401: tests for mode_2d_arearat mode_3d_ratio ctc

metcalcpy/util/ctc_statistics.py

@@ -114,7 +114,7 @@ def calculate_fbias(input_data, columns_names, logger=None):
         result = oyn / oy
         result = round_half_up(result, PRECISION)
         safe_log(logger, "info", f"ACC calculation successful: {result}")
-    except (TypeError, ZeroDivisionError, Warning, ValueError):
+    except (TypeError, ZeroDivisionError, Warning, ValueError) as e:
         safe_log(logger, "error", f"Error in ACC calculation: {e}")
         result = None
     warnings.filterwarnings('ignore')
@@ -150,7 +150,7 @@ def calculate_fmean(input_data, columns_names, logger=None):
         result = oyn / total
         result = round_half_up(result, PRECISION)
         safe_log(logger, "info", f"FMEAN calculation successful: {result}")
-    except (TypeError, ZeroDivisionError, Warning, ValueError):
+    except (TypeError, ZeroDivisionError, Warning, ValueError) as e:
         safe_log(logger, "error", f"Error in FMEAN calculation: {e}")
         result = None
     warnings.filterwarnings('ignore')
@@ -317,7 +317,7 @@ def calculate_podn(input_data, columns_names, logger=None):
     return result
 
 
-def calculate_far(input_data, column_names, logger=None):
+def calculate_far(input_data, columns_names, logger=None):
     """Performs calculation of FAR - false alarms
 
         Args:
@@ -941,11 +941,15 @@ def calculate_ctc_on(input_data, columns_names, logger=None):
             calculated ON as float
             or None if some data values are missing or invalid
     """
-    safe_log(logger, "debug", "Starting calculation of CTC ON.")
-    fy_on = sum_column_data_by_name(input_data, columns_names, 'fy_on')
-    fn_on = sum_column_data_by_name(input_data, columns_names, 'fn_on')
-    return round_half_up(fy_on + fn_on, PRECISION)
-
+    try:
+        safe_log(logger, "debug", "Starting calculation of CTC ON.")
+        fy_on = sum_column_data_by_name(input_data, columns_names, 'fy_on')
+        fn_on = sum_column_data_by_name(input_data, columns_names, 'fn_on')
+        result = round_half_up(fy_on + fn_on, PRECISION)
+    except (TypeError, ZeroDivisionError, Warning, ValueError) as e:
+        result = None
+        safe_log(logger, "error", f"Error in calculating CTC ON: {str(e)}")
+    return result
 
 def calculate_ctc_fy(input_data, columns_names, logger=None):
     """Calculates the Total Number of forecast yes and observation no plus

test/test_ctc_statistics.py

@@ -1,7 +1,10 @@
 import pytest
 import pandas as pd
+import numpy as np
 import sys
 import os
+from unittest.mock import patch
+from functools import cache
 sys.path.append("../../")
 from metcalcpy.util import ctc_statistics as ctc
 
@@ -139,6 +142,63 @@ def test_CTC_ROC_thresh():
     # if we get here, then all elements matched in value and position
     assert True
 
+@cache
+def ctc_data():
+    df = pd.read_csv(f"{cwd}/data/ROC_CTC.data", sep='\t', header='infer')
+    return df.to_numpy(), np.array(df.columns)
+
+
+# TODO: Some of the functions here that
+# are evaluating to None could be better 
+# tested with a different dataset. e.g. acc
+# needs a 'total' column in the test data.
+@pytest.mark.parametrize(
+    "func_name,expected",
+    [
+        ("fbias", 0.8549794),
+        ("acc", None),
+        ("fmean", None),
+        ("pody", 0.5603757),
+        ("pofd", 0.0368738),
+        ("podn", 0.9631262),
+        ("far", 0.3445741),
+        ("csi", 0.432855),
+        ("gss", None),
+        ("hk", 0.5235019),
+        ("hss", None),
+        ("odds", 33.2937647),
+        ("lodds", 3.5053691),
+        ("bagss", None),
+        ("eclv", None),
+        ("ctc_total", None),
+        ("cts_total", None),
+        ("ctc_fn_on", 44984491.0),
+        ("ctc_fn_oy", 2570049.0),
+        ("ctc_fy_on", 1722257.0),
+        ("ctc_fy_oy", 3275963.0),
+        ("ctc_oy", 5846012.0),
+        ("ctc_on", 46706748.0),
+        ("ctc_fy", 46706748.0),
+        ("ctc_fn", 47554540.0),
+        ("odds1", 33.2937647),
+        ("orss", 0.9416803),
+        ("sedi", 0.7397156),
+        ("seds", None),
+        ("edi", 0.7014241),
+        ("eds", None),
+    ]
+)
+def test_ctc_statistics(func_name, expected):
+    func_str = f"calculate_{func_name}"
+    func = getattr(ctc, func_str)
+    actual = func(*ctc_data())
+    assert actual == expected
+
+    # Check None is returned on Exception
+    with patch.object(ctc, "sum_column_data_by_name", side_effect=TypeError):
+        actual = func(*ctc_data())
+    assert actual == None
+
 
 
 if __name__ == "__main__":

