deriva-ml 1.14.33__tar.gz → 1.14.34__tar.gz

{deriva_ml-1.14.33 → deriva_ml-1.14.34}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: deriva-ml
-Version: 1.14.33
+Version: 1.14.34
 Summary: Utilities to simplify use of Dervia and Pandas to create reproducable ML pipelines
 Author-email: ISRD <isrd-dev@isi.edu>
 Requires-Python: >=3.10
@@ -25,10 +25,9 @@ Deriva-ML is a python library to simplify the process of creating and executing
 using a deriva catalog.
 
 
-## Installing the GitHub CLI
+Complete on-line documentation for DerivaML can be found [here](https://informatics-isi-edu.github.io/deriva-ml/)
 
-The script release.sh will create a new release tag in GitHub.  This script requires the 
-GitHUB CLI be installed. 
+To get started using DerivaML, you can clone the [model template repository](https://github.com/informatics-isi-edu/deriva-ml-model-template), and modify it to suite your requirements.
 
-See [https://cli.github.com](https://cli.github.com) for instructions on how to install and configure the CLI.
 
+## References

deriva_ml-1.14.34/README.md

@@ -0,0 +1,11 @@
+# DerivaML
+Deriva-ML is a python library to simplify the process of creating and executing reproducible machine learning workflows
+using a deriva catalog.
+
+
+Complete on-line documentation for DerivaML can be found [here](https://informatics-isi-edu.github.io/deriva-ml/)
+
+To get started using DerivaML, you can clone the [model template repository](https://github.com/informatics-isi-edu/deriva-ml-model-template), and modify it to suite your requirements.
+
+
+## References

{deriva_ml-1.14.33 → deriva_ml-1.14.34}/pyproject.toml

@@ -64,6 +64,9 @@ setup_hooks = []
 pre_commit_hooks = []
 post_commit_hooks = []
 
+[tool.pytest]
+mock_use_standalone_module = true
+
 [tool.pytest.ini_options]
 testpaths = ["tests"]
 python_files = ["test_*.py"]
@@ -97,6 +100,7 @@ dev = [
     "mkdocstrings[python]",
     "mkdocs-material",
     "pytest>=8.4.1",
+    "pytest-mock",
     "pytest-coverage>=0.0",
     "ruff"
 ]

{deriva_ml-1.14.33 → deriva_ml-1.14.34}/src/deriva_ml/feature.py

@@ -52,6 +52,7 @@ class FeatureRecord(BaseModel):
 
     class Config:
         arbitrary_types_allowed = True
+        extra = "forbid"
 
     @classmethod
     def feature_columns(cls) -> set[Column]:

{deriva_ml-1.14.33 → deriva_ml-1.14.34}/src/deriva_ml/run_notebook.py

@@ -95,7 +95,7 @@ class DerivaMLRunNotebookCLI(BaseCLI):
             exit(1)
 
         os.environ["DERIVA_HOST"] = args.host
-        os.environ["DERIVA_CATALOG_ID"] = args.catalog
+        os.environ["DERIVA_CATALOG"] = args.catalog
 
         # Create a workflow instance for this specific version of the script.
         # Return an existing workflow if one is found.
@@ -106,14 +106,14 @@ class DerivaMLRunNotebookCLI(BaseCLI):
             return
         else:
             notebook_parameters = (
-                {"host": args.host, "catalog_id": args.catalog, "catalog": args.catalog}
-                | {k: v["default"] for k, v in notebook_parameters.items()}
+                {k: v["default"] for k, v in notebook_parameters.items()}
+                | {"host": args.host, "hostname": args.host, "catalog_id": args.catalog, "catalog": args.catalog}
                 | parameters
             )
             print(f"Running notebook {notebook_file.name} with parameters:")
             for param, value in notebook_parameters.items():
                 print(f"  {param}:{value}")
-            self.run_notebook(notebook_file.resolve(), parameters, kernel=args.kernel, log=args.log_output)
+            self.run_notebook(notebook_file.resolve(), parameters, kernel=args.kernel[0], log=args.log_output)
 
     def run_notebook(self, notebook_file, parameters, kernel=None, log=False):
         url, checksum = Workflow.get_url_and_checksum(Path(notebook_file))

