Update DevOps to cache conda and fix attributes not being preserved with xarray > 2023.3.0

.github/workflows/build_workflow.yml

@@ -62,32 +62,44 @@ jobs:
             - if: ${{ steps.skip_check.outputs.should_skip != 'true' }}
               uses: actions/checkout@v3
 
-            - if: ${{ steps.skip_check.outputs.should_skip != 'true' }}
-              name: Cache Conda
-              uses: actions/cache@v3
-              env:
-                  # Increase this value to reset cache if conda-env/ci.yml has not changed in the workflow
-                  CACHE_NUMBER: 0
-              with:
-                  path: ~/conda_pkgs_dir
-                  key: ${{ runner.os }}-${{ matrix.python-version }}-conda-${{ env.CACHE_NUMBER }}
-
             - if: ${{ steps.skip_check.outputs.should_skip != 'true' }}
               name: Set up Conda Environment
               uses: conda-incubator/setup-miniconda@v2
               with:
-                  activate-environment: "xcdat_ci"
                   miniforge-variant: Mambaforge
                   miniforge-version: latest
+                  activate-environment: "xcdat_ci"
                   use-mamba: true
                   mamba-version: "*"
-                  environment-file: conda-env/ci.yml
                   channel-priority: strict
                   auto-update-conda: true
-                  # IMPORTANT: This needs to be set for caching to work properly!
-                  use-only-tar-bz2: true
                   python-version: ${{ matrix.python-version }}
 
+            # Refresh the cache every 24 hours to avoid inconsistencies of package versions
+            # between the CI pipeline and local installations.
+            - if: ${{ steps.skip_check.outputs.should_skip != 'true' }}
+              id: get-date
+              name: Get Date
+              run: echo "today=$(/bin/date -u '+%Y%m%d')" >> $GITHUB_OUTPUT
+              shell: bash
+
+            - if: ${{ steps.skip_check.outputs.should_skip != 'true' }}
+              id: cache
+              name: Cache Conda env
+              uses: actions/cache@v3
+              with:
+                  path: ${{ env.CONDA }}/envs
+                  key:
+                      conda-${{ runner.os }}-${{ runner.arch }}-${{ matrix.python-version }}-${{
+                      steps.get-date.outputs.today }}-${{hashFiles('conda-env/ci.yml') }}-${{ env.CACHE_NUMBER}}
+              env:
+                  # Increase this value to reset cache if conda-env/ci.yml has not changed in the workflow
+                  CACHE_NUMBER: 0
+
+            - if: $${{ steps.skip_check.outputs.should_skip != 'true' && steps.cache.outputs.cache-hit != 'true' }}
+              name: Update environment
+              run: mamba env update -n xcdat_ci -f conda-env/ci.yml
+
             - if: ${{ steps.skip_check.outputs.should_skip != 'true' }}
               name: Install xcdat
               # Source: https://github.com/conda/conda-build/issues/4251#issuecomment-1053460542

conda-env/ci.yml

@@ -19,7 +19,9 @@ dependencies:
     - pandas
     - python-dateutil
     - xarray
-    - xesmf
+    # Constrained because 0.6.3 breaks with import ESMF
+    # Source: https://github.com/pangeo-data/xESMF/issues/212
+    - xesmf >0.6.3
     # Quality Assurance
     # ==================
     - types-python-dateutil

tests/test_dataset.py

@@ -8,16 +8,16 @@
 from lxml import etree
 
 from tests.fixtures import generate_dataset
+from xcdat._logger import _setup_custom_logger
 from xcdat.dataset import (
     _keep_single_var,
     _postprocess_dataset,
     decode_time,
     open_dataset,
     open_mfdataset,
 )
-from xcdat.logger import setup_custom_logger
 
