Merge pull request #1672 from andrew-s28/logging-updates

Implement warnings in Parcels

Author: VeckoTheGecko

docs/examples/documentation_indexing.ipynb

@@ -56,6 +56,7 @@
    "source": [
     "from glob import glob\n",
     "from os import path\n",
+    "import warnings\n",
     "\n",
     "import numpy as np\n",
     "\n",
@@ -179,17 +180,17 @@
     "    \"W\": c_grid_dimensions,\n",
     "}\n",
     "\n",
-    "fieldsetC = parcels.FieldSet.from_nemo(\n",
-    "    filenames, variables, dimensions, netcdf_decodewarning=False\n",
-    ")"
+    "with warnings.catch_warnings():\n",
+    "    warnings.simplefilter(\"ignore\", parcels.FileWarning)\n",
+    "    fieldsetC = parcels.FieldSet.from_nemo(filenames, variables, dimensions)"
    ]
   },
   {
    "attachments": {},
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "Note by the way, that we used `netcdf_decodewarning=False` in the `FieldSet.from_nemo()` call above. This is to silence an expected warning because the time dimension in the `coordinates.nc` file can't be decoded by `xarray`.\n"
+    "Note by the way, that we used `warnings.catch_warnings()` with `warnings.simplefilter(\"ignore\", parcels.FileWarning)` to wrap the `FieldSet.from_nemo()` call above. This is to silence an expected warning because the time dimension in the `coordinates.nc` file can't be decoded by `xarray`.\n"
    ]
   },
   {
@@ -293,7 +294,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.11.6"
+   "version": "3.12.4"
   }
  },
  "nbformat": 4,

docs/examples/tutorial_nemo_3D.ipynb

@@ -54,15 +54,16 @@
    "source": [
     "from datetime import timedelta\n",
     "from glob import glob\n",
+    "import warnings\n",
     "\n",
     "import matplotlib.pyplot as plt\n",
     "import xarray as xr\n",
     "\n",
     "import parcels\n",
-    "from parcels import logger\n",
+    "from parcels import FileWarning\n",
     "\n",
     "# Add a filter for the xarray decoding warning\n",
-    "logger.addFilter(parcels.XarrayDecodedFilter())\n",
+    "warnings.simplefilter(\"ignore\", FileWarning)\n",
     "\n",
     "example_dataset_folder = parcels.download_example_dataset(\n",
     "    \"NemoNorthSeaORCA025-N006_data\"\n",
@@ -234,7 +235,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.4"
+   "version": "3.12.5"
   }
  },
  "nbformat": 4,

docs/examples/tutorial_timestamps.ipynb

@@ -29,6 +29,12 @@ parcels.tools.loggers module
     :members:
     :undoc-members:
 
+parcels.tools.warnings module
+
+.. automodule:: parcels.tools.warnings
+    :members:
+    :undoc-members:
+
 parcels.tools.exampledata_utils module
 --------------------------------------
 

docs/reference/misc.rst

@@ -2,6 +2,7 @@
 import collections
 import math
 import random
+import warnings
 from abc import ABC
 from copy import copy
 
@@ -10,8 +11,8 @@
 from parcels.field import Field, NestedField, VectorField
 from parcels.grid import Grid
 from parcels.particle import JITParticle
-from parcels.tools.loggers import logger
 from parcels.tools.statuscodes import StatusCode
+from parcels.tools.warnings import KernelWarning
 
 
 class IntrinsicNode(ast.AST):
@@ -178,9 +179,11 @@ def __init__(self, obj, attr):
 