{deriva_ml-1.14.33 → deriva_ml-1.14.34}/src/deriva_ml.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: deriva-ml
-Version: 1.14.33
+Version: 1.14.34
 Summary: Utilities to simplify use of Dervia and Pandas to create reproducable ML pipelines
 Author-email: ISRD <isrd-dev@isi.edu>
 Requires-Python: >=3.10
@@ -25,10 +25,9 @@ Deriva-ML is a python library to simplify the process of creating and executing
 using a deriva catalog.
 
 
-## Installing the GitHub CLI
+Complete on-line documentation for DerivaML can be found [here](https://informatics-isi-edu.github.io/deriva-ml/)
 
-The script release.sh will create a new release tag in GitHub.  This script requires the 
-GitHUB CLI be installed. 
+To get started using DerivaML, you can clone the [model template repository](https://github.com/informatics-isi-edu/deriva-ml-model-template), and modify it to suite your requirements.
 
-See [https://cli.github.com](https://cli.github.com) for instructions on how to install and configure the CLI.
 
+## References

{deriva_ml-1.14.33 → deriva_ml-1.14.34}/tests/conftest.py

@@ -5,7 +5,7 @@ Pytest configuration and shared fixtures.
 import os
 
 import pytest
-from test_utils import MLCatalog, MLDatasetCatalog
+from test_utils import MLCatalog, MLDatasetCatalog, create_jupyter_kernel, destroy_jupyter_kernel
 
 from deriva_ml import DerivaML
 from deriva_ml.demo_catalog import (
@@ -61,6 +61,16 @@ def dataset_test(catalog_with_datasets):
     return catalog_with_datasets
 
 
+@pytest.fixture(scope="function")
+def notebook_test(deriva_catalog, tmp_path):
+    deriva_catalog.reset_demo_catalog()
+    create_jupyter_kernel("test_kernel", tmp_path)
+    yield DerivaML(deriva_catalog.hostname, deriva_catalog.catalog_id, use_minid=False, working_dir=tmp_path)
+    print("Resetting catalog... ", end="")
+    deriva_catalog.reset_demo_catalog()
+    destroy_jupyter_kernel("test_kernel")
+
+
 @pytest.fixture(scope="function")
 def test_ml_demo_catalog(ml_catalog, tmp_path):
     # reset_demo_catalog(ml_catalog.catalog)

{deriva_ml-1.14.33 → deriva_ml-1.14.34}/tests/execution/test_execution.py

@@ -3,6 +3,7 @@ Tests for the execution module.
 """
 
 import subprocess
+from pathlib import Path
 from tempfile import TemporaryDirectory
 
 from deriva_ml import (
@@ -23,20 +24,23 @@ class TestWorkflow:
         ml_instance.add_term(vc.asset_type, "Test Model", description="Model for our Test workflow")
         ml_instance.add_term(vc.workflow_type, "Test Workflow", description="A ML Workflow that uses Deriva ML API")
         print("Running workflow-test.py ...")
+        workflow_script = Path(__file__).parent / "workflow-test.py"
+
         workflow_table = ml_instance.pathBuilder.schemas[ml_instance.ml_schema].Workflow
         workflows = list(workflow_table.entities().fetch())
         assert 0 == len(workflows)
         result = subprocess.run(
             [
                 "python",
-                "execution/workflow-test.py",
+                workflow_script.as_posix(),
                 ml_instance.catalog.deriva_server.server,
                 ml_instance.catalog_id,
             ],
             capture_output=True,
             text=True,
         )
-
+        print(result.stdout)
+        print(result.stderr)
         workflows = list(workflow_table.entities().fetch())
         assert 1 == len(workflows)
         workflow_rid = workflows[0]["RID"]
@@ -50,18 +54,22 @@ class TestWorkflow:
         result = subprocess.run(
             [
                 "python",
-                "execution/workflow-test.py",
+                workflow_script.as_posix(),
                 ml_instance.catalog.deriva_server.server,
                 ml_instance.catalog_id,
             ],
             capture_output=True,
             text=True,
         )
+        print(result.stdout)
+        print(result.stderr)
         new_workflow = result.stdout.strip()
         assert new_workflow == workflow_rid
 
