From dfc114a51b9d675c1a08852ecc02a2d7dad7fbf7 Mon Sep 17 00:00:00 2001
From: janezd <janez.demsar@fri.uni-lj.si>
Date: Fri, 22 Dec 2023 19:55:42 +0100
Subject: [PATCH] Network File: Fix crash when two variables are equally
 preferrable for label

---
 orangecontrib/network/widgets/OWNxFile.py     |  7 +++++--
 .../network/widgets/tests/test_OWNxFile.py    | 19 +++++++++++++++++++
 2 files changed, 24 insertions(+), 2 deletions(-)

diff --git a/orangecontrib/network/widgets/OWNxFile.py b/orangecontrib/network/widgets/OWNxFile.py
index 5d85ebe..8e82294 100755
--- a/orangecontrib/network/widgets/OWNxFile.py
+++ b/orangecontrib/network/widgets/OWNxFile.py
@@ -1,3 +1,4 @@
+from operator import itemgetter
 from os import path
 from itertools import product
 from traceback import format_exception_only
@@ -243,13 +244,15 @@ def _vars_for_label(self, data: Table):
             values= data.get_column(var)
             values = values[values != ""]
             set_values = set(values)
+            # values have to be unique, and have to include all labels
             if len(values) != len(set_values) \
                     or not original_nodes <= set_values:
                 continue
-            vars_and_overs.append((len(set_values - original_nodes), var))
+            vars_and_overs.append((len(set_values), var))
         if not vars_and_overs:
             return None, []
-        _, best_var = min(vars_and_overs)
+        # Prefer variables with less extra values
+        _, best_var = min(vars_and_overs, key=itemgetter(0))
         useful_string_vars = [var for _, var in vars_and_overs]
         return best_var, useful_string_vars
 
diff --git a/orangecontrib/network/widgets/tests/test_OWNxFile.py b/orangecontrib/network/widgets/tests/test_OWNxFile.py
index 236da3b..9783691 100644
--- a/orangecontrib/network/widgets/tests/test_OWNxFile.py
+++ b/orangecontrib/network/widgets/tests/test_OWNxFile.py
@@ -54,6 +54,25 @@ def test_vars_for_label(self):
         self.assertIs(best_var, domain["label"])
         self.assertEqual(useful_vars, [domain["with_extras"], domain["label"]])
 
+        data = Orange.data.Table.from_list(
+            Orange.data.Domain(
+                [], None, [Orange.data.StringVariable(x) for x in "abcde"]),
+            [["aa", "", "cc", "aa", ""],
+             ["bb", "bb", "cc", "bb", "aa"],
+             ["cc", "", "aa", "cc", "bb"],
+             ["dd", "aa", "bb", "dd", "cc"],
+             ["ee", "cc", "dd", "ee", ""],
+             ["ff", "ee", "ee", "ff", "dd"],
+             ["gg", "dd", "ff", "", "ee"],
+             ["hh", "ff", "gg", "", "ff"],
+             ["ii", "gg", "", "", "gg"]]
+        )
+        domain = data.domain
+        best_var, useful_vars = self.widget._vars_for_label(data)
+        self.assertIs(best_var, domain["b"])
+        # c is not unique and d doesn't cover all values
+        self.assertEqual(useful_vars, [domain["a"], domain["b"], domain["e"]])
+
     def test_label_combo_contents(self):
         widget = self.widget
         widget.read_auto_data = Mock()