From f23e4d396640406d81f3cc07f0d433c0986d15d9 Mon Sep 17 00:00:00 2001
From: Luca Foppiano <Foppiano.Luca@nims.go.jp>
Date: Fri, 29 Mar 2024 07:37:34 +0900
Subject: [PATCH] add more tests

---
 .../engines/utilities/LabellingUtilsTest.kt   | 93 +++++++++----------
 1 file changed, 44 insertions(+), 49 deletions(-)

diff --git a/src/test/kotlin/org/grobid/core/engines/utilities/LabellingUtilsTest.kt b/src/test/kotlin/org/grobid/core/engines/utilities/LabellingUtilsTest.kt
index 2d363b5e..b5fab41e 100644
--- a/src/test/kotlin/org/grobid/core/engines/utilities/LabellingUtilsTest.kt
+++ b/src/test/kotlin/org/grobid/core/engines/utilities/LabellingUtilsTest.kt
@@ -96,66 +96,61 @@ class LabellingUtilsTest {
     }
 
     @Test
-    fun testCorrectingRangeValues() {
+    fun testCorrectingRangeValues1() {
         val input = "70\t70\t7\t70\t70\t70\t0\t70\t70\t70\tNOCAPS\tALLDIGIT\t0\tNOPUNCT\tdd\td\t0\t0\tI-<valueBase>\n" +
             "±\t±\t±\t±\t±\t±\t±\t±\t±\t±\tALLCAPS\tNODIGIT\t1\tNOPUNCT\t±\t±\t0\t0\t<other>\n" +
             "9\t9\t9\t9\t9\t9\t9\t9\t9\t9\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
-            "kg\tkg\tk\tkg\tkg\tkg\tg\tkg\tkg\tkg\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txx\tx\t1\t0\t<unitLeft>\n" +
-            ",\t,\t,\t,\t,\t,\t,\t,\t,\t,\tALLCAPS\tNODIGIT\t1\tCOMMA\t,\t,\t0\t0\t<other>\n" +
-            "15\t15\t1\t15\t15\t15\t5\t15\t15\t15\tNOCAPS\tALLDIGIT\t0\tNOPUNCT\tdd\td\t0\t0\tI-<valueBase>\n" +
-            "±\t±\t±\t±\t±\t±\t±\t±\t±\t±\tALLCAPS\tNODIGIT\t1\tNOPUNCT\t±\t±\t0\t0\t<other>\n" +
-            "5\t5\t5\t5\t5\t5\t5\t5\t5\t5\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
-            "%\t%\t%\t%\t%\t%\t%\t%\t%\t%\tALLCAPS\tNODIGIT\t1\tNOPUNCT\t%\t%\t0\t0\t<unitLeft>\n" +
-            "of\tof\to\tof\tof\tof\tf\tof\tof\tof\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txx\tx\t1\t0\t<other>\n" +
-            "fat\tfat\tf\tfa\tfat\tfat\tt\tat\tfat\tfat\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txxx\tx\t1\t0\t<other>\n" +
-            "mass\tmass\tm\tma\tmas\tmass\ts\tss\tass\tmass\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txxxx\tx\t1\t0\t<other>\n" +
-            ",\t,\t,\t,\t,\t,\t,\t,\t,\t,\tALLCAPS\tNODIGIT\t1\tCOMMA\t,\t,\t0\t0\t<other>\n" +
-            "VO\tvo\tV\tVO\tVO\tVO\tO\tVO\tVO\tVO\tALLCAPS\tNODIGIT\t0\tNOPUNCT\tXX\tX\t0\t0\t<other>\n" +
-            "2\t2\t2\t2\t2\t2\t2\t2\t2\t2\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<other>\n" +
-            "max\tmax\tm\tma\tmax\tmax\tx\tax\tmax\tmax\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txxx\tx\t0\t0\t<other>\n" +
-            ":\t:\t:\t:\t:\t:\t:\t:\t:\t:\tALLCAPS\tNODIGIT\t1\tPUNCT\t:\t:\t0\t0\t<other>\n" +
-            "50\t50\t5\t50\t50\t50\t0\t50\t50\t50\tNOCAPS\tALLDIGIT\t0\tNOPUNCT\tdd\td\t0\t0\tI-<valueBase>\n" +
+            "kg\tkg\tk\tkg\tkg\tkg\tg\tkg\tkg\tkg\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txx\tx\t1\t0\t<unitLeft>"
+
+        var output = LabellingUtils.correctLabelling(input)
+
+        assertEquals(output, input)
+    }
+
+    @Test
+    fun testCorrectingRangeValues2() {
+        val input = "70\t70\t7\t70\t70\t70\t0\t70\t70\t70\tNOCAPS\tALLDIGIT\t0\tNOPUNCT\tdd\td\t0\t0\tI-<valueBase>\n" +