-    def test_workflow_creation_notebook(self, test_ml):
-        ml_instance = test_ml
+    def test_workflow_creation_notebook(self, notebook_test):
+        ml_instance = notebook_test
+
+        notebook_path = Path(__file__).parent / "workflow-test.ipynb"  # directory where this test lives
         ml_instance.add_term(vc.asset_type, "Test Model", description="Model for our Test workflow")
         ml_instance.add_term(vc.workflow_type, "Test Workflow", description="A ML Workflow that uses Deriva ML API")
         workflow_table = ml_instance.pathBuilder.schemas[ml_instance.ml_schema].Workflow
@@ -69,14 +77,17 @@ class TestWorkflow:
         assert 0 == len(workflows)
 
         print("Running notebook...")
+
         result = subprocess.run(
             [
                 "deriva-ml-run-notebook",
-                "execution/workflow-test.ipynb",
+                notebook_path.as_posix(),
                 "--host",
                 ml_instance.catalog.deriva_server.server,
                 "--catalog",
                 ml_instance.catalog_id,
+                "--kernel",
+                "test_kernel",
                 "--log-output",
             ],
             capture_output=True,

{deriva_ml-1.14.33 → deriva_ml-1.14.34}/tests/execution/workflow-test.ipynb

@@ -18,9 +18,7 @@
    "outputs": [],
    "source": [
     "from deriva_ml import DerivaML, MLVocab as vc\n",
-    "import os\n",
-    "import logging\n",
-    "logger = logging.getLogger()"
+    "import os"
    ]
   },
   {
@@ -42,8 +40,8 @@
    },
    "outputs": [],
    "source": [
-    "hostname = None\n",
-    "catalog_id = None"
+    "host = None\n",
+    "catalog = None"
    ]
   },
   {
@@ -54,24 +52,21 @@
    "outputs": [],
    "source": [
     "# Modify these to your desired server and catalog.\n",
-    "hostname = hostname or os.environ.get(\"DERIVA_HOST\")\n",
-    "catalog_id = catalog_id or os.environ.get(\"DERIVA_CATALOG_ID\")\n",
+    "host = host or os.environ.get(\"DERIVA_HOST\")\n",
+    "catalog = catalog or os.environ.get(\"DERIVA_CATALOG\")\n",
     "\n",
     "# Change this line to call the domain specific class derived from DerivaML\n",
-    "ml_instance = DerivaML(hostname, catalog_id)\n",
-    "logger.info(\"Got ML instance:\")\n",
+    "ml_instance = DerivaML(host, catalog)\n",
     "\n",
     "ml_instance.add_term(vc.asset_type, \"Test Model\", description=\"Model for our Test workflow\")\n",
     "ml_instance.add_term(vc.workflow_type, \"Test Workflow\", description=\"A ML Workflow that uses Deriva ML API\")\n",
-    "print(\"Added terms to ML instance\")\n",
+    "\n",
     "api_workflow = ml_instance.create_workflow(\n",
-    "    name=\"Test Workflow One\",\n",
-    "    workflow_type=\"Test Workflow\",\n",
-    "    description=\"A test operation\",\n",
-    ")\n",
-    "logger.info(f\"URL: {api_workflow.url}\")\n",
-    "rid = ml_instance.add_workflow(api_workflow)\n",
-    "logger.info(f\"RID {rid}\")\n"
+    "     name=\"Test Workflow One\",\n",
+    "     workflow_type=\"Test Workflow\",\n",
+    "     description=\"A test operation\",\n",
+    " )\n",
+    "rid = ml_instance.add_workflow(api_workflow)"
    ]
   }
  ],

