climate-ref-pmp 0.5.0__tar.gz → 0.5.2__tar.gz

{climate_ref_pmp-0.5.0 → climate_ref_pmp-0.5.2}/.gitignore

@@ -74,7 +74,6 @@ coverage.xml
 *.pot
 
 # Django stuff:
-*.log
 local_settings.py
 db.sqlite3
 db.sqlite3-journal

{climate_ref_pmp-0.5.0 → climate_ref_pmp-0.5.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: climate-ref-pmp
-Version: 0.5.0
+Version: 0.5.2
 Summary: PMP diagnostic provider for the Rapid Evaluation Framework
 Author-email: Jiwoo Lee <jwlee@llnl.gov>
 License: Apache-2.0
@@ -18,7 +18,6 @@ Classifier: Programming Language :: Python :: 3.13
 Classifier: Topic :: Scientific/Engineering
 Requires-Python: >=3.11
 Requires-Dist: climate-ref-core
-Requires-Dist: pcmdi-metrics
 Description-Content-Type: text/markdown
 
 # climate-ref-pmp

{climate_ref_pmp-0.5.0 → climate_ref_pmp-0.5.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "climate-ref-pmp"
-version = "0.5.0"
+version = "0.5.2"
 description = "PMP diagnostic provider for the Rapid Evaluation Framework"
 readme = "README.md"
 authors = [
@@ -21,12 +21,8 @@ classifiers = [
 ]
 dependencies = [
     "climate-ref-core",
-    "pcmdi_metrics",
 ]
 
-[tool.uv.sources]
-pcmdi_metrics = { git = "https://github.com/PCMDI/pcmdi_metrics", rev="v3.9.1" }
-
 [project.license]
 text = "Apache-2.0"
 

{climate_ref_pmp-0.5.0 → climate_ref_pmp-0.5.2}/src/climate_ref_pmp/diagnostics/annual_cycle.py

@@ -12,6 +12,7 @@ from climate_ref_core.diagnostics import (
     ExecutionDefinition,
     ExecutionResult,
 )
+from climate_ref_core.pycmec.metric import remove_dimensions
 from climate_ref_pmp.pmp_driver import build_glob_pattern, build_pmp_command, process_json_result
 
 
@@ -22,7 +23,16 @@ class AnnualCycle(CommandLineDiagnostic):
 
     name = "Annual Cycle"
     slug = "annual-cycle"
-    facets = ("model", "realization", "reference", "mode", "season", "method", "statistic")
+    facets = (
+        "source_id",
+        "member_id",
+        "experiment_id",
+        "variable_id",
+        "reference_source_id",
+        "region",
+        "statistic",
+        "season",
+    )
     data_requirements = (
         # Surface temperature
         (
@@ -101,16 +111,17 @@ class AnnualCycle(CommandLineDiagnostic):
             Command arguments to execute in the PMP environment
         """
         input_datasets = definition.datasets[SourceDatasetType.CMIP6]
+        reference_datasets = definition.datasets[SourceDatasetType.PMPClimatology]
+        selector = input_datasets.selector_dict()
+        reference_selector = reference_datasets.selector_dict()
+        logger.debug(f"selector: {selector}")
+        logger.debug(f"reference selector: {reference_selector}")
+
         source_id = input_datasets["source_id"].unique()[0]
         experiment_id = input_datasets["experiment_id"].unique()[0]
         member_id = input_datasets["member_id"].unique()[0]
         variable_id = input_datasets["variable_id"].unique()[0]
 
