add more exceptions (close #263)

gnames · Jun 5, 2024 · 26615eb · 26615eb
1 parent 42f35e0
commit 26615eb
Show file tree

Hide file tree

Showing 5 changed files with 120 additions and 30 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -2,6 +2,11 @@
 
 ## Unreleased
 
+## [v1.10.1] - 2024-06-05 Wed
+
+- Add [#263]: add more exceptions with speciific epithets like
+              "complex", "do", "spec".
+
 ## [v1.10.0] - 2024-06-04 Tue
 
 - Add [#260]: add `candidatus` field for parsed data.

diff --git a/ent/internal/preprocess/preprocess.go b/ent/internal/preprocess/preprocess.go
@@ -39,35 +39,69 @@ var VirusException = map[string]string{
 }
 
 var AmbiguousException = map[string][]string{
-	"Aeolesthes":     {"mihi"},
-	"Agnetina":       {"den"},
-	"Anisochaeta":    {"mihi"},
-	"Antaplaga":      {"dela"},
-	"Baeolidia":      {"dela"},
-	"Bolitoglossa":   {"la"},
-	"Campylosphaera": {"dela"},
-	"Desmoxytes":     {"des"},
-	"Dicentria":      {"dela"},
-	"Eucyclops":      {"mihi"},
-	"Eulaira":        {"dela"},
-	"Gnathopleustes": {"den"},
-	"Gobiosoma":      {"spec"},
-	"Helophorus":     {"ser"},
-	"Lampona":        {"spec"},
-	"Leptonetela":    {"la"},
-	"Malamatidia":    {"zu"},
-	"Meteorus":       {"dos"},
-	"Nocaracris":     {"van"},
-	"Paralvinella":   {"dela"},
-	"Ruteloryctes":   {"bis"},
-	"Scoparia":       {"dela"},
-	"Selenops":       {"ab"},
-	"Semiothisa":     {"da"},
-	"Serina":         {"ser", "subser"},
-	"Stenoecia":      {"dos"},
-	"Sympycnus":      {"du"},
-	"Tortolena":      {"dela"},
-	"Zodarion":       {"van"},
+	"Aeolesthes":        {"mihi"},
+	"Agnetina":          {"den"},
+	"Aleuroclava":       {"complex"},
+	"Allawrencius":      {"complex"},
+	"Anisochaeta":       {"mihi"},
+	"Antaplaga":         {"dela"},
+	"Baeolidia":         {"dela"},
+	"Bolbodeomyia":      {"complex"},
+	"Bolitoglossa":      {"la"},
+	"Campylosphaera":    {"dela"},
+	"Castelnaudia":      {"spec"},
+	"Cicada":            {"complex"},
+	"Desmoxytes":        {"des"},
+	"Dicentria":         {"dela"},
+	"Dichostasia":       {"complex"},
+	"Dimorphoceras":     {"complex"},
+	"Dischidia":         {"complex"},
+	"Ecnomus":           {"complex"},
+	"Eucyclops":         {"mihi"},
+	"Eulaira":           {"dela"},
+	"Fusinus":           {"complex"},
+	"Gnathopleustes":    {"den"},
+	"Gobiosoma":         {"spec"},
+	"Gonatobotrys":      {"complex"},
+	"Heizmannia":        {"complex"},
+	"Helophorus":        {"ser"},
+	"Hemicloeina":       {"spec"},
+	"Lampona":           {"spec"},
+	"Leptonetela":       {"la"},
+	"Libystica":         {"complex"},
+	"Malamatidia":       {"zu"},
+	"Meteorus":          {"dos"},
+	"Nocaracris":        {"van"},
+	"Notozomus":         {"spec"},
+	"Ochodaeus":         {"complex"},
+	"Odontella":         {"do"},
+	"Oecetis":           {"complex"},
+	"Oedipina":          {"complex"},
+	"Oedipus":           {"complex"},
+	"Oedopinola":        {"complex"},
+	"Paradimorphoceras": {"complex"},
+	"Paralvinella":      {"dela"},
+	"Parentia":          {"do"},
+	"Phyllospongia":     {"complex"},
+	"Plectrocnemia":     {"complex"},
+	"Rubus":             {"complex"},
+	"Ruteloryctes":      {"bis"},
+	"Sceliphron":        {"complex"},
+	"Scopaeus":          {"complex"},
+	"Scoparia":          {"dela"},
+	"Selenops":          {"ab"},
+	"Semiothisa":        {"da"},
+	"Serina":            {"ser", "subser"},
+	"Sigipinius":        {"complex"},
+	"Stegosoladidus":    {"complex"},
+	"Stenoecia":         {"dos"},
+	"Sympycnus":         {"du"},
+	"Tetracis":          {"complex"},
+	"Tetramorium":       {"do"},
+	"Tortolena":         {"dela"},
+	"Trichosternus":     {"spec"},
+	"Trisephena":        {"complex"},
+	"Zodarion":          {"van"},
 }
 
