add tests for translator funcs

src/access_nri_intake/catalog/translators.py

@@ -6,7 +6,6 @@
 like the ACCESS-NRI catalog
 """
 
-import re
 from functools import partial
 
 import pandas as pd
@@ -172,13 +171,13 @@ def _realm_translator(self):
         """
         Return realm, fixing a few issues
         """
-        return _cmip_realm_translator(self.source.df)
+        return _cmip_realm_translator(self.source.df["realm"])
 
     def _frequency_translator(self):
         """
         Return frequency, fixing a few issues
         """
-        return _cmip_frequency_translator(self.source.df)
+        return _to_tuple(_cmip_frequency_translator(self.source.df["frequency"]))
 
     def _variable_translator(self):
         """
@@ -220,13 +219,13 @@ def _realm_translator(self):
         """
         Return realm, fixing a few issues
         """
-        return _cmip_realm_translator(self.source.df)
+        return _cmip_realm_translator(self.source.df["realm"])
 
     def _frequency_translator(self):
         """
         Return frequency, fixing a few issues
         """
-        return _cmip_frequency_translator(self.source.df)
+        return _to_tuple(_cmip_frequency_translator(self.source.df["frequency"]))
 
     def _variable_translator(self):
         """
@@ -249,7 +248,7 @@ def __init__(self, source, columns):
         source: :py:class:`~intake.DataSource`
             The NCI ERA-Interim intake-esm datastore
         columns: list of str
-            The columns to translate (these are the core columns in the intake-dataframe-catalog)
+            The columns to translate to (these are the core columns in the intake-dataframe-catalog)
         """
 
         super().__init__(source, columns)
@@ -262,58 +261,66 @@ def _variable_translator(self):
         return _to_tuple(self.source.df["variable"])
 
 
-def _cmip_frequency_translator(df):
+def _cmip_frequency_translator(series):
     """
     Return frequency from CMIP frequency metadata
     """
 
-    def _parse(string):
-        for remove in ["Pt", "C.*"]:  # Remove Pt, C, and Clim
-            string = re.sub(remove, "", string)
-        string = string.replace("daily", "day")  # Some incorrect metadata
-        string = string.replace("sem", "3mon")  # CORDEX for seasonal mean
-        return (f"1{string}",) if string[0] in ["m", "d", "y"] else (string,)
+    def _translate(string):
+        translations = {
+            "3hrPt": "3hr",
+            "6hrPt": "6hr",
+            "daily": "1day",
+            "day": "1day",
+            "mon": "1mon",
+            "monC": "1mon",
+            "monClim": "1mon",
+            "monPt": "1mon",
+            "sem": "3mon",
+            "subhrPt": "subhr",
+            "yr": "1yr",
+            "yrPt": "1yr",
+        }
+
+        try:
+            return translations[string]
+        except KeyError:
+            return string
 
-    return df["frequency"].apply(lambda string: _parse(string))
+    return series.apply(lambda string: _translate(string))
 
 
-def _cmip_realm_translator(df):
+def _cmip_realm_translator(series):
     """
-    Return realm from CMIP realm metadata, fixing some issues
+    Return realm from CMIP realm metadata, fixing some issues. This function returns
+    a tuple as there are sometimes multiple realms per cmip asset
     """
 
-    def _parse(string):
+    def _translate(string):
+        translations = {
+            "na": "none",
+            "landonly": "land",
+            "ocnBgChem": "ocnBgchem",
+            "seaice": "seaIce",
+        }
+
         raw_realms = string.split(" ")
         realms = []
         for realm in raw_realms:
-            if re.match("na", realm, flags=re.I):
-                realms.append("none")
-            elif re.match("seaIce", realm, flags=re.I):
-                realms.append("seaIce")
-            elif re.match("landIce", realm, flags=re.I):
-                realms.append("landIce")
-            elif re.match("ocnBgchem", realm, flags=re.I):
-                realms.append("ocnBgchem")
-            elif re.match("atmos", realm, flags=re.I):
-                realms.append("atmos")
-            elif re.match("atmosChem", realm, flags=re.I):
-                realms.append("atmosChem")
-            elif re.match("aerosol", realm, flags=re.I):
-                realms.append("aerosol")
-            elif re.match("land", realm, flags=re.I):
-                realms.append("land")
-            elif re.match("ocean", realm, flags=re.I):
-                realms.append("ocean")
-            else:
-                realms.append("unknown")
-        return tuple(set(realms))
-
-    return df["realm"].apply(lambda string: _parse(string))
+            try:
+                realm = translations[realm]
+            except KeyError:
+                pass
+            if realm not in realms:
+                realms.append(realm)
+        return tuple(realms)
+
+    return series.apply(lambda string: _translate(string))
 
