Add error testing

Author: Jacobluke-

dabest/_dabest_object.py

@@ -471,7 +471,9 @@ def _check_errors(self, x, y, idx, experiment, experiment_label, x1_level):
 
             # Check if all types are valid
             if not valid_types:
-                raise TypeError("Invalid type found in idx. Expected str, tuple, or list.")
+                err0 = "`mini_meta` is True, but `idx` ({})".format(idx)
+                err1 = "does not contain exactly 2 unique columns."
+                raise ValueError(err0 + err1)
 
             # Handling str type condition
             if is_str_condition_met:

dabest/plot_tools.py

@@ -494,7 +494,7 @@ def single_sankey(
                 myD["top"] = 1
             else:
                 myD["right"] -= 0.02
-                myD["bottom"] = rightWidths_norm[right_labelsss[i - 1]]["top"] + 0.02
+                myD["bottom"] = rightWidths_norm[right_labels[i - 1]]["top"] + 0.02
                 myD["top"] = myD["bottom"] + myD["right"]
                 topEdge = myD["top"]
         else:

nbs/API/dabest_object.ipynb

@@ -541,7 +541,9 @@
     "\n",
     "            # Check if all types are valid\n",
     "            if not valid_types:\n",
-    "                raise TypeError(\"Invalid type found in idx. Expected str, tuple, or list.\")\n",
+    "                err0 = \"`mini_meta` is True, but `idx` ({})\".format(idx)\n",
+    "                err1 = \"does not contain exactly 2 unique columns.\"\n",
+    "                raise ValueError(err0 + err1)\n",
     "\n",
     "            # Handling str type condition\n",
     "            if is_str_condition_met:\n",

nbs/API/plot_tools.ipynb

@@ -549,7 +549,7 @@
     "                myD[\"top\"] = 1\n",
     "            else:\n",
     "                myD[\"right\"] -= 0.02\n",
-    "                myD[\"bottom\"] = rightWidths_norm[right_labelsss[i - 1]][\"top\"] + 0.02\n",
+    "                myD[\"bottom\"] = rightWidths_norm[right_labels[i - 1]][\"top\"] + 0.02\n",
     "                myD[\"top\"] = myD[\"bottom\"] + myD[\"right\"]\n",
     "                topEdge = myD[\"top\"]\n",
     "        else:\n",

nbs/tests/data/mocked_data_test_load_errors.py

@@ -5,5 +5,13 @@
 random.seed(88888)
 N = 10
 c1 = sp.stats.norm.rvs(loc=100, scale=5, size=N)
-t1 = sp.stats.norm.rvs(loc=115, scale=5, size=N)
-dummy_df = pd.DataFrame({"Control 1": c1, "Test 1": t1})
+c2 = sp.stats.norm.rvs(loc=115, scale=5, size=N)
+c3 = sp.stats.norm.rvs(loc=3.25, scale=0.4, size=N)
+
+t1 = sp.stats.norm.rvs(loc=3.5, scale=0.5, size=N)
+t2 = sp.stats.norm.rvs(loc=2.5, scale=0.6, size=N)
+id_col = pd.Series(range(1, N+1))
+dummy_df = pd.DataFrame({'Control 1' : c1,     'Test 1' : t1,
+                   'Control 2' : c2,     'Test 2' : t2,
+                   'Control 3' : c3,     'ID'  : id_col
+                  })

nbs/tests/test_load_errors.py

@@ -19,7 +19,7 @@ def test_wrong_params_combinations():
         my_data = load(
             dummy_df,
             idx=("Control 1", "Test 1"),
-            delta2=True,
+            delta2=True
         )
     assert error_msg in str(excinfo.value)
 