 class ParticleXiYiZiTiAttributeNode(IntrinsicNode):
     def __init__(self, obj, attr):
-        logger.warning_once(
+        warnings.warn(
             f"Be careful when sampling particle.{attr}, as this is updated in the kernel loop. "
-            "Best to place the sampling statement before advection."
+            "Best to place the sampling statement before advection.",
+            KernelWarning,
+            stacklevel=2,
         )
         self.obj = obj.ccode
         self.attr = attr
@@ -309,8 +312,10 @@ def visit_Subscript(self, node):
     def visit_AugAssign(self, node):
         node.target = self.visit(node.target)
         if isinstance(node.target, ParticleAttributeNode) and node.target.attr in ["lon", "lat", "depth", "time"]:
-            logger.warning_once(
-                "Don't change the location of a particle directly in a Kernel. Use particle_dlon, particle_dlat, etc."
+            warnings.warn(
+                "Don't change the location of a particle directly in a Kernel. Use particle_dlon, particle_dlat, etc.",
+                KernelWarning,
+                stacklevel=2,
             )
         node.op = self.visit(node.op)
         node.value = self.visit(node.value)
@@ -439,7 +444,11 @@ def generate(self, py_ast, funcvars: list[str]):
         for kvar in funcvars:
             if kvar in used_vars + ["particle_dlon", "particle_dlat", "particle_ddepth"]:
                 if kvar not in ["particle", "fieldset", "time", "particle_dlon", "particle_dlat", "particle_ddepth"]:
-                    logger.warning(kvar + " declared in multiple Kernels")
+                    warnings.warn(
+                        kvar + " declared in multiple Kernels",
+                        KernelWarning,
+                        stacklevel=2,
+                    )
                 funcvars_copy.remove(kvar)
             else:
                 used_vars.append(kvar)

parcels/compilation/codegenerator.py

@@ -1,6 +1,7 @@
 import collections
 import datetime
 import math
+import warnings
 from ctypes import POINTER, Structure, c_float, c_int, pointer
 from pathlib import Path
 from typing import TYPE_CHECKING, Iterable, Type
@@ -18,14 +19,14 @@
     UnitConverter,
     unitconverters_map,
 )
-from parcels.tools.loggers import logger
 from parcels.tools.statuscodes import (
     AllParcelsErrorCodes,
     FieldOutOfBoundError,
     FieldOutOfBoundSurfaceError,
     FieldSamplingError,
     TimeExtrapolationError,
 )
+from parcels.tools.warnings import FieldSetWarning, _deprecated_param_netcdf_decodewarning
 
 from .fieldfilebuffer import (
     DaskFileBuffer,
@@ -163,6 +164,10 @@ def __init__(
         to_write=False,
         **kwargs,
     ):
+        if kwargs.get("netcdf_decodewarning") is not None:
+            _deprecated_param_netcdf_decodewarning()
+            kwargs.pop("netcdf_decodewarning")
+
         if not isinstance(name, tuple):
             self.name = name
             self.filebuffername = name
@@ -211,8 +216,10 @@ def __init__(
             GridType.RectilinearSGrid,
             GridType.CurvilinearSGrid,
         ]:
-            logger.warning_once(  # type: ignore
-                "General s-levels are not supported in B-grid. RectilinearSGrid and CurvilinearSGrid can still be used to deal with shaved cells, but the levels must be horizontal."
+            warnings.warn(
+                "General s-levels are not supported in B-grid. RectilinearSGrid and CurvilinearSGrid can still be used to deal with shaved cells, but the levels must be horizontal.",
+                FieldSetWarning,
+                stacklevel=2,
             )
 
         self.fieldset: "FieldSet" | None = None
@@ -223,9 +230,10 @@ def __init__(
 
         self.time_periodic = time_periodic
         if self.time_periodic is not False and self.allow_time_extrapolation:
-            logger.warning_once(  # type: ignore
-                "allow_time_extrapolation and time_periodic cannot be used together.\n \
-                                 allow_time_extrapolation is set to False"
+            warnings.warn(
+                "allow_time_extrapolation and time_periodic cannot be used together. allow_time_extrapolation is set to False",
+                FieldSetWarning,
+                stacklevel=2,
             )
             self.allow_time_extrapolation = False
         if self.time_periodic is True:
@@ -275,9 +283,8 @@ def __init__(
             self.dataFiles = np.append(self.dataFiles, self.dataFiles[0])
         self._field_fb_class = kwargs.pop("FieldFileBuffer", None)
         self.netcdf_engine = kwargs.pop("netcdf_engine", "netcdf4")
-        self.netcdf_decodewarning = kwargs.pop("netcdf_decodewarning", True)
-        self.loaded_time_indices: Iterable[int] = []
-        self.creation_log: str = kwargs.pop("creation_log", "")
+        self.loaded_time_indices: Iterable[int] = []  # type: ignore
+        self.creation_log = kwargs.pop("creation_log", "")
         self.chunksize = kwargs.pop("chunksize", None)
         self.netcdf_chunkdims_name_map = kwargs.pop("chunkdims_name_map", None)
         self.grid.depth_field = kwargs.pop("depth_field", None)
@@ -315,8 +322,10 @@ def get_dim_filenames(cls, filenames, dim):
 