-logger = setup_custom_logger("xcdat.dataset", propagate=True)
+logger = _setup_custom_logger("xcdat.dataset", propagate=True)
 
 
 class TestOpenDataset:
@@ -29,6 +29,9 @@ def setup(self, tmp_path):
         self.file_path = f"{dir}/file.nc"
 
     def test_raises_warning_if_decode_times_but_no_time_coords_found(self, caplog):
+        # Silence warning to not pollute test suite output
+        caplog.set_level(logging.CRITICAL)
+
         ds = generate_dataset(decode_times=False, cf_compliant=True, has_bounds=True)
         ds = ds.drop_dims("time")
         ds.to_netcdf(self.file_path)
@@ -42,16 +45,10 @@ def test_raises_warning_if_decode_times_but_no_time_coords_found(self, caplog):
         expected = expected.drop_dims("time")
 
         assert result.identical(expected)
-        assert (
-            "No time coordinates were found in this dataset to decode. If time "
-            "coordinates were expected to exist, make sure they are detectable by "
-            "setting the CF 'axis' or 'standard_name' attribute (e.g., "
-            "ds['time'].attrs['axis'] = 'T' or "
-            "ds['time'].attrs['standard_name'] = 'time'). Afterwards, try decoding "
-            "again with `xcdat.decode_time`."
-        ) in caplog.text
-
-    def test_skip_decoding_time_explicitly(self):
+
+    def test_skip_decoding_time_explicitly(
+        self,
+    ):
         ds = generate_dataset(decode_times=False, cf_compliant=True, has_bounds=True)
         ds.to_netcdf(self.file_path)
 
@@ -602,6 +599,9 @@ def setUp(self, tmp_path):
         self.file_path2 = f"{self.dir}/file2.nc"
 
     def test_raises_warning_if_decode_times_but_no_time_coords_found(self, caplog):
+        # Silence warning to not pollute test suite output
+        caplog.set_level(logging.CRITICAL)
+
         ds = generate_dataset(decode_times=False, cf_compliant=True, has_bounds=True)
         ds = ds.drop_dims("time")
         ds.to_netcdf(self.file_path1)
@@ -615,14 +615,6 @@ def test_raises_warning_if_decode_times_but_no_time_coords_found(self, caplog):
         expected = expected.drop_dims("time")
 
         assert result.identical(expected)
-        assert (
-            "No time coordinates were found in this dataset to decode. If time "
-            "coordinates were expected to exist, make sure they are detectable by "
-            "setting the CF 'axis' or 'standard_name' attribute (e.g., "
-            "ds['time'].attrs['axis'] = 'T' or "
-            "ds['time'].attrs['standard_name'] = 'time'). "
-            "Afterwards, try decoding again with `xcdat.decode_time`."
-        ) in caplog.text
 
     def test_skip_decoding_times_explicitly(self):
         ds1 = generate_dataset(decode_times=False, cf_compliant=False, has_bounds=True)
@@ -1024,6 +1016,7 @@ def setup(self):
                 "axis": "T",
                 "long_name": "time",
                 "standard_name": "time",
+                "calendar": "standard",
             },
         )
         time_bnds = xr.DataArray(

tests/test_temporal.py

@@ -1,3 +1,5 @@
+import logging
+
 import cftime
 import numpy as np
 import pytest
@@ -6,14 +8,14 @@
 from xarray.tests import requires_dask
 
 from tests.fixtures import generate_dataset
-from xcdat.logger import setup_custom_logger
+from xcdat._logger import _setup_custom_logger
 from xcdat.temporal import (
     TemporalAccessor,
     _contains_datetime_like_objects,
     _get_datetime_like_type,
 )
 
-logger = setup_custom_logger("xcdat.temporal", propagate=True)
+logger = _setup_custom_logger("xcdat.temporal", propagate=True)
 
 
 class TestTemporalAccessor:
@@ -41,23 +43,18 @@ def test_raises_error_if_time_coords_are_not_decoded(self):
         with pytest.raises(TypeError):
             ds.temporal.average("ts")
 
-    def test_raises_warning_if_calendar_encoding_attr_not_found_on_data_var_time_coords(
-        self, caplog
-    ):
+    def test_defaults_calendar_attribute_to_standard_if_missing(self, caplog):
+        # Silence warning to not pollute test suite output
+        caplog.set_level(logging.CRITICAL)
+
         ds: xr.Dataset = generate_dataset(
             decode_times=True, cf_compliant=False, has_bounds=True
         )
         ds.ts.time.encoding = {}
 
         ds.temporal.average("ts")
 