test/test_mode_2d_arearat_statistics.py

@@ -0,0 +1,70 @@
+import pytest
+from unittest.mock import patch
+import numpy as np
+import metcalcpy.util.mode_2d_arearat_statistics as m2as
+
+column_names = np.array(
+    ["object_type", "area", "fcst_flag", "simple_flag", "matched_flag"]
+)
+
+data = np.array(
+    [
+        ["2d", 100, 1, 1, 1],
+        ["2d", 30, 0, 1, 1],
+        ["2d", 120, 1, 1, 0],
+        ["2d", 12, 0, 1, 1],
+        ["2d", 1, 1, 0, 1],
+        ["2d", 17, 0, 1, 1],
+        ["2d", 200, 1, 1, 1],
+        ["3d", 66, 0, 1, 0],
+    ]
+)
+
+@pytest.mark.parametrize(
+    "func_name,data_values,expected",
+    [
+        ("arearat_fsa_asa", data, 0.8768267),
+        ("arearat_osa_asa", data, 0.1231733),
+        ("arearat_asm_asa", data, 0.7494781),
+        ("arearat_asu_asa", data, 0.2505219),
+        ("arearat_fsm_fsa", data, 0.7142857),
+        ("arearat_fsu_fsa", data, 0.2857143),
+        ("arearat_osm_osa", data, 1.0),
+        ("arearat_osu_osa", data, None),
+        ("arearat_fsm_asm", data, 0.2857143),
+        ("arearat_osm_asm", data, 0.1643454),
+        ("arearat_osu_asu", data, None),
+        ("arearat_fsa_aaa", data, 0.875),
+        ("arearat_osa_aaa", data, 0.1229167),
+        ("arearat_fsa_faa", data, 0.9976247),
+        ("arearat_fca_faa", data, 0.0023753),
+        ("arearat_osa_oaa", data, 1.0),
+        ("arearat_oca_oaa", data, None),
+        ("arearat_fca_aca", data, 1.0),
+        ("arearat_oca_aca", data, None),
+        ("arearat_fsa_osa", data, 7.1186441),
+        ("arearat_osa_fsa", data, 1.0),
+        ("arearat_aca_asa", data, 0.0020877),
+        ("arearat_asa_aca", data, 479.0),
+        ("arearat_fca_fsa", data, 0.002381),
+        ("arearat_fsa_fca", data, 420.0),
+        ("arearat_oca_osa", data, None),
+        ("arearat_osa_oca", data, None),
+        ("objahits", data, 179.5),
+        ("objamisses", data, None),
+        ("objafas", data, 120.0),
+        ("objacsi", data, 0.5993322),
+        ("objapody", data, None),
+        ("objafar", data, 0.1431981),
+            ],
+)
+def test_calculate_3d_ratio(func_name, data_values, expected):
+    func_str = f"calculate_2d_{func_name}"
+    func = getattr(m2as, func_str)
+    actual = func(data_values, column_names)
+    assert actual == expected
+
+    # Check None is returned on Exception
+    with patch.object(m2as, "column_data_by_name_value", side_effect=TypeError):
+        actual = func(data_values, column_names)
+    assert actual == None

test/test_mode_3d_ratio_statistics.py

@@ -0,0 +1,69 @@
+import pytest
+from unittest.mock import patch
+import numpy as np
+import metcalcpy.util.mode_3d_ratio_statistics as m3rs
+
+column_names = np.array(
+    ["object_type", "volume", "fcst_flag", "simple_flag", "matched_flag"]
+)
+
+data = np.array(
+    [
+        ["3d", 100, 1, 1, 1],
+        ["3d", 30, 0, 1, 1],
+        ["3d", 120, 1, 1, 0],
+        ["3d", 12, 0, 1, 1],
+        ["3d", 1, 1, 0, 1],
+        ["3d", 17, 0, 1, 1],
+        ["2d", 200, 1, 1, 1],
+        ["3d", 66, 0, 1, 0],
+    ]
+)
+
+@pytest.mark.parametrize(
+    "func_name,data_values,expected",
+    [
+        ("ratio_fsa_asa", data, 0.3333333),
+        ("ratio_osa_asa", data, 0.6666667),
+        ("ratio_asm_asa", data, 0.6666667),
+        ("ratio_asu_asa", data, 0.3333333),
+        ("ratio_fsm_fsa", data, 1.0),
+        ("ratio_fsu_fsa", data, 0.5),
+        ("ratio_osm_osa", data, 0.75),
+        ("ratio_osu_osa", data, 0.25),
+        ("ratio_fsm_asm", data, 0.25),
+        ("ratio_osm_asm", data, 0.25),
+        ("ratio_fsu_asu", data, 0.5),
+        ("ratio_osu_asu", data, 0.5),
+        ("ratio_fsa_aaa", data, 0.2857143),
+        ("ratio_osa_aaa", data, 0.5714286),
+        ("ratio_fsa_faa", data, 0.6666667),
+        ("ratio_fca_faa", data, 0.3333333),
+        ("ratio_osa_oaa", data, 1.0),
+        ("ratio_oca_oaa", data, 0.0),
+        ("ratio_fca_aca", data, 1.0),
+        ("ratio_fsa_osa", data, 0.5),
+        ("ratio_osa_fsa", data, 2.0),
+        ("ratio_asa_aca", data, 6.0),
+        ("ratio_fca_fsa", data, 0.5),
+        ("ratio_fsa_fca", data, 2.0),
+        ("ratio_oca_osa", data, 0.0),
+        ("ratio_osa_oca", data, None),
+        ("objhits", data, 2.0),
+        ("objmisses", data, 1.0),
+        ("objfas", data, 1.0),
+        ("objcsi", data, 0.5),
+        ("objpody", data, 0.6666667),
+        ("objfar", data, 0.3333333),
+    ],
+)
+def test_calculate_3d_ratio(func_name, data_values, expected):
+    func_str = f"calculate_3d_{func_name}"
+    func = getattr(m3rs, func_str)
+    actual = func(data_values, column_names)
+    assert actual == expected
+
+    # Check None is returned on Exception
+    with patch.object(m3rs, "column_data_by_name_value", side_effect=TypeError):
+        actual = func(data_values, column_names)
+    assert actual == None