G4X sample dataset

g4x_io/README.md

@@ -0,0 +1,24 @@
+# G4X Sample Data
+
+Download and convert a sample dataset from the Singular Genomics G4X Platform to SpatialData. The data will include:
+
+- images:
+  - `h_and_e`: H&E image in RGB
+  - `nuclear`: nuclear stain
+  - `eosin`: eosin stain
+  - `protein`: multi-channel stack of 20 protein images
+- labels:
+  - `nuclei`: Nuclei segmentation instance labels
+  - `nuclei_exp`: Cell segmentation instance labels
+- points:
+  - `transcripts`: Single-molecule transcript locations, gene identities, and metadata
+- shapes:
+  - `nuclei_shapes`: Nuclei segmentation polygons
+  - `nuclei_exp`: Cell segmentation polygons
+- table:
+  - `table`: the cell by gene expression count matrix with cell metadata
+
+### Download
+1. Make a folder `g4x_io` and cd into it.
+1. Download sample dataset with `download.py`
+3. Convert the data into the SpatialData format with `to_zarr.py`

g4x_io/download.py

@@ -0,0 +1,25 @@
+##
+import os
+from pathlib import Path
+import subprocess
+
+# Singular Genomics sample data
+
+urls = [
+    "https://singular-public-repo.s3.us-west-1.amazonaws.com/g4x_tutorial_dataset.zip"
+]
+
+# download the data
+for url in urls:
+    filename = Path(url).name
+    os.makedirs("data", exist_ok=True)
+    command = f"curl -o {'data/' + filename} {url}"
+    subprocess.run(command, shell=True, check=True)
+
+##
+# unzip the data
+subprocess.run(
+    "unzip -o data/g4x_tutorial_dataset.zip -d data/",
+    shell=True,
+    check=True,
+)

g4x_io/to_zarr.py

@@ -0,0 +1,36 @@
+##
+from spatialdata_io import g4x
+import spatialdata as sd
+
+##
+from pathlib import Path
+import shutil
+
+##
+path = Path().resolve()
+# luca's workaround for pycharm
+if not str(path).endswith("g4x_io"):
+    path /= "g4x_io"
+    assert path.exists()
+
+path_read = path / "data"
+path_write = path / "data.zarr"
+
+##
+print("parsing the data... ", end="")
+sdata = g4x(
+    input_path=str(path_read),
+    output_path=str(path_write),
+    include_he=True,
+    include_segmentation=True,
+    include_protein=True,
+    include_transcripts=True,
+    include_tables=True,
+    mode="append",
+)
+print("done")
+
+##
+sdata = sd.SpatialData.read("./data.zarr/")
+print(sdata)
+##