Merge pull request #19 from crs4/data_entity_types

Snakemake: add SoftwareSourceCode type to some entities
crs4 · Nov 16, 2023 · 789b35f · 789b35f
2 parents 7856d96 + 342a37f
commit 789b35f
Show file tree

Hide file tree

Showing 4 changed files with 18 additions and 15 deletions.
diff --git a/repo2rocrate/common.py b/repo2rocrate/common.py
@@ -16,6 +16,7 @@
 from pathlib import Path
 
 from rocrate.rocrate import ROCrate
+from rocrate.utils import as_list
 from .utils import get_ci_wf_endpoint
 
 GH_API_URL = "https://api.github.com"
@@ -127,11 +128,13 @@ def add_data_entities(self, data_entities=None):
             if self.crate.get(str(relpath)):
                 continue  # existing entity is more specific
             source = self.root / relpath
-            properties = {"description": description} if description else None
-            if type_ == "File":
+            properties = {"@type": type_}
+            if description:
+                properties["description"] = description
+            if "File" in as_list(type_):
                 if source.is_file():
                     self.crate.add_file(source, relpath, properties=properties)
-            elif type_ == "Dataset":
+            elif "Dataset" in as_list(type_):
                 if source.is_dir():
                     self.crate.add_dataset(source, relpath, properties=properties)
             else:

diff --git a/repo2rocrate/snakemake.py b/repo2rocrate/snakemake.py
@@ -51,16 +51,16 @@ def get_lang_version(workflow_path):
 class SnakemakeCrateBuilder(CrateBuilder):
 
     DATA_ENTITIES = [
-        (".tests/integration", "Dataset", "Integration tests for the workflow"),
-        (".tests/unit", "Dataset", "Unit tests for the workflow"),
-        ("workflow", "Dataset", "Workflow folder"),
+        (".tests/integration", ["Dataset", "SoftwareSourceCode"], "Integration tests for the workflow"),
+        (".tests/unit", ["Dataset", "SoftwareSourceCode"], "Unit tests for the workflow"),
+        ("workflow", ["Dataset", "SoftwareSourceCode"], "Workflow folder"),
         ("config", "Dataset", "Configuration folder"),
         ("results", "Dataset", "Output files"),
         ("resources", "Dataset", "Retrieved resources"),
-        ("workflow/rules", "Dataset", "Workflow rule modules"),
+        ("workflow/rules", ["Dataset", "SoftwareSourceCode"], "Workflow rule modules"),
         ("workflow/envs", "Dataset", "Conda environments"),
-        ("workflow/scripts", "Dataset", "Scripts folder"),
-        ("workflow/notebooks", "Dataset", "Notebooks folder"),
+        ("workflow/scripts", ["Dataset", "SoftwareSourceCode"], "Scripts folder"),
+        ("workflow/notebooks", ["Dataset", "SoftwareSourceCode"], "Notebooks folder"),
         ("workflow/report", "Dataset", "Report caption files"),
         ("workflow/schemas", "Dataset", "Validation files"),
     ]

diff --git a/test/test_cli.py b/test/test_cli.py
@@ -54,10 +54,10 @@ def test_default(data_dir, tmpdir, monkeypatch, to_zip):
         ("LICENSE", "File"),
         ("README.md", "File"),
         ("config", "Dataset"),
-        (".tests/integration", "Dataset"),
-        ("workflow/rules", "Dataset"),
+        (".tests/integration", ["Dataset", "SoftwareSourceCode"]),
+        ("workflow/rules", ["Dataset", "SoftwareSourceCode"]),
         ("workflow/schemas", "Dataset"),
-        ("workflow/scripts", "Dataset"),
+        ("workflow/scripts", ["Dataset", "SoftwareSourceCode"]),
     ]
     for relpath, type_ in expected_data_entities:
         entity = crate.get(relpath)

diff --git a/test/test_snakemake.py b/test/test_snakemake.py
@@ -113,10 +113,10 @@ def test_make_crate(data_dir, defaults):
         ("LICENSE", "File", ""),
         ("README.md", "File", ""),
         ("config", "Dataset", "Configuration folder"),
-        (".tests/integration", "Dataset", "Integration tests for the workflow"),
-        ("workflow/rules", "Dataset", "Workflow rule modules"),
+        (".tests/integration", ["Dataset", "SoftwareSourceCode"], "Integration tests for the workflow"),
+        ("workflow/rules", ["Dataset", "SoftwareSourceCode"], "Workflow rule modules"),
         ("workflow/schemas", "Dataset", "Validation files"),
-        ("workflow/scripts", "Dataset", "Scripts folder"),
+        ("workflow/scripts", ["Dataset", "SoftwareSourceCode"], "Scripts folder"),
     ]
     for relpath, type_, desc in expected_data_entities:
         entity = crate.get(relpath)