{deriva_ml-1.14.33 → deriva_ml-1.14.34}/tests/execution/workflow-test.py

@@ -6,7 +6,6 @@ from deriva_ml import MLVocab as vc
 hostname = sys.argv[1]
 catalog_id = sys.argv[2]
 
-
 ml_instance = DerivaML(hostname, catalog_id)
 
 ml_instance.add_term(vc.asset_type, "Test Model", description="Model for our Test workflow")

{deriva_ml-1.14.33 → deriva_ml-1.14.34}/tests/feature/test_features.py

@@ -2,9 +2,8 @@
 Tests for feature functionality.
 """
 
-from unittest.mock import Mock
-
 import pytest
+from pydantic import ValidationError
 
 from deriva_ml import (
     BuiltinTypes,
@@ -21,14 +20,14 @@ from deriva_ml.feature import FeatureRecord
 class TestFeatureRecord:
     """Test cases for the FeatureRecord base class."""
 
-    def test_feature_record_creation(self):
+    def test_feature_record_creation(self, mocker):
         """Test basic FeatureRecord creation."""
         # Create a mock feature
-        mock_feature = Mock()
-        mock_feature.feature_columns = {Mock(name="value"), Mock(name="confidence")}
-        mock_feature.asset_columns = {Mock(name="image_file")}
-        mock_feature.term_columns = {Mock(name="category")}
-        mock_feature.value_columns = {Mock(name="score")}
+        mock_feature = mocker.Mock()
+        mock_feature.feature_columns = {mocker.Mock(name="value"), mocker.Mock(name="confidence")}
+        mock_feature.asset_columns = {mocker.Mock(name="image_file")}
+        mock_feature.term_columns = {mocker.Mock(name="category")}
+        mock_feature.value_columns = {mocker.Mock(name="score")}
 
         # Create a test class that inherits from FeatureRecord
         class TestFeature(FeatureRecord):
@@ -58,17 +57,17 @@ class TestFeatureRecord:
         assert record.category == "good"
         assert record.score == 0.8
 
-    def test_feature_record_column_methods(self):
+    def test_feature_record_column_methods(self, mocker):
         """Test the column access methods of FeatureRecord."""
         # Create mock columns
-        value_col = Mock(name="value")
-        confidence_col = Mock(name="confidence")
-        asset_col = Mock(name="image_file")
-        term_col = Mock(name="category")
-        value_only_col = Mock(name="score")
+        value_col = mocker.Mock(name="value")
+        confidence_col = mocker.Mock(name="confidence")
+        asset_col = mocker.Mock(name="image_file")
+        term_col = mocker.Mock(name="category")
+        value_only_col = mocker.Mock(name="score")
 
         # Create a mock feature
-        mock_feature = Mock()
+        mock_feature = mocker.Mock()
         mock_feature.feature_columns = {value_col, confidence_col, asset_col, term_col, value_only_col}
         mock_feature.asset_columns = {asset_col}
         mock_feature.term_columns = {term_col}
@@ -141,6 +140,22 @@ class TestFeatures:
         with pytest.raises(DerivaMLException):
             ml_instance.lookup_feature("Subject", "SubjectHealth1")
 
+    def test_feature_record(self, dataset_test, tmp_path):
+        ml_instance = DerivaML(
+            dataset_test.catalog.hostname, dataset_test.catalog.catalog_id, working_dir=tmp_path, use_minid=False
+        )
+        SubjectHealthFeature = ml_instance.feature_record_class("Subject", "Health")
+        print(SubjectHealthFeature.model_fields.keys())
+
+        print(SubjectHealthFeature.feature_columns())
+
+        with pytest.raises(ValidationError):
+            SubjectHealthFeature(Subject="SubjectRID", Health="Good", Scale=23, Foo="Bar")
+        print(SubjectHealthFeature.value_columns())
+        print(SubjectHealthFeature.term_columns())
+        print(SubjectHealthFeature.asset_columns())
+        print(SubjectHealthFeature.feature_columns())
+
     def test_add_feature(self, dataset_test, tmp_path):
         ml_instance = DerivaML(
             dataset_test.catalog.hostname, dataset_test.catalog.catalog_id, working_dir=tmp_path, use_minid=False
@@ -166,7 +181,8 @@ class TestFeatures:
 
