5ad(config['full_datapath'], backed='r')
     dirpath = config['dirpath']
     dge_type = config['dge_type']
     assert (dge_type == 'DgePseudoBulk') and ('psedobulk_params' in config), (
diff --git a/tutorials/pipeline/config_celltype.yaml b/tutorials/pipeline/config_celltype.yaml
@@ -1,7 +1,7 @@
 # Config file for pipeline run for cell type classification.
 
 # DEVICE SETUP.
-device: 'cuda'
+device: 'cpu'
 
 # EXPERIMENT.
 experiment:
@@ -15,8 +15,8 @@ data:
     sample_chunksize: 20000
 
     train_val_test:
-        full_datapath: 'data/modified_adata.h5ad'
-        num_workers: 2
+        full_datapath: 'path/to/adata.h5ad'
+        num_workers: 4
 
         splitter_config:
             name: GroupSplitter
diff --git a/tutorials/pipeline/config_clinical.yaml b/tutorials/pipeline/config_clinical.yaml
@@ -1,7 +1,7 @@
 # Config file for pipeline run for clinical condition specific biomarker identification.
 
 # DEVICE SETUP.
-device: 'cuda'
+device: 'cpu'
 
 # EXPERIMENT.
 experiment:
@@ -15,7 +15,7 @@ data:
     sample_chunksize: 20000
 
     train_val_test:
-        full_datapath: 'data/modified_adata.h5ad'
+        full_datapath: 'path/to/adata.h5ad'
         num_workers: 2
 
         splitter_config:
diff --git a/tutorials/pipeline/scalr_pipeline.ipynb b/tutorials/pipeline/scalr_pipeline.ipynb
@@ -358,7 +358,7 @@
    "outputs": [],
    "source": [
     "#Gene expression values of first 5 cells and 10 genes.\n",
-    "adata.X[:5,:10].A"
+    "adata.X[:5,:10]\n"
    ]
   },
   {
@@ -385,7 +385,7 @@
    "source": [
     "# Verifying normalized values in X\n",
     "# Getting the sum of gene expression values for the first 10 cells (should be floating-point values).\n",
-    "adata.X[:10,:].A.sum(axis=1)"
+    "adata.X[:10,:].sum(axis=1)"
    ]
   },
   {
@@ -411,8 +411,8 @@
    "outputs": [],
    "source": [
     "# Getting the maximum and minimum gene expression values for the first 1000 cells.\n",
-    "max_val = np.max(adata.X[:1000, :].A)\n",
-    "min_val = np.min(adata.X[:1000, :].A)\n",
+    "max_val = np.max(adata.X[:1000, :])\n",
+    "min_val = np.min(adata.X[:1000, :])\n",
     "print(f'Max value : {max_val} | Min value : {min_val}')\n",
     "# Raising a warning if the values are outside the 0-10 range\n",
     "if max_val > 10 or min_val < 0:\n",
diff --git a/tutorials/pipeline/scalr_pipeline_local_run.ipynb b/tutorials/pipeline/scalr_pipeline_local_run.ipynb
