Add regrid operator (#111)

* add regrid operation

* Updated tests to include fixture; got test_regrid running.

* update conda env with clisops >= 0.12

* update requirements for clisops

* added roocs-grids conda package

* test with conda on github

* added xesmf

* fix errors in docs build

* update rtd config

* updated history

* pep8

* enabled lint

* test also on python 3.10 and 3.11

---------

Co-authored-by: ellesmith88 <[email protected]>
Co-authored-by: Ag Stephens <[email protected]>

.github/workflows/main.yml

@@ -5,37 +5,46 @@ on: [ push, pull_request ]
 jobs:
   build:
 
+    name: Conda Build with Python${{ matrix.python-version }}
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        python-version: ["3.8", "3.9", "3.10", "3.11"]
+        python-version: ["3.9", "3.10", "3.11"]
+    defaults:
+      run:
+        shell: bash -l {0}
     steps:
-    - uses: actions/checkout@v2
+    - uses: actions/checkout@v3
     - name: Install packages
       run: |
         sudo apt-get -y install pandoc
-      if: matrix.python-version == 3.8
-    - name: Set up Python ${{ matrix.python-version }}
-      uses: actions/setup-python@v2
+      if: matrix.python-version == 3.9
+    - name: Setup Conda (Micromamba) with Python${{ matrix.python-version }}
+      uses: mamba-org/setup-micromamba@v1
       with:
-        python-version: ${{ matrix.python-version }}
+        cache-downloads: true
+        cache-environment: true
+        environment-file: environment.yml
+        create-args: >-
+          conda
+          python=${{ matrix.python-version }}
+    - name: Conda and Mamba versions
+      run: |
+        conda --version
+        echo "micromamba: $(micromamba --version)"
     - name: Install dependencies
       run: |
-        python -m pip install --upgrade pip
-        # pip install flake8 black pytest
-        if [ -f requirements.txt ]; then pip install -r requirements.txt; fi
+        python -m pip install --no-user --editable "."
         if [ -f requirements_dev.txt ]; then pip install -r requirements_dev.txt; fi
-        # pip install the package, command-line unit tests work 
-        pip install --no-deps -e .
     - name: Lint with flake8
       run: make lint
-      if: matrix.python-version == 3.8
+      if: matrix.python-version == 3.9
 #    - name: Check formatting with black
-#      run: black --check --target-version py38 daops tests
-#      if: matrix.python-version == 3.8
+#      run: black --check --target-version py39 daops tests
+#      if: matrix.python-version == 3.9
     - name: Test with pytest
       run: |
         python -m pytest -v tests
     - name: Build docs 🏗️
       run: make docs
-      if: matrix.python-version == 3.8
+      if: matrix.python-version == 3.9