 var NoParseException = map[string]string{

diff --git a/gnparser_test.go b/gnparser_test.go
@@ -149,6 +149,20 @@ func TestOutputRestore(t *testing.T) {
 	assert.Equal(t, "sapiens", sp.Species.Species)
 }
 
+func TestExceptions(t *testing.T) {
+	assert := assert.New(t)
+	cfg := gnparser.NewConfig()
+	gnp := gnparser.New(cfg)
+	f, err := os.Open("testdata/exceptions.txt")
+	assert.Nil(err)
+	scanner := bufio.NewScanner(f)
+	for scanner.Scan() {
+		name := scanner.Text()
+		parsed := gnp.ParseName(name)
+		assert.True(parsed.ParseQuality == 1, name)
+	}
+}
+
 func getTestData(t *testing.T, filename string) []testData {
 	var res []testData
 	path := filepath.Join("testdata", filename)

diff --git a/testdata/exceptions.txt b/testdata/exceptions.txt
@@ -0,0 +1,37 @@
+Aleuroclava complex Singh, 1931
+Allawrencius complex Lawrence, 1953
+Bolbodeomyia complex Theobald, 1910
+Castelnaudia spec (Darlington, 1962)
+Cicada complex Walker, 1850
+Dichostasia complex Yochelson, 1956
+Dimorphoceras complex (Moore, 1939)
+Dischidia complex Griff.
+Ecnomus complex Mosely, 1932
+Fusinus complex M. A. Snyder, 2000
+Fusinus pauciliratus complex M. A. Snyder, 2000
+Gonatobotrys complex Jane Walker & Minter
+Heizmannia (Heizmannia) complex (Theobald, 1910)
+Hemicloeina spec Platnick, 2002
+Libystica complex Holland, 1894
+Notozomus spec (Harvey, 1992)
+Ochodaeus complex LeConte, 1868
+Odontella do J Najt, & WM Weiner
+Oecetis complex Hwang, 1957
+Oedipina complex (Dunn, 1924)
+Oedipus complex Dunn, 1924
+Oedopinola complex (Dunn, 1924)
+Paradimorphoceras complex (Moore, 1939)
+Parentia do Bickel, 2002
+Phyllospongia complex de Laubenfels, 1954
+Plectrocnemia complex Hwang, 1958
+Rubus complex L. H. Bailey
+Sceliphron complex Kohl, 1918
+Sceliphron fossuliferum complex Kohl, 1918
+Scopaeus (Scopaeus) complex Sharp, 1874
+Scopaeus complex Sharp, 1874
+Sigipinius complex Golovatch, 2013
+Stegosoladidus complex Berge, 2001
+Tetracis complex Sharp, 1874
+Tetramorium do Forel, 1914
+Trichosternus spec Darlington, 1962
+Trisephena complex Medler, 1990
diff --git a/version.go b/version.go
@@ -3,7 +3,7 @@ package gnparser
 var (
 	// Version is the version of the gnparser package. When Makefile is
 	// used, the version is calculated out of Git tags.
-	Version = "v1.10.0"
+	Version = "v1.10.1"
 	// Build is a timestamp of when Makefile was used to compile
 	// the gnparser code. If go build was used, Build stays empty.
 	Build string