-        assert (
-            "'time' does not have a calendar encoding attribute set, "
-            "which is used to determine the `cftime.datetime` object type for the "
-            "output time coordinates. Defaulting to CF 'standard' calendar. "
-            "Otherwise, set the calendar type (e.g., "
-            "ds['time'].encoding['calendar'] = 'noleap') and try again."
-        ) in caplog.text
+        assert ds.temporal.calendar == "standard"
 
     def test_averages_for_yearly_time_series(self):
         ds = xr.Dataset(
@@ -460,23 +457,18 @@ def test_raises_error_if_time_coords_are_not_decoded(self):
         with pytest.raises(TypeError):
             ds.temporal.group_average("ts", freq="year")
 
-    def test_raises_warning_if_calendar_encoding_attr_not_found_on_data_var_time_coords(
-        self, caplog
-    ):
+    def test_defaults_calendar_attribute_to_standard_if_missing(self, caplog):
+        # Silence warning to not pollute test suite output
+        caplog.set_level(logging.CRITICAL)
+
         ds: xr.Dataset = generate_dataset(
             decode_times=True, cf_compliant=False, has_bounds=True
         )
         ds.ts.time.encoding = {}
 
         ds.temporal.group_average("ts", freq="year")
 
-        assert (
-            "'time' does not have a calendar encoding attribute set, "
-            "which is used to determine the `cftime.datetime` object type for the "
-            "output time coordinates. Defaulting to CF 'standard' calendar. "
-            "Otherwise, set the calendar type (e.g., "
-            "ds['time'].encoding['calendar'] = 'noleap') and try again."
-        ) in caplog.text
+        assert ds.temporal.calendar == "standard"
 
     def test_weighted_annual_averages(self):
         ds = self.ds.copy()
@@ -1039,23 +1031,18 @@ def test_raises_error_if_time_coords_are_not_decoded(self):
         with pytest.raises(TypeError):
             ds.temporal.climatology("ts", freq="year")
 
-    def test_raises_warning_if_calendar_encoding_attr_not_found_on_data_var_time_coords(
-        self, caplog
-    ):
+    def test_defaults_calendar_attribute_to_standard_if_missing(self, caplog):
+        # Silence warning to not pollute test suite output
+        caplog.set_level(logging.CRITICAL)
+
         ds: xr.Dataset = generate_dataset(
             decode_times=True, cf_compliant=False, has_bounds=True
         )
         ds.ts.time.encoding = {}
 
         ds.temporal.climatology("ts", freq="season")
 
-        assert (
-            "'time' does not have a calendar encoding attribute set, "
-            "which is used to determine the `cftime.datetime` object type for the "
-            "output time coordinates. Defaulting to CF 'standard' calendar. "
-            "Otherwise, set the calendar type (e.g., "
-            "ds['time'].encoding['calendar'] = 'noleap') and try again."
-        ) in caplog.text
+        assert ds.temporal.calendar == "standard"
 
     def test_raises_error_if_reference_period_arg_is_incorrect(self):
         ds = self.ds.copy()
@@ -1742,23 +1729,18 @@ def test_raises_error_if_time_coords_are_not_decoded(self):
         with pytest.raises(TypeError):
             ds.temporal.departures("ts", freq="season")
 
-    def test_raises_warning_if_calendar_encoding_attr_not_found_on_data_var_time_coords(
-        self, caplog
-    ):
+    def test_defaults_calendar_attribute_to_standard_if_missing(self, caplog):
+        # Silence warning to not pollute test suite output
+        caplog.set_level(logging.CRITICAL)
+
         ds: xr.Dataset = generate_dataset(
             decode_times=True, cf_compliant=False, has_bounds=True
         )
         ds.ts.time.encoding = {}
 
         ds.temporal.departures("ts", freq="season")
 