     @staticmethod
     def collect_timeslices(
-        timestamps, data_filenames, _grid_fb_class, dimensions, indices, netcdf_engine, netcdf_decodewarning=True
+        timestamps, data_filenames, _grid_fb_class, dimensions, indices, netcdf_engine, netcdf_decodewarning=None
     ):
+        if netcdf_decodewarning is not None:
+            _deprecated_param_netcdf_decodewarning()
         if timestamps is not None:
             dataFiles = []
             for findex in range(len(data_filenames)):
@@ -329,9 +338,7 @@ def collect_timeslices(
             timeslices = []
             dataFiles = []
             for fname in data_filenames:
-                with _grid_fb_class(
-                    fname, dimensions, indices, netcdf_engine=netcdf_engine, netcdf_decodewarning=netcdf_decodewarning
-                ) as filebuffer:
+                with _grid_fb_class(fname, dimensions, indices, netcdf_engine=netcdf_engine) as filebuffer:
                     ftime = filebuffer.time
                     timeslices.append(ftime)
                     dataFiles.append([fname] * len(ftime))
@@ -408,7 +415,7 @@ def from_netcdf(
         chunksize :
             size of the chunks in dask loading
         netcdf_decodewarning : bool
-            Whether to show a warning id there is a problem decoding the netcdf files.
+            (DEPRECATED - v3.1.0) Whether to show a warning if there is a problem decoding the netcdf files.
             Default is True, but in some cases where these warnings are expected, it may be useful to silence them
             by setting netcdf_decodewarning=False.
         grid :
@@ -423,6 +430,10 @@ def from_netcdf(
         * `Timestamps <../examples/tutorial_timestamps.ipynb>`__
 
         """
+        if kwargs.get("netcdf_decodewarning") is not None:
+            _deprecated_param_netcdf_decodewarning()
+            kwargs.pop("netcdf_decodewarning")
+
         # Ensure the timestamps array is compatible with the user-provided datafiles.
         if timestamps is not None:
             if isinstance(filenames, list):
@@ -475,7 +486,6 @@ def from_netcdf(
             depth_filename = depth_filename[0]
 
         netcdf_engine = kwargs.pop("netcdf_engine", "netcdf4")
-        netcdf_decodewarning = kwargs.pop("netcdf_decodewarning", True)
 
         indices = {} if indices is None else indices.copy()
         for ind in indices:
@@ -498,9 +508,7 @@ def from_netcdf(
 
         _grid_fb_class = NetcdfFileBuffer
 
-        with _grid_fb_class(
-            lonlat_filename, dimensions, indices, netcdf_engine, netcdf_decodewarning=netcdf_decodewarning
-        ) as filebuffer:
+        with _grid_fb_class(lonlat_filename, dimensions, indices, netcdf_engine) as filebuffer:
             lon, lat = filebuffer.lonlat
             indices = filebuffer.indices
             # Check if parcels_mesh has been explicitly set in file
@@ -514,7 +522,6 @@ def from_netcdf(
                 indices,
                 netcdf_engine,
                 interp_method=interp_method,
-                netcdf_decodewarning=netcdf_decodewarning,
             ) as filebuffer:
                 filebuffer.name = filebuffer.parse_name(variable[1])
                 if dimensions["depth"] == "not_yet_set":
@@ -537,7 +544,7 @@ def from_netcdf(
             # Concatenate time variable to determine overall dimension
             # across multiple files
             time, time_origin, timeslices, dataFiles = cls.collect_timeslices(
-                timestamps, data_filenames, _grid_fb_class, dimensions, indices, netcdf_engine, netcdf_decodewarning
+                timestamps, data_filenames, _grid_fb_class, dimensions, indices, netcdf_engine
             )
             grid = Grid.create_grid(lon, lat, depth, time, time_origin=time_origin, mesh=mesh)
             grid.timeslices = timeslices
@@ -546,15 +553,17 @@ def from_netcdf(
             # ==== means: the field has a shared grid, but may have different data files, so we need to collect the
             # ==== correct file time series again.
             _, _, _, dataFiles = cls.collect_timeslices(
-                timestamps, data_filenames, _grid_fb_class, dimensions, indices, netcdf_engine, netcdf_decodewarning
+                timestamps, data_filenames, _grid_fb_class, dimensions, indices, netcdf_engine
             )
             kwargs["dataFiles"] = dataFiles
 