         with feature_execution.execute() as exe:
             SubjectHealthFeature = ml_instance.feature_record_class("Subject", "Health")
-            exe.add_features([SubjectHealthFeature(Subject=subject_rids[0], Health="Good", Scale=23)])
+            print(SubjectHealthFeature.feature_columns())
+            exe.add_features([SubjectHealthFeature(Subject=subject_rids[0], SubjectHealth="Sick", Scale=23)])
 
         feature_execution.upload_execution_outputs()
         features = list(ml_instance.list_feature_values("Subject", "Health"))

{deriva_ml-1.14.33 → deriva_ml-1.14.34}/tests/test_utils.py

@@ -1,8 +1,12 @@
+import os
+import shutil
 from tempfile import TemporaryDirectory
 from urllib.parse import quote as urlquote
 
 from demo_catalog import create_demo_features
 from deriva.core.datapath import DataPathException
+from ipykernel.kernelspec import install
+from jupyter_client.kernelspec import KernelSpecManager
 
 from deriva_ml import DerivaML
 from deriva_ml.demo_catalog import (
@@ -42,7 +46,14 @@ class MLCatalog:
         pb = self.catalog.getPathBuilder()
         ml_path = pb.schemas["deriva-ml"]
         domain_path = pb.schemas[self.domain_schema]
-        for t in ["Dataset_Execution", "Dataset_Version", "Dataset_Dataset", "Workflow", "Workflow_Execution"]:
+        for t in [
+            "Dataset_Execution",
+            "Dataset_Version",
+            "Dataset_Dataset",
+            "Execution",
+            "Workflow_Execution",
+            "Workflow",
+        ]:
             try:
                 ml_path.tables[t].path.delete()
             except DataPathException:
@@ -110,3 +121,60 @@ class MLDatasetCatalog:
         with TemporaryDirectory() as tmp_dir:
             ml_instance = DerivaML(self.catalog.hostname, self.catalog.catalog_id, working_dir=tmp_dir, use_minid=False)
             self.dataset_description: DatasetDescription = create_demo_datasets(ml_instance)
+
+
+def create_jupyter_kernel(name: str, kernel_dir, display_name: str = None, user: bool = True) -> None:
+    """
+    Create and install a Jupyter kernel spec using ipykernel.
+
+    Parameters
+    ----------
+    name : str
+        The internal name of the kernel (used in `--kernel`).
+    display_name : str, optional
+        The label shown in Jupyter’s kernel chooser (defaults to name).
+    user : bool, default=True
+        If True, install for the current user only.
+        If False, requires admin rights (system-wide).
+    """
+    if display_name is None:
+        display_name = name
+
+    os.environ["JUPYTER_PATH"] = f"{kernel_dir}/share/jupyter"
+
+    print(f"Installing Jupyter kernel '{name}' with display name '{display_name}'")
+    install(
+        kernel_name=name,
+        display_name=display_name,
+        prefix=kernel_dir,  # ensures it uses the current environment
+    )
+    print("✅ Kernel installed successfully.")
+
+
+def destroy_jupyter_kernel(name: str, user: bool = True) -> None:
+    """
+    Remove a Jupyter kernel spec by name.
+
+    Parameters
+    ----------
+    name : str
+        The internal kernel name (the same name used in create_jupyter_kernel).
+    user : bool, default=True
+        If True, remove from the user-level kernels directory.
+        If False, attempt system-wide removal (requires permissions).
+    """
+    ksm = KernelSpecManager()
+    kernels = ksm.find_kernel_specs()
+
+    if name not in kernels:
+        print(f"❌ Kernel '{name}' not found.")
+        return
+
+    kernel_path = kernels[name]
+    print(f"Removing kernel '{name}' at {kernel_path}")
+
+    try:
+        shutil.rmtree(kernel_path)
+        print(f"✅ Kernel '{name}' removed successfully.")
+    except Exception as e:
+        print(f"⚠️ Failed to remove kernel '{name}': {e}")