Skip to content

climate_ref_pmp #

Rapid evaluating CMIP data

PMPDiagnosticProvider #

Bases: CondaDiagnosticProvider

Provider for PMP diagnostics.

Source code in packages/climate-ref-pmp/src/climate_ref_pmp/__init__.py 
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
class PMPDiagnosticProvider(CondaDiagnosticProvider):
    """
    Provider for PMP diagnostics.
    """

    def configure(self, config: Config) -> None:
        """Configure the provider."""
        super().configure(config)
        self.env_overrides["PCMDI_CONDA_EXE"] = str(self.conda_exe_path)
        # 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'")
            self.env_overrides["FI_PROVIDER"] = "tcp"

        # If any budgets are set, don't set default values
        if not any(name in os.environ for name in _THREAD_LIMIT_VARS):
            logger.debug(f"Limiting threads to {_DEFAULT_THREAD_LIMIT}")
            for name in _THREAD_LIMIT_VARS:
                self.env_overrides[name] = _DEFAULT_THREAD_LIMIT

    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 resolve_cache_dir(_REGISTRY_NAME)

    def ingest_data(self, config: Config, db: Any) -> None:
        """
        Ingest PMP climatology data into the database.

        This registers the climatology datasets so they can be used by diagnostics.

        Note: This method requires the climate-ref package to be installed.
        When using climate-ref-pmp standalone (without climate-ref), ingestion
        will be skipped with a warning message.
        """
        try:
            from climate_ref.datasets import ingest_datasets  # noqa: PLC0415
            from climate_ref.datasets.pmp_climatology import PMPClimatologyDatasetAdapter  # noqa: PLC0415
        except ImportError:
            logger.info(
                f"Skipping {self.slug} data ingestion: climate-ref package not installed. "
                "Run `ref datasets ingest --source-type pmp-climatology` manually if needed."
            )
            return

        data_path = self.get_data_path()
        if data_path is None or not data_path.exists():
            logger.warning(
                f"PMP data path does not exist. Run `ref providers setup --provider {self.slug}` first."
            )
            return

        # Find the pmp-climatology subdirectory
        climatology_path = data_path / "PMP_obs4MIPsClims"
        if not climatology_path.exists():
            logger.warning(f"PMP climatology data not found at {climatology_path}")
            return

        adapter = PMPClimatologyDatasetAdapter()

        try:
            stats = ingest_datasets(adapter, climatology_path, db, skip_invalid=True)
            stats.log_summary("PMP climatology ingestion complete:")
        except ValueError as e:
            logger.warning(f"No valid PMP climatology datasets found: {e}")

configure(config) #

Configure the provider.

Source code in packages/climate-ref-pmp/src/climate_ref_pmp/__init__.py 
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
def configure(self, config: Config) -> None:
    """Configure the provider."""
    super().configure(config)
    self.env_overrides["PCMDI_CONDA_EXE"] = str(self.conda_exe_path)
    # 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'")
        self.env_overrides["FI_PROVIDER"] = "tcp"

    # If any budgets are set, don't set default values
    if not any(name in os.environ for name in _THREAD_LIMIT_VARS):
        logger.debug(f"Limiting threads to {_DEFAULT_THREAD_LIMIT}")
        for name in _THREAD_LIMIT_VARS:
            self.env_overrides[name] = _DEFAULT_THREAD_LIMIT

fetch_data(config) #

Fetch PMP climatology data.

Source code in packages/climate-ref-pmp/src/climate_ref_pmp/__init__.py 
70
71
72
73
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)

get_data_path() #

Get the path where PMP data is cached.

Source code in packages/climate-ref-pmp/src/climate_ref_pmp/__init__.py 
94
95
96
def get_data_path(self) -> Path | None:
    """Get the path where PMP data is cached."""
    return resolve_cache_dir(_REGISTRY_NAME)

ingest_data(config, db) #

Ingest PMP climatology data into the database.

This registers the climatology datasets so they can be used by diagnostics.

Note: This method requires the climate-ref package to be installed. When using climate-ref-pmp standalone (without climate-ref), ingestion will be skipped with a warning message.

Source code in packages/climate-ref-pmp/src/climate_ref_pmp/__init__.py 
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
def ingest_data(self, config: Config, db: Any) -> None:
    """
    Ingest PMP climatology data into the database.

    This registers the climatology datasets so they can be used by diagnostics.

    Note: This method requires the climate-ref package to be installed.
    When using climate-ref-pmp standalone (without climate-ref), ingestion
    will be skipped with a warning message.
    """
    try:
        from climate_ref.datasets import ingest_datasets  # noqa: PLC0415
        from climate_ref.datasets.pmp_climatology import PMPClimatologyDatasetAdapter  # noqa: PLC0415
    except ImportError:
        logger.info(
            f"Skipping {self.slug} data ingestion: climate-ref package not installed. "
            "Run `ref datasets ingest --source-type pmp-climatology` manually if needed."
        )
        return

    data_path = self.get_data_path()
    if data_path is None or not data_path.exists():
        logger.warning(
            f"PMP data path does not exist. Run `ref providers setup --provider {self.slug}` first."
        )
        return

    # Find the pmp-climatology subdirectory
    climatology_path = data_path / "PMP_obs4MIPsClims"
    if not climatology_path.exists():
        logger.warning(f"PMP climatology data not found at {climatology_path}")
        return

    adapter = PMPClimatologyDatasetAdapter()

    try:
        stats = ingest_datasets(adapter, climatology_path, db, skip_invalid=True)
        stats.log_summary("PMP climatology ingestion complete:")
    except ValueError as e:
        logger.warning(f"No valid PMP climatology datasets found: {e}")

validate_setup(config) #

Validate conda environment and data checksums.

Source code in packages/climate-ref-pmp/src/climate_ref_pmp/__init__.py 
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
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

sub-packages#

Sub-package Description
diagnostics PMP diagnostics.
pmp_driver