Slurm runner

README.md

@@ -12,7 +12,7 @@ datasets but can be generalized to store any 2-dimensional experimental data.
 
 To get started, install the package from [PyPI](https://pypi.org/project/cellarr/)
 
-```bash
+```sh
 pip install cellarr
 
 ## to include optional dependencies
@@ -123,6 +123,52 @@ print(dataset)
 
 Check out the [documentation](https://biocpy.github.io/cellarr/tutorial.html) for more details.
 
+### Building on HPC environments with `slurm`
+
+To simplify building TileDB files on HPC environments that use `slurm`, there are a few steps you need to follow.
+
+- Step 1: Construct a manifest file
+A minimal manifest file (json) must contain the following fields
+- `"files"`: A list of file path to the input `h5ad` objects.
+- `"python_env"`: A set of commands to activate the Python environment containing this package and its dependencies.
+
+Here’s an example of the manifest file:
+
+```py
+manifest = {
+    "files": your/list/of/files,
+    "python_env": """
+ml Miniforge3
+conda activate cellarr
+
+python --version
+which python
+    """,
+    "matrix_options": [
+        {
+            "matrix_name": "non_zero_cells",
+            "dtype": "uint32"
+        },
+        {
+            "matrix_name": "pseudo_bulk_log_normed",
+            "dtype": "float32"
+        }
+    ],
+}
+
+import json
+json.dump(manifest, open("your/path/to/manifest.json", "w"))
+```
+
+For more options, check out the [README](./src/cellarr/slurm/README.md).
+
+- Step 2: Submit the job
+Once your manifest file is ready, you can submit the necessary jobs using the `cellarr_build` CLI. Run the following command:
+
+```sh
+cellarr_build --input-manifest your/path/to/manifest.json --output-dir your/path/to/output --memory-per-job 8 --cpus-per-task 2
+```
+
 ### Query a `CellArrDataset`
 
 Users have the option to reuse the `dataset` object returned when building the dataset or by creating a `CellArrDataset` object by initializing it to the path where the files were created.

setup.cfg

@@ -80,8 +80,8 @@ testing =
 
 [options.entry_points]
 # Add here console scripts like:
-# console_scripts =
-#     script_name = cellarr.module:function
+console_scripts =
+    cellarr_build = cellarr.slurm.build_cellarr_steps:main
 # For example:
 # console_scripts =
 #     fibonacci = cellarr.skeleton:run

src/cellarr/buildutils_tiledb_frame.py

@@ -76,7 +76,7 @@ def create_tiledb_frame_from_column_names(
     )
 
 
-def create_tiledb_frame_from_dataframe(tiledb_uri_path: str, frame: List[str], column_types=dict):
+def create_tiledb_frame_from_dataframe(tiledb_uri_path: str, frame: List[str], column_types: dict = None):
     """Create a TileDB file with the provided attributes to persistent storage.
 
     This will materialize the array directory and all

src/cellarr/slurm/README.md

@@ -0,0 +1,65 @@
+
+# manifest json
+
+```json
+{
+    "files": [
+        "/path/to/dataset1.h5ad",
+        "/path/to/dataset2.h5ad"
+    ],
+    "matrix_options": [
+        {
+            "matrix_name": "counts",
+            "dtype": "uint32"
+        },
+        {
+            "matrix_name": "normalized",
+            "dtype": "float32"
+        }
+    ],
+    "gene_options": {
+        "feature_column": "index"
+    },
+    "sample_options": {
+        "metadata": {
+            "sample_1": {
+                "condition": "control",
+                "batch": "1"
+            },
+            "sample_2": {
+                "condition": "treatment",
+                "batch": "1"
+            }
+        }
+    },
+    "cell_options": {
+        "column_types": {
+            "cell_type": "ascii",
+            "quality_score": "float32"
+        },
+    },
+    "python_env": """
+. /system/gredit/clientos/etc/profile
+
+ml Miniforge3
+conda activate biocpy_miniforge
+
+~/.conda/envs/biocpy_miniforge/bin/python --version
+which python
+python --version
+    """,
+}
+```
+
+
+Run
+
+```sh
+
+python build_cellarr_steps.py \
+    --input-manifest manifest.json \
+    --output-dir /path/to/output \
+    --memory-per-job 64 \
+    --cpus-per-task 4
+
+```