add ACCESS-OM3 builder and basic tests

src/access_nri_intake/source/builders.py

@@ -266,6 +266,91 @@ def parser(file):
             return {INVALID_ASSET: file, TRACEBACK: traceback.format_exc()}
 
 
+class AccessOm3Builder(BaseBuilder):
+    """Intake-ESM datastore builder for ACCESS-OM3 COSIMA datasets"""
+
+    def __init__(self, path):
+        """
+        Initialise a AccessOm3Builder
+
+        Parameters
+        ----------
+        path : str or list of str
+            Path or list of paths to crawl for assets/files.
+        """
+
+        kwargs = dict(
+            path=path,
+            depth=2,
+            exclude_patterns=[
+                "*restart*",
+                "*MOM_IC.nc",
+                "*ocean_geometry.nc",
+                "*ocean.stats.nc",
+                "*Vertical_coordinate.nc",
+            ],
+            include_patterns=["*.nc"],
+            data_format="netcdf",
+            groupby_attrs=["file_id", "frequency"],
+            aggregations=[
+                {
+                    "type": "join_existing",
+                    "attribute_name": "start_date",
+                    "options": {
+                        "dim": "time",
+                        "combine": "by_coords",
+                    },
+                },
+            ],
+        )
+
+        super().__init__(**kwargs)
+
+    @staticmethod
+    def parser(file):
+        try:
+            (
+                filename,
+                file_id,
+                _,
+                frequency,
+                start_date,
+                end_date,
+                variable_list,
+                variable_long_name_list,
+                variable_standard_name_list,
+                variable_cell_methods_list,
+                variable_units_list,
+            ) = parse_access_ncfile(file)
+
+            if ("mom6" in filename) or ("ww3" in filename):
+                realm = "ocean"
+            elif "cice" in filename:
+                realm = "seaIce"
+            else:
+                raise ParserError(f"Cannot determine realm for file {file}")
+
+            info = {
+                "path": str(file),
+                "realm": realm,
+                "variable": variable_list,
+                "frequency": frequency,
+                "start_date": start_date,
+                "end_date": end_date,
+                "variable_long_name": variable_long_name_list,
+                "variable_standard_name": variable_standard_name_list,
+                "variable_cell_methods": variable_cell_methods_list,
+                "variable_units": variable_units_list,
+                "filename": filename,
+                "file_id": file_id,
+            }
+
+            return info
+
+        except Exception:
+            return {INVALID_ASSET: file, TRACEBACK: traceback.format_exc()}
+
+
 class AccessEsm15Builder(BaseBuilder):
     """Intake-ESM datastore builder for ACCESS-ESM1.5 datasets"""
 

src/access_nri_intake/source/utils.py

@@ -169,7 +169,10 @@ def parse_access_filename(filename):
     """
 
     # ACCESS output file patterns
+    # TODO: these should be defined per driver to prevent new patterns from breaking old drivers
     not_multi_digit = "(?:\\d(?!\\d)|[^\\d](?=\\d)|[^\\d](?!\\d))"
+    om3_components = "(?:cice|mom6|ww3)"
+    ymds = "\\d{4}[_,-]\\d{2}[_,-]\\d{2}[_,-]\\d{5}"
     ymd = "\\d{4}[_,-]\\d{2}[_,-]\\d{2}"
     ym = "\\d{4}[_,-]\\d{2}"
     y = "\\d{4}"
@@ -181,6 +184,7 @@ def parse_access_filename(filename):
         r"^ocean.*[^\d]_(\d{2})$",  # A few wierd files in ACCESS-OM2 01deg_jra55v13_ryf9091
         r"^.*\.p.(\d{6})_.*",  # ACCESS-CM2 atmosphere
         r"^.*\.p.-(\d{6})_.*",  # ACCESS-ESM1.5 atmosphere
+        rf"[^\.]*\.{om3_components}\..*({ymds}|{ymd})$",  # ACCESS-OM3
     ]
     # Frequency translations
     frequencies = {

tests/test_source_utils.py

@@ -104,6 +104,47 @@
             "rregionPrydz_temp_xflux_adv",
             ("rregionPrydz_temp_xflux_adv", None, None),
         ),
+        # Example ACCESS-OM3 filenames
+        (
+            "GMOM_JRA_WD.ww3.hi.1958-01-02-00000",
+            (
+                "GMOM_JRA_WD_ww3_hi_XXXX_XX_XX_XXXXX",
+                "1958-01-02-00000",
+                None,
+            ),
+        ),
+        (
+            "GMOM_JRA.cice.h.1900-01-01",
+            (
+                "GMOM_JRA_cice_h_XXXX_XX_XX",
+                "1900-01-01",
+                None,
+            ),
+        ),
+        (
+            "GMOM_JRA.mom6.ocean_sfc_1900_01_01",
+            (
+                "GMOM_JRA_mom6_ocean_sfc_XXXX_XX_XX",
+                "1900_01_01",
+                None,
+            ),
+        ),
+        (
+            "GMOM_JRA.mom6.sfc_1900_01_01",
+            (
+                "GMOM_JRA_mom6_sfc_XXXX_XX_XX",
+                "1900_01_01",
+                None,
+            ),
+        ),
+        (
+            "GMOM_JRA.mom6.static",
+            (
+                "GMOM_JRA_mom6_static",
+                None,
+                None,
+            ),
+        ),
     ],
 )
 def test_parse_access_filename(filename, expected):