@@ -30,7 +30,7 @@ def test_wrong_params_combinations():
             x=["Control 1", "Control 1"],
             y="Test 1",
             delta2=True,
-            mini_meta=True,
+            mini_meta=True
         )
 
     assert error_msg in str(excinfo.value)
@@ -42,7 +42,7 @@ def test_wrong_params_combinations():
             x=["Control 1", "Control 1"],
             y="Test 1",
             delta2=True,
-            proportional=True,
+            proportional=True
         )
 
     assert error_msg in str(excinfo.value)
@@ -53,7 +53,164 @@ def test_wrong_params_combinations():
             dummy_df,
             x=["Control 1", "Control 1"],
             idx=("Control 1", "Test 1"),
+            delta2=True
+        )
+
+    assert error_msg in str(excinfo.value)
+
+    error_msg = "`id_col` must be specified if `paired` is assigned with a not NoneType value."
+    with pytest.raises(IndexError) as excinfo:
+        my_data = load(
+            dummy_df, idx=("Control 1", "Test 1"), paired="baseline"
+        )
+
+    assert error_msg in str(excinfo.value)
+
+    error_msg = "`delta2` is True but `y` is not indicated."
+    with pytest.raises(ValueError) as excinfo:
+        my_data = load(
+            dummy_df,
+            x=["Control 1", "Control 1"],
+            delta2=True
+        )
+
+
+def test_param_validations():
+    error_msg = "`idx` contains duplicated groups. Please remove any duplicates and try again.".format(N)
+    with pytest.raises(ValueError) as excinfo:
+        my_data = load(
+            dummy_df, idx=("Control 1", "Control 1")
+        )
+
+    assert error_msg in str(excinfo.value)
+
+    err0 = "Groups are repeated across tuples,"
+    err1 = " or a tuple has repeated groups in it."
+    err2 = " Please remove any duplicates and try again."
+    error_msg = err0 + err1 + err2
+    with pytest.raises(ValueError) as excinfo:
+        my_data = load(
+            dummy_df, idx=(("Control 1", "Control 1", "Test 1"), ("Control 2", "Test 2"))
+        )
+
+    assert error_msg in str(excinfo.value)
+
+    wrong_idx = ("Control 1", ("Control 1", "Test 1"))
+    error_msg = "There seems to be a problem with the idx you " "entered--{}.".format(wrong_idx)
+    with pytest.raises(ValueError) as excinfo:
+        my_data = load(
+            dummy_df, idx=wrong_idx
+        )
+
+    assert error_msg in str(excinfo.value)
+
+    wrong_paired = 'not_valid'
+    error_msg = "{} assigned for `paired` is not valid.".format(wrong_paired)
+    with pytest.raises(ValueError) as excinfo:
+        my_data = load(
+            dummy_df, idx=("Control 1", "Test 1"), paired=wrong_paired, id_col="ID"
+        )
+
+    assert error_msg in str(excinfo.value)
+
+
+    wrong_id_col = 'not_valid'
+    error_msg = "{} is not a column in `data`. ".format(wrong_id_col)
+    with pytest.raises(IndexError) as excinfo:
+        my_data = load(
+            dummy_df, idx=("Control 1", "Test 1"), paired="baseline", id_col=wrong_id_col
+        )
+
+    assert error_msg in str(excinfo.value)
+
+    wrong_idx_mmeta = ("Control 1", "Test 1", "Test 2")
+    err0 = "`mini_meta` is True, but `idx` ({})".format(wrong_idx_mmeta)
+    err1 = "does not contain exactly 2 unique columns."
+    error_msg = err0 + err1
+    with pytest.raises(ValueError) as excinfo:
+        my_data = load(
+            dummy_df, idx=wrong_idx_mmeta, mini_meta=True
+        )
+
+    assert error_msg in str(excinfo.value)
+
+    wrong_idx_mmeta = (("Control 1", "Test 1", "Test 2"), ("Control 1", "Control 2", "Test 3"))
+    err0 = "`mini_meta` is True, but `idx` ({})".format(wrong_idx_mmeta)
+    err1 = "does not contain exactly 2 unique columns."
+    error_msg = err0 + err1
+    with pytest.raises(ValueError) as excinfo:
+        my_data = load(
+            dummy_df, idx=wrong_idx_mmeta, mini_meta=True
+        )
+
+    assert error_msg in str(excinfo.value)
+
+    wrong_x = ["Control 1", "Control 1", "Control 2"]
+    error_msg = "`delta2` is True but the number of variables indicated by `x` is {}.".format(len(wrong_x))
+    with pytest.raises(ValueError) as excinfo:
+        my_data = load(
+            dummy_df, x=wrong_x, y="Test 1", delta2=True
+        )
+
+    assert error_msg in str(excinfo.value)
+
+    wrong_x = ["Control 4", "Control 5"]
+    error_msg = "is not a column in `data`. Please check."
+    with pytest.raises(IndexError) as excinfo:
+        my_data = load(
+            dummy_df, x=wrong_x, y="Test 1", delta2=True
+        )
+
+    assert error_msg in str(excinfo.value)
+
+    wrong_y = "Test 3"
+    error_msg = "is not a column in `data`. Please check."
+    with pytest.raises(IndexError) as excinfo:
+        my_data = load(
+            dummy_df, x=["Control 1", "Control 2"], y=wrong_y, delta2=True
+        )
+
+    assert error_msg in str(excinfo.value)
+
+    wrong_experiment = "not_valid"
+    error_msg = "is not a column in `data`. Please check."
+    with pytest.raises(IndexError) as excinfo:
+        my_data = load(
+            dummy_df,
+            x=["Control 1", "Control 1"],
+            y="Test 1",
+            delta2=True,
+            experiment=wrong_experiment
+        )
+
+    assert error_msg in str(excinfo.value)
+
+    #TODO experiment and experiment_label are different
+
+    wrong_experiment_label = ["A", "B", "C"]
+    error_msg = "`experiment_label` does not have a length of 2."
+    with pytest.raises(ValueError) as excinfo:
+        my_data = load(
+            dummy_df,
+            x=["Control 1", "Control 1"],
+            y="Test 1",
             delta2=True,
+            experiment="Control 1",
+            experiment_label=wrong_experiment_label
         )
 