             "±\t±\t±\t±\t±\t±\t±\t±\t±\t±\tALLCAPS\tNODIGIT\t1\tNOPUNCT\t±\t±\t0\t0\t<other>\n" +
-            "8\t8\t8\t8\t8\t8\t8\t8\t8\t8\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
-            "ml\tml\tm\tml\tml\tml\tl\tml\tml\tml\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txx\tx\t1\t0\t<unitLeft>\n" +
-            "•\t•\t•\t•\t•\t•\t•\t•\t•\t•\tALLCAPS\tNODIGIT\t1\tNOPUNCT\t•\t•\t0\t0\t<unitLeft>\n" +
+            "9\t9\t9\t9\t9\t9\t9\t9\t9\t9\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
+            "9\t9\t9\t9\t9\t9\t9\t9\t9\t9\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
+            "9\t9\t9\t9\t9\t9\t9\t9\t9\t9\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
             "kg\tkg\tk\tkg\tkg\tkg\tg\tkg\tkg\tkg\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txx\tx\t1\t0\t<unitLeft>\n" +
-            "-\t-\t-\t-\t-\t-\t-\t-\t-\t-\tALLCAPS\tNODIGIT\t1\tHYPHEN\t-\t-\t0\t0\t<unitLeft>\n" +
-            "1\t1\t1\t1\t1\t1\t1\t1\t1\t1\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<unitLeft>\n" +
-            "•\t•\t•\t•\t•\t•\t•\t•\t•\t•\tALLCAPS\tNODIGIT\t1\tNOPUNCT\t•\t•\t0\t0\t<unitLeft>\n" +
-            "min\tmin\tm\tmi\tmin\tmin\tn\tin\tmin\tmin\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txxx\tx\t1\t0\t<unitLeft>\n" +
-            "-\t-\t-\t-\t-\t-\t-\t-\t-\t-\tALLCAPS\tNODIGIT\t1\tHYPHEN\t-\t-\t0\t0\t<unitLeft>\n" +
-            "1\t1\t1\t1\t1\t1\t1\t1\t1\t1\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<unitLeft>\n" +
-            "and\tand\ta\tan\tand\tand\td\tnd\tand\tand\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txxx\tx\t0\t0\t<other>\n" +
-            "21\t21\t2\t21\t21\t21\t1\t21\t21\t21\tNOCAPS\tALLDIGIT\t0\tNOPUNCT\tdd\td\t0\t0\tI-<valueAtomic>\n" +
-            "of\tof\to\tof\tof\tof\tf\tof\tof\tof\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txx\tx\t1\t0\t<other>\n" +
-            "race\trace\tr\tra\trac\trace\te\tce\tace\trace\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txxxx\tx\t0\t0\t<other>\n" +
-            "A\ta\tA\tA\tA\tA\tA\tA\tA\tA\tALLCAPS\tNODIGIT\t1\tNOPUNCT\tX\tX\t1\t0\t<other>\n" +
-            "(\t(\t(\t(\t(\t(\t(\t(\t(\t(\tALLCAPS\tNODIGIT\t1\tOPENBRACKET\t(\t(\t0\t0\t<other>\n" +
-            "6\t6\t6\t6\t6\t6\t6\t6\t6\t6\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\tI-<valueAtomic>\n" +
-            "women\twomen\tw\two\twom\twome\tn\ten\tmen\tomen\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txxxx\tx\t0\t0\t<other>\n" +
-            "and\tand\ta\tan\tand\tand\td\tnd\tand\tand\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txxx\tx\t0\t0\t<other>\n" +
-            "15\t15\t1\t15\t15\t15\t5\t15\t15\t15\tNOCAPS\tALLDIGIT\t0\tNOPUNCT\tdd\td\t0\t0\tI-<valueAtomic>\n" +
-            "men\tmen\tm\tme\tmen\tmen\tn\ten\tmen\tmen\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txxx\tx\t0\t0\t<other>\n" +
-            ",\t,\t,\t,\t,\t,\t,\t,\t,\t,\tALLCAPS\tNODIGIT\t1\tCOMMA\t,\t,\t0\t0\t<other>\n" +
-            "40\t40\t4\t40\t40\t40\t0\t40\t40\t40\tNOCAPS\tALLDIGIT\t0\tNOPUNCT\tdd\td\t0\t0\tI-<valueBase>\n" +
-            "±\t±\t±\t±\t±\t±\t±\t±\t±\t±\tALLCAPS\tNODIGIT\t1\tNOPUNCT\t±\t±\t0\t0\t<other>\n" +
