climate-ref-pmp 0.8.0__tar.gz → 0.9.0__tar.gz

{climate_ref_pmp-0.8.0 → climate_ref_pmp-0.9.0}/.gitignore

@@ -150,7 +150,7 @@ dmypy.json
 
 # Generated output
 out
-.ref
+.ref*
 
 # Ignore copied LICENCE/NOTICE files
 packages/*/LICENCE
@@ -158,3 +158,12 @@ packages/*/NOTICE
 
 # Local directory for data
 /data
+
+# Generated SDK
+/climate_ref_client
+
+# User-specific catalog paths (test data)
+*.paths.yaml
+
+# Helm dependencies
+helm/charts/*

{climate_ref_pmp-0.8.0 → climate_ref_pmp-0.9.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: climate-ref-pmp
-Version: 0.8.0
+Version: 0.9.0
 Summary: PMP diagnostic provider for the Rapid Evaluation Framework
 Author-email: Jiwoo Lee <jwlee@llnl.gov>, Jared Lewis <jared.lewis@climate-resource.com>
 License-Expression: Apache-2.0

{climate_ref_pmp-0.8.0 → climate_ref_pmp-0.9.0}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "climate-ref-pmp"
-version = "0.8.0"
+version = "0.9.0"
 description = "PMP diagnostic provider for the Rapid Evaluation Framework"
 readme = "README.md"
 authors = [

{climate_ref_pmp-0.8.0 → climate_ref_pmp-0.9.0}/src/climate_ref_pmp/__init__.py

@@ -6,11 +6,18 @@ from __future__ import annotations
 
 import importlib.metadata
 import os
+from pathlib import Path
 from typing import TYPE_CHECKING
 
+import pooch
 from loguru import logger
 
-from climate_ref_core.dataset_registry import DATASET_URL, dataset_registry_manager
+from climate_ref_core.dataset_registry import (
+    DATASET_URL,
+    dataset_registry_manager,
+    fetch_all_files,
+    validate_registry_cache,
+)
 from climate_ref_core.providers import CondaDiagnosticProvider
 from climate_ref_pmp.diagnostics import ENSO, AnnualCycle, ExtratropicalModesOfVariability
 
@@ -19,6 +26,8 @@ if TYPE_CHECKING:
 
 __version__ = importlib.metadata.version("climate-ref-pmp")
 
+_REGISTRY_NAME = "pmp-climatology"
+
 
 # Create the PMP diagnostics provider
 # PMP uses a conda environment to run the diagnostics
@@ -38,6 +47,34 @@ class PMPDiagnosticProvider(CondaDiagnosticProvider):
             logger.debug("Setting env variable 'FI_PROVIDER=tcp'")
             self.env_vars["FI_PROVIDER"] = "tcp"
 
+    def fetch_data(self, config: Config) -> None:
+        """Fetch PMP climatology data."""
+        registry = dataset_registry_manager[_REGISTRY_NAME]
+        fetch_all_files(registry, _REGISTRY_NAME, output_dir=None)
+
+    def validate_setup(self, config: Config) -> bool:
+        """Validate conda environment and data checksums."""
+        # First check conda environment
+        if not super().validate_setup(config):
+            return False
+
+        # Then check data checksums
+        registry = dataset_registry_manager[_REGISTRY_NAME]
+        errors = validate_registry_cache(registry, _REGISTRY_NAME)
+        if errors:
+            for error in errors:
+                logger.error(f"{self.slug} validation failed: {error}")
+            logger.error(
+                f"Data for {self.slug} is missing or corrupted. "
+                f"Please run `ref providers setup --provider {self.slug}` to fetch data."
+            )
+            return False
+        return True
+
+    def get_data_path(self) -> Path | None:
+        """Get the path where PMP data is cached."""
+        return Path(pooch.os_cache("climate_ref"))
+
 
 provider = PMPDiagnosticProvider("PMP", __version__)
 