-        logger.debug(f"input_datasets['source_id'].unique(): {input_datasets['source_id'].unique()}")
-        logger.debug(f"input_datasets['experiment_id'].unique(): {input_datasets['experiment_id'].unique()}")
-        logger.debug(f"input_datasets['member_id'].unique(): {input_datasets['member_id'].unique()}")
-        logger.debug(f"input_datasets['variable_id'].unique(): {input_datasets['variable_id'].unique()}")
-
         model_files_raw = input_datasets.path.to_list()
         if len(model_files_raw) == 1:
             model_files = model_files_raw[0]  # If only one file, use it directly
@@ -123,23 +134,11 @@ class AnnualCycle(CommandLineDiagnostic):
 
         logger.debug(f"input_datasets: {input_datasets}")
         logger.debug(f"input_datasets.keys(): {input_datasets.keys()}")
-        logger.debug(f"input_datasets['variable_id']: {input_datasets['variable_id']}")
-
-        logger.debug(f"source_id: {source_id}")
-        logger.debug(f"experiment_id: {experiment_id}")
-        logger.debug(f"member_id: {member_id}")
-        logger.debug(f"variable_id: {variable_id}")
 
-        reference_dataset = definition.datasets[SourceDatasetType.PMPClimatology]
-        reference_dataset_name = reference_dataset["source_id"].unique()[0]
-        reference_dataset_path = reference_dataset.datasets.iloc[0]["path"]
-
-        logger.debug(f"reference_dataset.datasets: {reference_dataset.datasets}")
-        logger.debug(f"reference_dataset['source_id']: {reference_dataset['source_id']}")
-        logger.debug(
-            f"reference_dataset.datasets.iloc[0]['path']: {reference_dataset.datasets.iloc[0]['path']}"
-        )
+        reference_dataset_name = reference_datasets["source_id"].unique()[0]
+        reference_dataset_path = reference_datasets.datasets.iloc[0]["path"]
 
+        logger.debug(f"reference_dataset.datasets: {reference_datasets.datasets}")
         logger.debug(f"reference_dataset_name: {reference_dataset_name}")
         logger.debug(f"reference_dataset_path: {reference_dataset_path}")
 
@@ -154,14 +153,18 @@ class AnnualCycle(CommandLineDiagnostic):
         data_name = f"{source_id}_{experiment_id}_{member_id}"
         data_path = model_files
         params = {
-            "driver_file": "mean_climate/pcmdi_compute_climatologies.py",
-            "parameter_file": self.parameter_file_1,
             "vars": variable_id,
             "infile": data_path,
             "outfile": f"{output_directory_path}/{variable_id}_{data_name}_clims.nc",
         }
 
-        cmds.append(build_pmp_command(**params))
+        cmds.append(
+            build_pmp_command(
+                driver_file="pcmdi_compute_climatologies.py",
+                parameter_file=self.parameter_file_1,
+                **params,
+            )
+        )
 
         # ----------------------------------------------
         # PART 2: Build the command to calculate diagnostics
@@ -184,8 +187,6 @@ class AnnualCycle(CommandLineDiagnostic):
         date = datetime.datetime.now().strftime("%Y%m%d")
 
         params = {
-            "driver_file": "mean_climate/mean_climate_driver.py",
-            "parameter_file": self.parameter_file_2,
             "vars": variable_id,
             "custom_observations": f"{output_directory_path}/obs_dict.json",
             "test_data_path": output_directory_path,
@@ -196,7 +197,13 @@ class AnnualCycle(CommandLineDiagnostic):
             "cmec": "",
         }
 
-        cmds.append(build_pmp_command(**params))
+        cmds.append(
+            build_pmp_command(
+                driver_file="mean_climate_driver.py",
+                parameter_file=self.parameter_file_2,
+                **params,
+            )
+        )
 
         return cmds
 
@@ -227,6 +234,7 @@ class AnnualCycle(CommandLineDiagnostic):
         # Find the executions file
         results_files = list(results_directory.glob("*_cmec.json"))
         if len(results_files) != 1:  # pragma: no cover
+            logger.error(f"More than one or no cmec file found: {results_files}")
             return ExecutionResult.build_from_failure(definition)
         else:
             results_file = results_files[0]
@@ -252,12 +260,27 @@ class AnnualCycle(CommandLineDiagnostic):
         png_files = list(png_directory.glob("*.png"))
         data_files = list(data_directory.glob("*.nc"))
 