-            "7\t7\t7\t7\t7\t7\t7\t7\t7\t7\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
-            "years\tyears\ty\tye\tyea\tyear\ts\trs\tars\tears\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txxxx\tx\t1\t0\t<unitLeft>\n" +
-            ",\t,\t,\t,\t,\t,\t,\t,\t,\t,\tALLCAPS\tNODIGIT\t1\tCOMMA\t,\t,\t0\t0\t<other>\n" +
-            "176\t176\t1\t17\t176\t176\t6\t76\t176\t176\tNOCAPS\tALLDIGIT\t0\tNOPUNCT\tddd\td\t0\t0\tI-<valueBase>\n" +
+            ",\t,\t,\t,\t,\t,\t,\t,\t,\t,\tALLCAPS\tNODIGIT\t1\tCOMMA\t,\t,\t0\t0\t<other>"
+
+        var output = LabellingUtils.correctLabelling(input)
+
+        assertEquals(output, input)
+    }
+
+    @Test
+    fun testCorrectingRangeValues3() {
+        val input = "70\t70\t7\t70\t70\t70\t0\t70\t70\t70\tNOCAPS\tALLDIGIT\t0\tNOPUNCT\tdd\td\t0\t0\tI-<valueBase>\n" +
             "±\t±\t±\t±\t±\t±\t±\t±\t±\t±\tALLCAPS\tNODIGIT\t1\tNOPUNCT\t±\t±\t0\t0\t<other>\n" +
-            "7\t7\t7\t7\t7\t7\t7\t7\t7\t7\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
-            "cm\tcm\tc\tcm\tcm\tcm\tm\tcm\tcm\tcm\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txx\tx\t1\t0\t<unitLeft>\n" +
-            ",\t,\t,\t,\t,\t,\t,\t,\t,\t,\tALLCAPS\tNODIGIT\t1\tCOMMA\t,\t,\t0\t0\t<other>\n" +
-            "72\t72\t7\t72\t72\t72\t2\t72\t72\t72\tNOCAPS\tALLDIGIT\t0\tNOPUNCT\tdd\td\t0\t0\tI-<valueBase>\n" +
+            "9\t9\t9\t9\t9\t9\t9\t9\t9\t9\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
+            "9\t9\t9\t9\t9\t9\t9\t9\t9\t9\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
+            "9\t9\t9\t9\t9\t9\t9\t9\t9\t9\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
+            "kg\tkg\tk\tkg\tkg\tkg\tg\tkg\tkg\tkg\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txx\tx\t1\t0\t<unitLeft>\n" +
+            ",\t,\t,\t,\t,\t,\t,\t,\t,\t,\tALLCAPS\tNODIGIT\t1\tCOMMA\t,\t,\t0\t0\t<other>"
+
+        var output = LabellingUtils.correctLabelling(input)
+
+        assertEquals(output, input)
+    }
+
+    @Test
+    fun testCorrectingRangeValues4() {
+        val input = "70\t70\t7\t70\t70\t70\t0\t70\t70\t70\tNOCAPS\tALLDIGIT\t0\tNOPUNCT\tdd\td\t0\t0\tI-<valueBase>\n" +
             "±\t±\t±\t±\t±\t±\t±\t±\t±\t±\tALLCAPS\tNODIGIT\t1\tNOPUNCT\t±\t±\t0\t0\t<other>\n" +
-            "10\t10\t1\t10\t10\t10\t0\t10\t10\t10\tNOCAPS\tALLDIGIT\t0\tNOPUNCT\tdd\td\t0\t0\t<valueRange>\n" +
-            "kg\tkg\tk\tkg\tkg\tkg\tg\tkg\tkg\tkg\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txx\tx\t1\t0\t<unitLeft>"
+            "9\t9\t9\t9\t9\t9\t9\t9\t9\t9\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
+            "9\t9\t9\t9\t9\t9\t9\t9\t9\t9\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
+            "9\t9\t9\t9\t9\t9\t9\t9\t9\t9\tNOCAPS\tALLDIGIT\t1\tNOPUNCT\td\td\t0\t0\t<valueRange>\n" +
+            "kg\tkg\tk\tkg\tkg\tkg\tg\tkg\tkg\tkg\tNOCAPS\tNODIGIT\t0\tNOPUNCT\txx\tx\t1\t0\tI-<unitLeft>\n" +
+            ",\t,\t,\t,\t,\t,\t,\t,\t,\t,\tALLCAPS\tNODIGIT\t1\tCOMMA\t,\t,\t0\t0\t<other>"
 
         var output = LabellingUtils.correctLabelling(input)
 
         assertEquals(output, input)
     }
 
+
 }
\ No newline at end of file