     assert error_msg in str(excinfo.value)
+
+    wrong_x1_level = "not_valid"
+    error_msg = "`x1_level` does not have a length of 2."
+    with pytest.raises(ValueError) as excinfo:
+        my_data = load(
+            dummy_df,
+            x=["Control 1", "Control 1"],
+            y="Test 1",
+            delta2=True,
+            experiment="Control 1",
+            x1_level=wrong_x1_level
+        )
+
+    assert error_msg in str(excinfo.value)

nbs/tests/test_plot_tools.py

@@ -1,5 +1,6 @@
 import pytest
-from dabest.plot_tools import get_swarm_spans, width_determine
+import pandas as pd
+from dabest.plot_tools import get_swarm_spans, width_determine, error_bar, check_data_matches_labels
 import numpy as np
 
 
@@ -23,3 +24,55 @@ def test_width_determine():
         width_determine("some_labels", None)
 
     assert error_msg in str(excinfo.value)
+
+
+def test_error_bar():
+    data = pd.DataFrame({
+        'group': ['A', 'A', 'B', 'B'],
+        'value': [1, 2, 3, 4]
+    })
+    error_msg = "`gap_width_percent` must be between 0 and 100."
+    with pytest.raises(ValueError) as excinfo:
+        error_bar(
+        data=data,
+        x='group',
+        y='value',
+        type='mean_sd',
+        gap_width_percent=-10  # Invalid as it's less than 0
+    )
+
+    assert error_msg in str(excinfo.value)
+
+    error_msg = "Invalid `method`. Must be one of 'gapped_lines', \
+                         'proportional_error_bar', or 'sankey_error_bar'."
+    with pytest.raises(ValueError) as excinfo:
+        error_bar(
+        data=data,
+        x='group',
+        y='value',
+        type='mean_sd',
+        method='invalid_method'  # Invalid as it's not one of the accepted values
+    )
+
+    assert error_msg in str(excinfo.value)
+
+    error_msg = "Only accepted values for type are ['mean_sd', 'median_quartiles']"
+    with pytest.raises(ValueError) as excinfo:
+        error_bar(
+        data=data,
+        x='group',
+        y='value',
+        type='invalid_type'
+    )
+
+    assert error_msg in str(excinfo.value)
+
+
+def test_check_data_matches_labels():
+    wrong_labels = ['A', 'B', 'C']
+    wrong_data = pd.Series(['A', 'B', 'D'])
+    error_msg = "labels and data do not match."
+    with pytest.raises(Exception) as excinfo:
+        check_data_matches_labels(wrong_labels, wrong_data, side='left')
+
+    assert error_msg in str(excinfo.value)