 
 def _to_tuple(series):
     """
-    Make entries in the provided series a tuple
+    Make each entry in the provided series a tuple
 
     Parameters
     ----------

tests/test_translators.py

@@ -0,0 +1,129 @@
+# Copyright 2023 ACCESS-NRI and contributors. See the top-level COPYRIGHT file for details.
+# SPDX-License-Identifier: Apache-2.0
+
+import pandas as pd
+import pytest
+
+from access_nri_intake.catalog.translators import (
+    _cmip_frequency_translator,
+    _cmip_realm_translator,
+    _to_tuple,
+)
+
+
+@pytest.mark.parametrize(
+    "input, expected",
+    [
+        (
+            [
+                "1hr",
+                "3hr",
+                "3hrPt",
+                "6hr",
+                "6hrPt",
+                "daily",
+                "day",
+                "fx",
+                "mon",
+                "monC",
+                "monClim",
+                "monPt",
+                "sem",
+                "subhr",
+                "subhrPt",
+                "yr",
+                "yrPt",
+            ],
+            [
+                "1hr",
+                "3hr",
+                "3hr",
+                "6hr",
+                "6hr",
+                "1day",
+                "1day",
+                "fx",
+                "1mon",
+                "1mon",
+                "1mon",
+                "1mon",
+                "3mon",
+                "subhr",
+                "subhr",
+                "1yr",
+                "1yr",
+            ],
+        ),
+        (["daily"], ["1day"]),
+    ],
+)
+def test_cmip_frequency_translator(input, expected):
+    """Test translation of entries in the CMIP frequency column"""
+    series = pd.Series(input)
+    translated = _cmip_frequency_translator(series)
+    assert list(translated) == expected
+
+
+@pytest.mark.parametrize(
+    "input, expected",
+    [
+        (
+            [
+                "aerosol",
+                "atmos",
+                "atmos atmosChem",
+                "atmos land",
+                "land",
+                "landIce",
+                "landIce land",
+                "landonly",
+                "na",
+                "ocean",
+                "ocean seaIce",
+                "ocnBgChem",
+                "ocnBgchem",
+                "seaIce",
+                "seaIce ocean",
+                "seaice",
+            ],
+            [
+                ("aerosol",),
+                ("atmos",),
+                ("atmos", "atmosChem"),
+                ("atmos", "land"),
+                ("land",),
+                ("landIce",),
+                ("landIce", "land"),
+                ("land",),
+                ("none",),
+                ("ocean",),
+                ("ocean", "seaIce"),
+                ("ocnBgchem",),
+                ("ocnBgchem",),
+                ("seaIce",),
+                ("seaIce", "ocean"),
+                ("seaIce",),
+            ],
+        ),
+        (["landonly"], [("land",)]),
+        (["atmos atmosChem"], [("atmos", "atmosChem")]),
+    ],
+)
+def test_cmip_realm_translator(input, expected):
+    """Test translation of entries in the CMIP realm column"""
+    series = pd.Series(input)
+    translated = _cmip_realm_translator(series)
+    assert list(translated) == expected
+
+
+@pytest.mark.parametrize(
+    "input",
+    [
+        ["a", "b", "c"],
+        [0, 1, 2],
+    ],
+)
+def test_to_tuple(input):
+    """Test the _to_tuple function"""
+    series = pd.Series(input)
+    assert all(_to_tuple(series).map(type) == tuple)