-        cmec_output, cmec_metric = process_json_result(results_file_transformed, png_files, data_files)
+        cmec_output_bundle, cmec_metric_bundle = process_json_result(
+            results_file_transformed, png_files, data_files
+        )
+
+        # Add missing dimensions to the output
+        input_selectors = input_datasets.selector_dict()
+        reference_selectors = definition.datasets[SourceDatasetType.PMPClimatology].selector_dict()
+        cmec_metric_bundle = cmec_metric_bundle.prepend_dimensions(
+            {
+                "source_id": input_selectors["source_id"],
+                "member_id": input_selectors["member_id"],
+                "experiment_id": input_selectors["experiment_id"],
+                "variable_id": input_selectors["variable_id"],
+                "reference_source_id": reference_selectors["source_id"],
+            }
+        )
 
         return ExecutionResult.build_from_output_bundle(
             definition,
-            cmec_output_bundle=cmec_output,
-            cmec_metric_bundle=cmec_metric,
+            cmec_output_bundle=cmec_output_bundle,
+            cmec_metric_bundle=cmec_metric_bundle,
         )
 
     def run(self, definition: ExecutionDefinition) -> ExecutionResult:
@@ -274,7 +297,6 @@ class AnnualCycle(CommandLineDiagnostic):
         :
             The result of running the diagnostic.
         """
-        logger.debug("PMP annual cycle run start")
         cmds = self.build_cmds(definition)
 
         runs = [self.provider.run(cmd) for cmd in cmds]
@@ -297,41 +319,18 @@ def _transform_results(data: dict[str, Any]) -> dict[str, Any]:
     dict
         The transformed executions dictionary.
     """
+    # Remove the model, reference, rip dimensions
+    # These are later replaced with a REF-specific naming convention
+    data = remove_dimensions(data, ["model", "reference", "rip"])
+
+    # TODO: replace this with the ability to capture series
     # Remove the "CalendarMonths" key from the nested structure
-    if "RESULTS" in data:
-        models = list(data["RESULTS"].keys())
-        for model in models:
-            if "default" in data["RESULTS"][model]:
-                realizations = list(data["RESULTS"][model]["default"].keys())
-                if "attributes" in realizations:
-                    realizations.remove("attributes")
-                for realization in realizations:
-                    regions = list(data["RESULTS"][model]["default"][realization].keys())
-                    for region in regions:
-                        stats = list(data["RESULTS"][model]["default"][realization][region].keys())
-                        for stat in stats:
-                            if (
-                                "CalendarMonths"
-                                in data["RESULTS"][model]["default"][realization][region][stat]
-                            ):
-                                calendar_months = data["RESULTS"][model]["default"][realization][region][
-                                    stat
-                                ].pop("CalendarMonths")
-                                for i, value in enumerate(calendar_months):
-                                    key_name = f"CalendarMonth-{i + 1:02d}"
-                                    data["RESULTS"][model]["default"][realization][region][stat][key_name] = (
-                                        value
-                                    )
+    for region, region_values in data["RESULTS"].items():
+        for stat, stat_values in region_values.items():
+            if "CalendarMonths" in stat_values:
+                stat_values.pop("CalendarMonths")
 
     # Remove the "CalendarMonths" key from the nested structure in "DIMENSIONS"
-    if (
-        "DIMENSIONS" in data
-        and "season" in data["DIMENSIONS"]
-        and "CalendarMonths" in data["DIMENSIONS"]["season"]
-    ):
-        calendar_months = data["DIMENSIONS"]["season"].pop("CalendarMonths")
-        for i in range(1, 13):
-            key_name = f"CalendarMonth-{i:02d}"
-            data["DIMENSIONS"]["season"][key_name] = {}
+    data["DIMENSIONS"]["season"].pop("CalendarMonths")
 
     return data

{climate_ref_pmp-0.5.0 → climate_ref_pmp-0.5.2}/src/climate_ref_pmp/diagnostics/variability_modes.py

@@ -20,7 +20,16 @@ class ExtratropicalModesOfVariability(CommandLineDiagnostic):
     ts_modes = ("PDO", "NPGO", "AMO")
     psl_modes = ("NAO", "NAM", "PNA", "NPO", "SAM")
 