         chunksize: bool | None = kwargs.get("chunksize", None)
         grid.chunksize = chunksize
 
         if "time" in indices:
-            logger.warning_once("time dimension in indices is not necessary anymore. It is then ignored.")  # type: ignore
+            warnings.warn(
+                "time dimension in indices is not necessary anymore. It is then ignored.", FieldSetWarning, stacklevel=2
+            )
 
         if "full_load" in kwargs:  # for backward compatibility with Parcels < v2.0.0
             deferred_load = not kwargs["full_load"]
@@ -587,7 +596,6 @@ def from_netcdf(
                     interp_method=interp_method,
                     data_full_zdim=data_full_zdim,
                     chunksize=chunksize,
-                    netcdf_decodewarning=netcdf_decodewarning,
                 ) as filebuffer:
                     # If Field.from_netcdf is called directly, it may not have a 'data' dimension
                     # In that case, assume that 'name' is the data dimension
@@ -632,7 +640,6 @@ def from_netcdf(
         kwargs["indices"] = indices
         kwargs["time_periodic"] = time_periodic
         kwargs["netcdf_engine"] = netcdf_engine
-        kwargs["netcdf_decodewarning"] = netcdf_decodewarning
 
         return cls(
             variable,
@@ -820,16 +827,13 @@ def calc_cell_edge_sizes(self):
                         self.grid.cell_edge_sizes["y"][y, x] = y_conv.to_source(dy, lon, lat, self.grid.depth[0])
                 self.cell_edge_sizes = self.grid.cell_edge_sizes
             else:
-                logger.error(
+                raise ValueError(
                     (
-                        "Field.cell_edge_sizes() not implemented for ",
-                        self.grid.gtype,
-                        "grids.",
-                        "You can provide Field.grid.cell_edge_sizes yourself",
-                        "by in e.g. NEMO using the e1u fields etc from the mesh_mask.nc file",
+                        f"Field.cell_edge_sizes() not implemented for {self.grid.gtype} grids. "
+                        "You can provide Field.grid.cell_edge_sizes yourself by in, e.g., "
+                        "NEMO using the e1u fields etc from the mesh_mask.nc file."
                     )
                 )
-                exit(-1)
 
     def cell_areas(self):
         """Method to calculate cell sizes based on cell_edge_sizes.
@@ -1347,8 +1351,10 @@ def time_index(self, time):
 
     def _check_velocitysampling(self):
         if self.name in ["U", "V", "W"]:
-            logger.warning_once(
-                "Sampling of velocities should normally be done using fieldset.UV or fieldset.UVW object; tread carefully"
+            warnings.warn(
+                "Sampling of velocities should normally be done using fieldset.UV or fieldset.UVW object; tread carefully",
+                RuntimeWarning,
+                stacklevel=2,
             )
 
     def __getitem__(self, key):
@@ -1653,7 +1659,6 @@ def computeTimeChunk(self, data, tindex):
             cast_data_dtype=self.cast_data_dtype,
             rechunk_callback_fields=rechunk_callback_fields,
             chunkdims_name_map=self.netcdf_chunkdims_name_map,
-            netcdf_decodewarning=self.netcdf_decodewarning,
         )
         filebuffer.__enter__()
         time_data = filebuffer.time