normalize stemmed canonical Aus bus bus to Aus bus (close #255)

CHANGELOG.md

@@ -4,6 +4,10 @@
 
 - Add: sorting uses `slices` package.
 
+## [v1.8.0] - 2023-10-11 Wed
+
+- Add [#255]: normalize stemmed canonical of `Aus bus bus` to `Aus bus`.
+
 ## [v1.7.5] - 2023-09-26 Tue
 
 - Add: CSV and TSV files provide now verbatim authorship instead of normalized

ent/parser/name.go

@@ -45,7 +45,7 @@ func (sn *scientificNameNode) Canonical() *parsed.Canonical {
 	}
 	c := sn.canonical()
 	return &parsed.Canonical{
-		Stemmed: stemmer.StemCanonical(c.Value),
+		Stemmed: stemmer.StemCanonical(c.Value, sn.cardinality),
 		Simple:  c.Value,
 		Full:    c.ValueRanked,
 	}

ent/stemmer/stemmer.go

@@ -130,7 +130,10 @@ type StemmedWord struct {
 //  3. All characters in the string are ASCII with exception of the
 //     hybrid sign.
 //  4. The string always starts with a capitalized word.
-func StemCanonical(c string) string {
+func StemCanonical(c string, card int) string {
+	if card == 3 {
+		c = normalizeSpGroup(c, card)
+	}
 	graftChimeraFormulaParts := strings.Split(c, " + ")
 	for gci, gcv := range graftChimeraFormulaParts {
 		hybridFormulaParts := strings.Split(gcv, " × ")
@@ -159,6 +162,19 @@ func StemCanonical(c string) string {
 	return str.TransliterateDiaereses(strings.Join(graftChimeraFormulaParts, " + "))
 }
 
+func normalizeSpGroup(c string, card int) string {
+	if card != 3 {
+		return c
+	}
+
+	es := strings.Split(c, " ")
+	if len(es) != 3 || es[1] != es[2] {
+		return c
+	}
+
+	return es[0] + " " + es[1]
+}
+
 // Stem takes a word and, assuming the word is noun, removes its latin suffix
 // if such suffix is detected.
 func Stem(wrd string) StemmedWord {

ent/stemmer/stemmer_test.go

@@ -26,20 +26,23 @@ func TestStemmer(t *testing.T) {
 
 	t.Run("StemCanonical", func(t *testing.T) {
 		data := []struct {
-			msg string
-			in  string
-			out string
+			msg  string
+			in   string
+			out  string
+			card int
 		}{
-			{"Uninomial", "Pomatomus", "Pomatomus"},
-			{"Binomial1", "Betula naturae", "Betula natur"},
-			{"Binomial2", "Betula alba", "Betula alb"},
-			{"Binomial3", "Leptochloöpsis virgata", "Leptochloopsis uirgat"},
-			{"Trinomial", "Betula alba naturae", "Betula alb natur"},
-			{"GraftChimeraFormula", "Crataegus + Mespilus", "Crataegus + Mespilus"},
-			{"GraftChimeraFormula2", "Cytisus purpureus + Laburnum anagyroides", "Cytisus purpure + Laburnum anagyroid"},
+			{"Uninomial", "Pomatomus", "Pomatomus", 1},
+			{"Binomial1", "Betula naturae", "Betula natur", 2},
+			{"Binomial2", "Betula alba", "Betula alb", 2},
+			{"Binomial3", "Leptochloöpsis virgata", "Leptochloopsis uirgat", 2},
+			{"Trinomial", "Betula alba naturae", "Betula alb natur", 3},
+			{"SpGroup", "Betula alba alba", "Betula alb", 3},
+			{"SpGroup", "Betula alba albus", "Betula alb alb", 3},
+			{"GraftChimeraFormula", "Crataegus + Mespilus", "Crataegus + Mespilus", 0},
+			{"GraftChimeraFormula2", "Cytisus purpureus + Laburnum anagyroides", "Cytisus purpure + Laburnum anagyroid", 0},
 		}
 		for _, v := range data {
-			assert.Equal(t, v.out, stemmer.StemCanonical(v.in), v.msg)
+			assert.Equal(t, v.out, stemmer.StemCanonical(v.in, v.card), v.msg)
 		}
 	})
 }