-    facets = ("model", "realization", "reference", "mode", "season", "method", "statistic")
+    facets = (
+        "source_id",
+        "member_id",
+        "experiment_id",
+        "reference_source_id",
+        "mode",
+        "season",
+        "method",
+        "statistic",
+    )
 
     def __init__(self, mode_id: str):
         self.mode_id = mode_id.upper()
@@ -32,7 +41,6 @@ class ExtratropicalModesOfVariability(CommandLineDiagnostic):
             obs_variable: str,
             cmip_variable: str,
             extra_experiments: str | tuple[str, ...] | list[str] = (),
-            remove_experiments: str | tuple[str, ...] | list[str] = (),
         ) -> tuple[DataRequirement, DataRequirement]:
             filters = [
                 FacetFilter(
@@ -55,6 +63,7 @@ class ExtratropicalModesOfVariability(CommandLineDiagnostic):
                 DataRequirement(
                     source_type=SourceDatasetType.CMIP6,
                     filters=tuple(filters),
+                    # TODO: remove unneeded variant_label
                     group_by=("source_id", "experiment_id", "variant_label", "member_id"),
                 ),
             )
@@ -95,7 +104,6 @@ class ExtratropicalModesOfVariability(CommandLineDiagnostic):
 
         reference_dataset = definition.datasets[SourceDatasetType.obs4MIPs]
         reference_dataset_name = reference_dataset["source_id"].unique()[0]
-        # reference_dataset_path = reference_dataset.datasets[0]["path"]
         reference_dataset_path = reference_dataset.datasets.iloc[0]["path"]
 
         logger.debug(f"reference_dataset: {reference_dataset}")
@@ -119,9 +127,7 @@ class ExtratropicalModesOfVariability(CommandLineDiagnostic):
             reference_data_path = reference_dataset_path
 
         # Build the command to run the PMP driver script
-        params = {
-            "driver_file": "variability_mode/variability_modes_driver.py",
-            "parameter_file": self.parameter_file,
+        params: dict[str, str | int | None] = {
             "variability_mode": self.mode_id,
             "modpath": modpath,
             "modpath_lf": "none",
@@ -141,7 +147,11 @@ class ExtratropicalModesOfVariability(CommandLineDiagnostic):
             params["oeyear"] = 2005
 
         # Pass the parameters using **kwargs
-        return build_pmp_command(**params)
+        return build_pmp_command(
+            driver_file="variability_modes_driver.py",
+            parameter_file=self.parameter_file,
+            **params,
+        )
 
     def build_execution_result(self, definition: ExecutionDefinition) -> ExecutionResult:
         """
@@ -165,10 +175,28 @@ class ExtratropicalModesOfVariability(CommandLineDiagnostic):
         png_files = [definition.as_relative_path(f) for f in definition.output_directory.glob("*.png")]
         data_files = [definition.as_relative_path(f) for f in definition.output_directory.glob("*.nc")]
 
-        cmec_output, cmec_metric = process_json_result(results_files[0], png_files, data_files)
+        cmec_output_bundle, cmec_metric_bundle = process_json_result(results_files[0], png_files, data_files)
+
+        # Add additional metadata to the metrics
+        input_selectors = definition.datasets[SourceDatasetType.CMIP6].selector_dict()
+        reference_selectors = definition.datasets[SourceDatasetType.obs4MIPs].selector_dict()
+        cmec_metric_bundle = cmec_metric_bundle.remove_dimensions(
+            [
+                "model",
+                "realization",
+                "reference",
+            ],
+        ).prepend_dimensions(
+            {
+                "source_id": input_selectors["source_id"],
+                "member_id": input_selectors["member_id"],
+                "experiment_id": input_selectors["experiment_id"],
+                "reference_source_id": reference_selectors["source_id"],
+            }
+        )
 
         return ExecutionResult.build_from_output_bundle(
             definition,
-            cmec_output_bundle=cmec_output,
-            cmec_metric_bundle=cmec_metric,
+            cmec_output_bundle=cmec_output_bundle,
+            cmec_metric_bundle=cmec_metric_bundle,
         )

