[python] Support obs_id_name and var_id_name in tiledbsoma.io.from_anndata

.github/workflows/r-ci.yml

@@ -67,6 +67,9 @@ jobs:
       #  if: ${{ matrix.os == 'ubuntu-latest' }}
       #  run: sudo ldconfig
 
+      - name: Update packages
+        run: Rscript -e 'update.packages(ask=FALSE)'
+
       - name: Test
         if: ${{ matrix.covr == 'no' }}
         run: cd apis/r && tools/r-ci.sh run_tests
@@ -75,6 +78,10 @@ jobs:
         run: cat $HOME/work/TileDB-SOMA/TileDB-SOMA/apis/r/tiledbsoma.Rcheck/00install.out
         if: failure()
 
+      - name: View Test Output
+        run: cat $HOME/work/TileDB-SOMA/TileDB-SOMA/apis/r/tiledbsoma.Rcheck/00check.log
+        if: failure()
+
       - name: Coverage
         if: ${{ matrix.os == 'ubuntu-latest' && matrix.covr == 'yes' }}
         run: cd apis/r && tools/r-ci.sh coverage

.github/workflows/r-python-interop-testing.yml

@@ -80,5 +80,8 @@ jobs:
           python -c 'import tiledbsoma; tiledbsoma.show_package_versions()'
           python scripts/show-versions.py
 
+      - name: Update Packages
+        run: Rscript -e 'update.packages(ask=FALSE)'
+
       - name: Interop Tests
         run: python -m pytest apis/system/tests/

apis/python/src/tiledbsoma/io/ingest.py

@@ -233,6 +233,8 @@ def from_h5ad(
     *,
     context: Optional[SOMATileDBContext] = None,
     platform_config: Optional[PlatformConfig] = None,
+    obs_id_name: str = "obs_id",
+    var_id_name: str = "var_id",
     ingest_mode: IngestMode = "write",
     use_relative_uri: Optional[bool] = None,
     X_kind: Union[Type[SparseNDArray], Type[DenseNDArray]] = SparseNDArray,
@@ -326,6 +328,8 @@ def from_h5ad(
             measurement_name,
             context=context,
             platform_config=platform_config,
+            obs_id_name=obs_id_name,
+            var_id_name=var_id_name,
             ingest_mode=ingest_mode,
             use_relative_uri=use_relative_uri,
             X_kind=X_kind,
@@ -346,6 +350,8 @@ def from_anndata(
     *,
     context: Optional[SOMATileDBContext] = None,
     platform_config: Optional[PlatformConfig] = None,
+    obs_id_name: str = "obs_id",
+    var_id_name: str = "var_id",
     ingest_mode: IngestMode = "write",
     use_relative_uri: Optional[bool] = None,
     X_kind: Union[Type[SparseNDArray], Type[DenseNDArray]] = SparseNDArray,
@@ -422,7 +428,7 @@ def from_anndata(
     with _write_dataframe(
         df_uri,
         conversions.decategoricalize_obs_or_var(anndata.obs),
-        id_column_name="obs_id",
+        id_column_name=obs_id_name,
         platform_config=platform_config,
         context=context,
         ingestion_params=ingestion_params,
@@ -751,6 +757,7 @@ def append_obs(
     exp: Experiment,
     new_obs: pd.DataFrame,
     *,
+    obs_id_name: str = "obs_id",
     registration_mapping: ExperimentAmbientLabelMapping,
     context: Optional[SOMATileDBContext] = None,
     platform_config: Optional[PlatformConfig] = None,
@@ -792,7 +799,7 @@ def append_obs(
     with _write_dataframe(
         exp.obs.uri,
         conversions.decategoricalize_obs_or_var(new_obs),
-        id_column_name="obs_id",
+        id_column_name=obs_id_name,
         platform_config=platform_config,
         context=context,
         ingestion_params=ingestion_params,
@@ -809,6 +816,7 @@ def append_var(
     new_var: pd.DataFrame,
     measurement_name: str,
     *,
+    var_id_name: str = "var_id",
     registration_mapping: ExperimentAmbientLabelMapping,
     context: Optional[SOMATileDBContext] = None,
     platform_config: Optional[PlatformConfig] = None,

apis/python/tests/test_basic_anndata_io.py

@@ -744,3 +744,77 @@ def test_obs_with_categorical_int_nan_enumeration(
     tiledbsoma.io.from_h5ad(
         output_path, h5ad_file_categorical_int_nan, measurement_name="RNA"
     )
+
+
+@pytest.mark.parametrize("obs_id_name", ["obs_id", "cells_are_great"])
+@pytest.mark.parametrize("var_id_name", ["var_id", "genes_are_nice_too"])
+@pytest.mark.parametrize("indexify_obs", [True, False])
+@pytest.mark.parametrize("indexify_var", [True, False])
+def test_id_names(tmp_path, obs_id_name, var_id_name, indexify_obs, indexify_var):
+    obs_ids = ["AAAT", "CATG", "CTGA", "TCTG", "TGAG", "TTTG"]
+    var_ids = ["AKT1", "APOE", "ESR1", "TP53", "VEGFA", "ZZZ3"]
+
+    n_obs = len(obs_ids)
+    n_var = len(var_ids)
+
+    obs = pd.DataFrame(
+        data={
+            obs_id_name: np.asarray(obs_ids),
+            "cell_type": pd.Categorical(
+                [["B cell", "T cell"][e % 2] for e in range(n_obs)],
+                categories=["B cell", "T cell"],
+                ordered=True,
+            ),
+        },
+        index=np.arange(n_obs).astype(str),
+    )
+    if indexify_obs:
+        obs.set_index(obs_id_name, inplace=True)
+
+    var = pd.DataFrame(
+        data={
+            var_id_name: np.asarray(var_ids),
+            "counter": np.asarray(range(n_var), dtype=np.float32),
+        },
+        index=np.arange(n_var).astype(str),
+    )
+    if indexify_var:
+        var.set_index(var_id_name, inplace=True)
+
+    X = np.zeros([n_obs, n_var])
+    for i in range(n_obs):
+        for j in range(n_var):
+            if (i + j) % 2 == 1:
+                X[i, j] = 100 + 10 * i + j
+
+    adata = anndata.AnnData(X=X, obs=obs, var=var, dtype=X.dtype)
+
+    uri = tmp_path.as_posix()
+
+    # Implicitly, a check for no-throw
+    tiledbsoma.io.from_anndata(
+        uri,
+        adata,
+        measurement_name="RNA",
+        obs_id_name=obs_id_name,
+        var_id_name=var_id_name,
+    )
+
+    with tiledbsoma.Experiment.open(uri) as exp:
+        assert obs_id_name in exp.obs.keys()
+        assert var_id_name in exp.ms["RNA"].var.keys()
+
+        # Implicitly, a check for no-throw
+        bdata = tiledbsoma.io.to_anndata(
+            exp,
+            measurement_name="RNA",
+            obs_id_name=obs_id_name,
+            var_id_name=var_id_name,
+        )
+
+        soma_obs = exp.obs.read(column_names=[obs_id_name]).concat().to_pandas()
+        soma_var = (
+            exp.ms["RNA"].var.read(column_names=[var_id_name]).concat().to_pandas()
+        )
+        assert list(bdata.obs.index) == list(soma_obs[obs_id_name])
+        assert list(bdata.var.index) == list(soma_var[var_id_name])