Merge branch 'main' into dependabot/submodules/bids-examples-d039879

.github/workflows/cd.yml

@@ -21,10 +21,10 @@ jobs:
           password: ${{ secrets.DOCKERHUB_TOKEN }}
       -
         name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v2
+        uses: docker/setup-buildx-action@v3
       -
         name: Build and push
-        uses: docker/build-push-action@v4
+        uses: docker/build-push-action@v5
         with:
           context: .
           file: ./Dockerfile

.github/workflows/test.yml

@@ -35,6 +35,6 @@ jobs:
         coverage run -m pytest
         coverage lcov -o ./coverage/lcov.info
     - name: Coveralls GitHub Action
-      uses: coverallsapp/github-action@v2.2.1
+      uses: coverallsapp/github-action@v2.2.3
       with:
         github-token: ${{ secrets.GITHUB_TOKEN }}

bagel/cli.py

@@ -9,7 +9,7 @@
 import bagel.bids_utils as butil
 import bagel.pheno_utils as putil
 from bagel import mappings, models
-from bagel.utility import load_json
+from bagel.utility import check_overwrite, load_json
 
 bagel = typer.Typer()
 
@@ -32,14 +32,6 @@ def pheno(
         dir_okay=False,
         resolve_path=True,
     ),
-    output: Path = typer.Option(
-        ...,
-        help="The directory where outputs should be created.",
-        exists=True,
-        file_okay=False,
-        dir_okay=True,
-        resolve_path=True,
-    ),
     name: str = typer.Option(
         ...,
         help="A descriptive name for the dataset the input belongs to. "
@@ -51,6 +43,18 @@ def pheno(
         callback=putil.validate_portal_uri,
         help="URL (HTTP/HTTPS) to a website or page that describes the dataset and access instructions (if available).",
     ),
+    output: Path = typer.Option(
+        default="pheno.jsonld",
+        help="The path for the output .jsonld file.",
+        file_okay=True,
+        dir_okay=False,
+        resolve_path=True,
+    ),
+    overwrite: bool = typer.Option(
+        False,
+        "--overwrite",
+        help="Overwrite output file if it already exists.",
+    ),
 ):
     """
     Process a tabular phenotypic file (.tsv) that has been successfully annotated
@@ -61,6 +65,9 @@ def pheno(
     graph datamodel for the provided phenotypic file in the .jsonld format.
     You can upload this .jsonld file to the Neurobagel graph.
     """
+    # Check if output file already exists
+    check_overwrite(output, overwrite)
+
     data_dictionary = load_json(dictionary)
     pheno_df = pd.read_csv(pheno, sep="\t", keep_default_na=False, dtype=str)
     putil.validate_inputs(data_dictionary, pheno_df)
@@ -140,11 +147,10 @@ def pheno(
     # TODO: we should revisit this because there may be reasons to have None be meaningful in the future
     context.update(**dataset.dict(exclude_none=True))
 
-    output_filename = output / "pheno.jsonld"
-    with open(output_filename, "w") as f:
+    with open(output, "w") as f:
         f.write(json.dumps(context, indent=2))
 
-    print(f"Saved output to:  {output_filename}")
+    print(f"Saved output to:  {output}")
 
 
 @bagel.command()
@@ -166,13 +172,17 @@ def bids(
         resolve_path=True,
     ),
     output: Path = typer.Option(
-        ...,
-        help="The directory where outputs should be created",
-        exists=True,
-        file_okay=False,
-        dir_okay=True,
+        help="The path for the output .jsonld file.",
+        default="pheno_bids.jsonld",
+        file_okay=True,
+        dir_okay=False,
         resolve_path=True,
     ),
+    overwrite: bool = typer.Option(
+        False,
+        "--overwrite",
+        help="Overwrite output file if it already exists.",
+    ),
 ):
     """
     Extract imaging metadata from a valid BIDS dataset and combine them
@@ -183,6 +193,9 @@ def bids(
     graph datamodel for the combined metadata in the .jsonld format.
     You can upload this .jsonld file to the Neurobagel graph.
     """
+    # Check if output file already exists
+    check_overwrite(output, overwrite)
+
     jsonld = load_json(jsonld_path)
     layout = BIDSLayout(bids_dir, validate=True)
 
@@ -264,8 +277,7 @@ def bids(
 
     merged_dataset = {**context, **pheno_dataset.dict(exclude_none=True)}
 
-    output_filename = output / "pheno_bids.jsonld"
-    with open(output_filename, "w") as f:
+    with open(output, "w") as f:
         f.write(json.dumps(merged_dataset, indent=2))
 
-    print(f"Saved output to:  {output_filename}")
+    print(f"Saved output to:  {output}")

bagel/tests/test_cli_bids.py

@@ -1,10 +1,21 @@
 from pathlib import Path
 
+import pytest
+
 from bagel.cli import bagel
 
 
+@pytest.fixture(scope="function")
+def default_pheno_bids_output_path(tmp_path):
+    "Return temporary bids command output filepath that uses the default filename."
+    return tmp_path / "pheno_bids.jsonld"
+
+
 def test_bids_valid_inputs_run_successfully(
-    runner, test_data_upload_path, bids_synthetic, tmp_path
+    runner,
+    test_data_upload_path,
+    bids_synthetic,
+    default_pheno_bids_output_path,
 ):
     """Basic smoke test for the "add-bids" subcommand"""
     result = runner.invoke(
@@ -16,20 +27,20 @@ def test_bids_valid_inputs_run_successfully(
             "--bids-dir",
             bids_synthetic,
             "--output",
-            tmp_path,
+            default_pheno_bids_output_path,
         ],
     )
     assert result.exit_code == 0, f"Errored out. STDOUT: {result.output}"
     assert (
-        tmp_path / "pheno_bids.jsonld"
+        default_pheno_bids_output_path
     ).exists(), "The pheno_bids.jsonld output was not created."
 
 
 def test_bids_sessions_have_correct_labels(
     runner,
     test_data_upload_path,
     bids_synthetic,
-    tmp_path,
+    default_pheno_bids_output_path,
     load_test_json,
 ):
     """Check that sessions added to pheno_bids.jsonld have the expected labels."""
@@ -42,11 +53,11 @@ def test_bids_sessions_have_correct_labels(
             "--bids-dir",
             bids_synthetic,
             "--output",
-            tmp_path,
+            default_pheno_bids_output_path,
         ],
     )
 
-    pheno_bids = load_test_json(tmp_path / "pheno_bids.jsonld")
+    pheno_bids = load_test_json(default_pheno_bids_output_path)
     for sub in pheno_bids["hasSamples"]:
         assert ["ses-01", "ses-02"] == [
             ses["hasLabel"] for ses in sub["hasSession"]
@@ -56,7 +67,7 @@ def test_bids_sessions_have_correct_labels(
 def test_bids_data_with_sessions_have_correct_paths(
     runner,
     test_data_upload_path,
-    tmp_path,
+    default_pheno_bids_output_path,
     load_test_json,
 ):
     """
@@ -72,11 +83,11 @@ def test_bids_data_with_sessions_have_correct_paths(
             "--bids-dir",
             Path(__file__).parent / "../../bids-examples/synthetic",
             "--output",
-            tmp_path,
+            default_pheno_bids_output_path,
         ],
     )
 
-    pheno_bids = load_test_json(tmp_path / "pheno_bids.jsonld")
+    pheno_bids = load_test_json(default_pheno_bids_output_path)
     for sub in pheno_bids["hasSamples"]:
         for ses in sub["hasSession"]:
             assert sub["hasLabel"] in ses["hasFilePath"]