{climate_ref_pmp-0.5.0 → climate_ref_pmp-0.5.2}/src/climate_ref_pmp/pmp_driver.py

@@ -83,19 +83,13 @@ def process_json_result(
         dimensions.update(dimensions["dimensions"])
         del dimensions["dimensions"]
 
-    if "statistic" in dimensions["json_structure"]:  # pragma: no branch
-        dimensions["json_structure"].remove("statistic")
-        dimensions.pop("statistic")
-
-    # Remove the "attributes" key from the RESULTS
-    # This isn't standard CMEC output, but it is what PMP produces
     results = json_result["RESULTS"]
 
     cmec_metric["RESULTS"] = results
     cmec_metric["DIMENSIONS"] = dimensions
 
     if "provenance" in json_result:  # pragma: no branch
-        cmec_metric["provenance"] = json_result["provenance"]
+        cmec_metric["PROVENANCE"] = json_result["provenance"]
 
     logger.info(f"cmec_output: {pretty_repr(cmec_output)}")
     logger.info(f"cmec_metric: {pretty_repr(cmec_metric)}")
@@ -140,7 +134,7 @@ def _get_resource(package: str, resource_name: str | pathlib.Path, use_resources
 def build_pmp_command(
     driver_file: str,
     parameter_file: str,
-    **kwargs: dict[str, str | int | float | list[str]],
+    **kwargs: str | int | float | list[str] | None,
 ) -> list[str]:
     """
     Run a PMP driver script via a conda environment
@@ -150,6 +144,9 @@ def build_pmp_command(
     The output consists of a JSON file that contains the executions of the PMP diagnostics,
     and a set of PNG and data files that are produced by the diagnostics.
 
+    The PMP driver scripts are installed in the PMP conda environment,
+    but absolute paths should be used for non-PMP scripts.
+
     Parameters
     ----------
     driver_file
@@ -160,13 +157,18 @@ def build_pmp_command(
         Additional arguments to pass to the driver script
     """
     # Note this uses the driver script from the REF env *not* the PMP conda env
-    _driver_script = _get_resource("pcmdi_metrics", driver_file, use_resources=False)
     _parameter_file = _get_resource("climate_ref_pmp.params", parameter_file, use_resources=True)
 
+    # This is a workaround for a fatal error in internal_Finalize of MPICH
+    # when running in a conda environment on MacOS.
+    # It is not clear if this is a bug in MPICH or a problem with the conda environment.
+    if "FI_PROVIDER" not in os.environ:  # pragma: no branch
+        logger.debug("Setting env variable 'FI_PROVIDER=tcp'")
+        os.environ["FI_PROVIDER"] = "tcp"
+
     # Run the driver script inside the PMP conda environment
     cmd = [
-        "python",
-        _driver_script,
+        driver_file,
         "-p",
         _parameter_file,
     ]
@@ -179,9 +181,7 @@ def build_pmp_command(
             else:
                 cmd.extend([f"--{key}"])
 
-    logger.info("-- PMP command to run --")
-    logger.info("[PMP] Command to run:", " ".join(map(str, cmd)))
-    logger.info("[PMP] Command generation for the driver completed.")
+    logger.info(f"PMP Command: {cmd}")
 
     return cmd
 

climate_ref_pmp-0.5.2/tests/integration/test_diagnostics.py

@@ -0,0 +1,12 @@
+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):
+    diagnostic_validation(diagnostic)

{climate_ref_pmp-0.5.0 → climate_ref_pmp-0.5.2}/tests/unit/conftest.py

@@ -5,8 +5,13 @@ import pytest
 
 
 @pytest.fixture(scope="module")
-def pdo_example_dir() -> Path:
-    return Path(__file__).parent / "test-data" / "pdo-example"
+def pdo_example_dir(regression_data_dir) -> Path:
+    return (
+        regression_data_dir
+        / "pmp"
+        / "extratropical-modes-of-variability-pdo"
+        / "cmip6_hist-GHG_r1i1p1f1_ACCESS-ESM1-5_r1i1p1f1__obs4mips_HadISST-1-1_ts"
+    )
 
 
 @pytest.fixture