daops/ops/regrid.py

@@ -0,0 +1,80 @@
+from clisops.ops.regrid import regrid as clisops_regrid
+from roocs_utils.parameter import collection_parameter
+from roocs_utils.parameter import dimension_parameter
+
+from daops.ops.base import Operation
+
+__all__ = [
+    "regrid",
+]
+
+
+class Regrid(Operation):
+    def _resolve_params(self, collection, **params):
+        """
+        Resolve the input parameters to `self.params` and parameterise
+        collection parameter and set to `self.collection`.
+        """
+        # need to verify method, grid and adaptive masking threshold are correct format - parameters?
+        collection = collection_parameter.CollectionParameter(collection)
+
+        self.collection = collection
+        self.params = {
+            "method": params.get("method"),
+            "adaptive_masking_threshold": params.get("adaptive_masking_threshold"),
+            "grid": params.get("grid"),
+        }
+
+    def get_operation_callable(self):
+        return clisops_regrid
+
+
+def regrid(
+    collection,
+    method="nn",
+    adaptive_masking_threshold=0.5,
+    grid="1deg",
+    output_dir=None,
+    output_type="netcdf",
+    split_method="time:auto",
+    file_namer="standard",
+    apply_fixes=True,
+):
+    """
+    Regrid input dataset according to specified method and output grid.
+    The adaptive masking threshold can also be specified.
+
+    Parameters
+    ----------
+    collection: Collection of datasets to process, sequence or string of comma separated dataset identifiers.
+    method: The method by which to regrid.
+    adaptive_masking_threshold:
+    grid: The desired output grid.
+    output_dir: str or path like object describing output directory for output files.
+    output_type: {"netcdf", "nc", "zarr", "xarray"}
+    split_method: {"time:auto"}
+    file_namer: {"standard", "simple"}
+    apply_fixes: Boolean. If True fixes will be applied to datasets if needed. Default is True.
+
+    Returns
+    -------
+    List of outputs in the selected type: a list of xarray Datasets or file paths.
+
+
+    Examples
+    --------
+    | collection: ("cmip6.ukesm1.r1.gn.tasmax.v20200101",)
+    | method: "nn"
+    | adaptive_masking_threshold: 0.5
+    | grid: "1deg"
+    | output_type: "netcdf"
+    | output_dir: "/cache/wps/procs/req0111"
+    | split_method: "time:auto"
+    | file_namer: "standard"
+    | apply_fixes: True
+
+    """
+
+    result_set = Regrid(**locals()).calculate()
+
+    return result_set

environment.yml

@@ -5,15 +5,17 @@ channels:
   - defaults
 dependencies:
   - pip
-  - numpy>=1.16
-  - xarray>=0.15
+  - numpy >=1.16
+  - xarray >=0.21,<2023.3.0 # https://github.com/pydata/xarray/issues/7794
   - dask
   - netcdf4
-  - clisops>=0.10.0
+  - clisops >=0.12.0
+  - roocs-grids >=0.1.2
+  - xesmf >=0.8.2
   - elasticsearch>=7.9.1
   - roocs-utils>=0.6.2
   # logging
-  - loguru>=0.5.3
+  - loguru >=0.5.3
   # catalog
   - intake
   - fsspec

requirements.txt

@@ -1,11 +1,14 @@
 numpy>=1.16
-xarray>=0.19
+xarray>=0.21
+xesmf>=0.8.2
 dask[complete]
 cftime
 netcdf4
 elasticsearch>=8.0.1
-clisops>=0.10.0
-roocs-utils>=0.6.2
+clisops>=0.12.0
+## clisops @ git+https://github.com/roocs/clisops.git@master#egg=clisops
+roocs-utils>=0.6.4
+roocs_grids>=0.1.2
 # logging
 loguru>=0.5.3
 # catalog

tests/test_operations/test_regrid.py

@@ -0,0 +1,36 @@
+import os
+
+import pytest
+import xarray as xr
+
+
+# from daops.ops.regrid import regrid
+
+# TODO: remove when upgraded to new clisops version
+# pytestmark = pytest.mark.xfail(reason="needs clisops>=0.12 with regrid operator")
+
+CMIP6_IDS = ["CMIP6.CMIP.MPI-M.MPI-ESM1-2-HR.historical.r1i1p1f1.Omon.tos.gn.v20190710"]
+
+
+def _check_output_nc(result, fname="output_001.nc"):
+    assert fname in [os.path.basename(_) for _ in result.file_uris]
+
+
+@pytest.mark.online
+def test_regrid(tmpdir, load_esgf_test_data):
+    from daops.ops.regrid import regrid
+
+    result = regrid(
+        CMIP6_IDS[0],
+        method="conservative",
+        adaptive_masking_threshold=0.5,
+        grid="1deg",
+        output_dir=tmpdir,
+        file_namer="simple",
+        apply_fixes=False,
+    )
+
+    _check_output_nc(result)
+    ds = xr.open_dataset(result.file_uris[0], use_cftime=True)
+    assert "time" in ds.dims
+    assert "tos" in ds