-        assert (
-            "'time' does not have a calendar encoding attribute set, "
-            "which is used to determine the `cftime.datetime` object type for the "
-            "output time coordinates. Defaulting to CF 'standard' calendar. "
-            "Otherwise, set the calendar type (e.g., "
-            "ds['time'].encoding['calendar'] = 'noleap') and try again."
-        ) in caplog.text
+        assert ds.temporal.calendar == "standard"
 
     def test_raises_error_if_reference_period_arg_is_incorrect(self):
         ds = self.ds.copy()

xcdat/logger.py → xcdat/_logger.py

@@ -2,10 +2,25 @@
 import logging
 import logging.handlers
 
+# Logging module setup
+log_format = (
+    "%(asctime)s [%(levelname)s]: %(filename)s(%(funcName)s:%(lineno)s) >> %(message)s"
+)
+logging.basicConfig(format=log_format, filemode="w", level=logging.INFO)
 
-def setup_custom_logger(name: str, propagate: bool = False) -> logging.Logger:
+# Console handler setup
+console_handler = logging.StreamHandler()
+console_handler.setLevel(logging.INFO)
+logFormatter = logging.Formatter(log_format)
+console_handler.setFormatter(logFormatter)
+logging.getLogger().addHandler(console_handler)
+
+
+def _setup_custom_logger(name, propagate=True) -> logging.Logger:
     """Sets up a custom logger.
 
+    Documentation on logging: https://docs.python.org/3/library/logging.html
+
     Parameters
     ----------
     name : str
@@ -43,18 +58,7 @@ def setup_custom_logger(name: str, propagate: bool = False) -> logging.Logger:
 
     >>> logger.critical("")
     """
-    log_format = "%(asctime)s [%(levelname)s]: %(filename)s(%(funcName)s:%(lineno)s) >> %(message)s"
-    log_filemode = "w"  # w: overwrite; a: append
-
-    # Setup
-    logging.basicConfig(format=log_format, filemode=log_filemode, level=logging.INFO)
     logger = logging.getLogger(name)
     logger.propagate = propagate
 
-    # Console output
-    consoleHandler = logging.StreamHandler()
-    logFormatter = logging.Formatter(log_format)
-    consoleHandler.setFormatter(logFormatter)
-    logger.addHandler(consoleHandler)
-
     return logger

xcdat/bounds.py

@@ -12,16 +12,16 @@
 from xarray.coding.cftime_offsets import get_date_type
 from xarray.core.common import contains_cftime_datetimes
 
+from xcdat._logger import _setup_custom_logger
 from xcdat.axis import CF_ATTR_MAP, CFAxisKey, get_dim_coords
 from xcdat.dataset import _get_data_var
-from xcdat.logger import setup_custom_logger
 from xcdat.temporal import (
     _contains_datetime_like_objects,
     _get_datetime_like_type,
     _infer_freq,
 )
 
-logger = setup_custom_logger(__name__)
+logger = _setup_custom_logger(__name__)
 
 
 @xr.register_dataset_accessor("bounds")

xcdat/dataset.py

@@ -16,12 +16,12 @@
 from xarray.core.variable import as_variable
 
 from xcdat import bounds as bounds_accessor  # noqa: F401
+from xcdat._logger import _setup_custom_logger
 from xcdat.axis import CFAxisKey, _get_all_coord_keys
 from xcdat.axis import center_times as center_times_func
 from xcdat.axis import swap_lon_axis
-from xcdat.logger import setup_custom_logger
 
-logger = setup_custom_logger(__name__)
+logger = _setup_custom_logger(__name__)
 
 #: List of non-CF compliant time units.
 NON_CF_TIME_UNITS: List[str] = ["month", "months", "year", "years"]

xcdat/regridder/base.py

@@ -4,9 +4,9 @@
 import xarray as xr
 
 import xcdat.bounds  # noqa: F401
-from xcdat.logger import setup_custom_logger
+from xcdat._logger import _setup_custom_logger
 
-logger = setup_custom_logger(__name__)
+logger = _setup_custom_logger(__name__)
 
 
 def preserve_bounds(