climate_ref_pmp-0.9.0/tests/integration/test_diagnostics.py

@@ -0,0 +1,137 @@
+from pathlib import Path
+
+import pytest
+from climate_ref_pmp import provider
+
+from climate_ref.testing import TestCaseRunner, validate_result
+from climate_ref_core.diagnostics import Diagnostic
+from climate_ref_core.testing import (
+    RegressionValidator,
+    TestCasePaths,
+    collect_test_case_params,
+    load_datasets_from_yaml,
+)
+
+
+@pytest.fixture(scope="session")
+def provider_test_data_dir() -> Path:
+    """Path to the package-local test data directory."""
+    return Path(__file__).parent.parent / "test-data"
+
+
+enso_skip_diagnostics = [
+    "enso-teleconnections",
+    "enso-clivar-teleconnections",
+]
+
+diagnostics = [
+    pytest.param(
+        diagnostic,
+        id=diagnostic.slug,
+        marks=[
+            *(
+                [pytest.mark.skip(reason="ENSO diagnostics do not support the sample data")]
+                if diagnostic.slug in enso_skip_diagnostics
+                else []
+            ),
+        ],
+    )
+    for diagnostic in provider.diagnostics()
+]
+
+# Test case params for parameterized test_case tests
+test_case_params = collect_test_case_params(provider)
+
+
+@pytest.mark.slow
+@pytest.mark.parametrize("diagnostic", diagnostics)
+def test_diagnostics(diagnostic: Diagnostic, diagnostic_validation):
+    if "enso" in diagnostic.slug:
+        pytest.skip("ENSO diagnostics do not support the sample data")
+
+    validator = diagnostic_validation(diagnostic)
+
+    definition = validator.get_definition()
+    validator.execute(definition)
+
+
+@pytest.mark.parametrize("diagnostic", diagnostics)
+def test_build_results(diagnostic: Diagnostic, diagnostic_validation):
+    validator = diagnostic_validation(diagnostic)
+
+    definition = validator.get_regression_definition()
+    validator.validate(definition)
+    validator.execution_regression.check(definition.key, definition.output_directory)
+
+
+@pytest.mark.parametrize("diagnostic,test_case_name", test_case_params)
+def test_validate_test_case_regression(
+    diagnostic: Diagnostic,
+    test_case_name: str,
+    provider_test_data_dir: Path,
+    config,
+    tmp_path: Path,
+):
+    """
+    Validate pre-stored test case regression outputs as CMEC bundles.
+
+    Each diagnostic/test_case is a separate parameterized test.
+    """
+    diagnostic.provider.configure(config)
+
+    paths = TestCasePaths.from_test_data_dir(
+        provider_test_data_dir,
+        diagnostic.slug,
+        test_case_name,
+    )
+
+    if not paths.catalog.exists():
+        pytest.skip(f"No catalog file for {diagnostic.slug}/{test_case_name}")
+    if not paths.regression.exists():
+        pytest.skip(f"No regression data for {diagnostic.slug}/{test_case_name}")
+
+    validator = RegressionValidator(
+        diagnostic=diagnostic,
+        test_case_name=test_case_name,
+        test_data_dir=provider_test_data_dir,
+    )
+
+    definition = validator.load_regression_definition(tmp_path / diagnostic.slug / test_case_name)
+    validator.validate(definition)
+
+
+@pytest.mark.slow
+@pytest.mark.test_cases
+@pytest.mark.parametrize("diagnostic,test_case_name", test_case_params)
+def test_run_test_cases(
+    diagnostic: Diagnostic,
+    test_case_name: str,
+    provider_test_data_dir: Path,
+    config,
+    tmp_path: Path,
+):
+    """
+    Run diagnostic test cases end-to-end with ESGF data.
+
+    Requires: `ref test-cases fetch --provider pmp` to have been run first.
+    """
+    diagnostic.provider.configure(config)
+
+    paths = TestCasePaths.from_test_data_dir(
+        provider_test_data_dir,
+        diagnostic.slug,
+        test_case_name,
+    )
+
+    if not paths.catalog.exists():
+        pytest.skip(f"No catalog file for {diagnostic.slug}/{test_case_name}")
+
+    datasets = load_datasets_from_yaml(paths.catalog)
+
+    runner = TestCaseRunner(config=config, datasets=datasets)
+    output_dir = tmp_path / diagnostic.slug / test_case_name
+
+    result = runner.run(diagnostic, test_case_name, output_dir)
+
+    assert result.successful, f"Diagnostic {diagnostic.slug} failed"
+    validate_result(diagnostic, config, result)

climate_ref_pmp-0.9.0/tests/unit/test_provider.py

@@ -0,0 +1,102 @@
+from pathlib import Path
+
+import pooch
+from climate_ref_pmp import PMPDiagnosticProvider, __version__, provider
+
+
+def test_provider():
+    assert provider.name == "PMP"
+    assert provider.slug == "pmp"
+    assert provider.version == __version__
+
+    assert len(provider) >= 1
+
+
+class TestPMPProviderHooks:
+    """Tests for PMPDiagnosticProvider lifecycle hooks."""
+
+    def test_get_data_path(self):
+        """Test that get_data_path returns the pooch cache path."""
+        data_path = provider.get_data_path()
+        assert data_path is not None
+        assert isinstance(data_path, Path)
+        assert data_path == Path(pooch.os_cache("climate_ref"))
+
+    def test_fetch_data(self, mocker):
+        """Test that fetch_data calls fetch_all_files."""
+        mock_fetch = mocker.patch("climate_ref_pmp.fetch_all_files")
+        mock_config = mocker.Mock()
+
+        provider.fetch_data(mock_config)
+
+        mock_fetch.assert_called_once()
+        # Check it's using the right registry name
+        call_args = mock_fetch.call_args
+        assert call_args[0][1] == "pmp-climatology"
+        assert call_args[1]["output_dir"] is None
+
+    def test_validate_setup_env_missing(self, mocker):
+        """Test validate_setup returns False when conda env is missing."""
+        mock_config = mocker.Mock()
+        # Mock the parent class to return False
+        mocker.patch.object(
+            PMPDiagnosticProvider.__bases__[0],
+            "validate_setup",
+            return_value=False,
+        )
+
+        result = provider.validate_setup(mock_config)
+        assert result is False
+
+    def test_validate_setup_data_invalid(self, mocker):
+        """Test validate_setup returns False when data validation fails."""
+        mock_config = mocker.Mock()
+        # Mock parent class to return True (conda env exists)
+        mocker.patch.object(
+            PMPDiagnosticProvider.__bases__[0],
+            "validate_setup",
+            return_value=True,
+        )
+        # Mock data validation to return errors
+        mocker.patch(
+            "climate_ref_pmp.validate_registry_cache",
+            return_value=["File missing: test.nc"],
+        )
+
+        result = provider.validate_setup(mock_config)
+        assert result is False
+
+    def test_validate_setup_all_valid(self, mocker):
+        """Test validate_setup returns True when all validation passes."""
+        mock_config = mocker.Mock()
+        # Mock parent class to return True (conda env exists)
+        mocker.patch.object(
+            PMPDiagnosticProvider.__bases__[0],
+            "validate_setup",
+            return_value=True,
+        )
+        # Mock data validation to return no errors
+        mocker.patch(
+            "climate_ref_pmp.validate_registry_cache",
+            return_value=[],
+        )
+
+        result = provider.validate_setup(mock_config)
+        assert result is True
+
+    def test_configure_sets_env_vars(self, mocker, tmp_path):
+        """Test that configure sets the required environment variables."""
+        test_provider = PMPDiagnosticProvider("PMP-Test", "1.0")
+        mock_config = mocker.Mock()
+        mock_config.paths.software = tmp_path / "software"
+        mock_config.ignore_datasets_file = tmp_path / "ignore.yaml"
+        mock_config.ignore_datasets_file.touch()
+
+        mocker.patch.object(test_provider, "get_conda_exe", return_value=Path("/path/to/conda"))
+
+        test_provider.configure(mock_config)
+
+        assert "PCMDI_CONDA_EXE" in test_provider.env_vars
+        assert test_provider.env_vars["PCMDI_CONDA_EXE"] == "/path/to/conda"
+        assert "FI_PROVIDER" in test_provider.env_vars
+        assert test_provider.env_vars["FI_PROVIDER"] == "tcp"

climate_ref_pmp-0.8.0/tests/integration/test_diagnostics.py

@@ -1,27 +0,0 @@
-import pytest
-from climate_ref_pmp import provider as pmp_provider
-
-from climate_ref_core.diagnostics import Diagnostic
-
-diagnostics = [pytest.param(diagnostic, id=diagnostic.slug) for diagnostic in pmp_provider.diagnostics()]
-
-
-@pytest.mark.slow
-@pytest.mark.parametrize("diagnostic", diagnostics)
-def test_diagnostics(diagnostic: Diagnostic, diagnostic_validation):
-    if "enso" in diagnostic.slug:
-        pytest.skip("ENSO diagnostics do not support the sample data")
-
-    validator = diagnostic_validation(diagnostic)
-
-    definition = validator.get_definition()
-    validator.execute(definition)
-
-
-@pytest.mark.parametrize("diagnostic", diagnostics)
-def test_build_results(diagnostic: Diagnostic, diagnostic_validation):
-    validator = diagnostic_validation(diagnostic)
-
-    definition = validator.get_regression_definition()
-    validator.validate(definition)
-    validator.execution_regression.check(definition.key, definition.output_directory)

climate_ref_pmp-0.8.0/tests/unit/test_provider.py

@@ -1,9 +0,0 @@
-from climate_ref_pmp import __version__, provider
-
-
-def test_provider():
-    assert provider.name == "PMP"
-    assert provider.slug == "pmp"
-    assert provider.version == __version__
-
-    assert len(provider) >= 1