From afe0c480afc76642753c307e08f31d474871a298 Mon Sep 17 00:00:00 2001
From: Ayman Bagabas <ayman.bagabas@gmail.com>
Date: Thu, 30 May 2024 15:56:59 -0400
Subject: [PATCH] Detect if grapheme is an emoji

This adds the necessary logic to detect if a grapheme cluster is an
emoji based on @rivo's [comment](https://github.com/rivo/uniseg/issues/27#issuecomment-1368204010)

Fixes: https://github.com/rivo/uniseg/issues/27
---
 grapheme.go      | 52 ++++++++++++++++++++++++++++++++++++++++++++++++
 grapheme_test.go | 22 ++++++++++++++++++++
 2 files changed, 74 insertions(+)

diff --git a/grapheme.go b/grapheme.go
index 1c17c27..cf94215 100644
--- a/grapheme.go
+++ b/grapheme.go
@@ -103,6 +103,11 @@ func (g *Graphemes) Bytes() []byte {
 	return []byte(g.cluster)
 }
 
+// IsEmoji returns true if the current grapheme cluster is an emoji.
+func (g *Graphemes) IsEmoji() bool {
+	return IsGraphemeClusterInStringEmoji(g.cluster, g.Width())
+}
+
 // Positions returns the interval of the current grapheme cluster as byte
 // positions into the original string. The first returned value "from" indexes
 // the first byte and the second returned value "to" indexes the first byte that
@@ -343,3 +348,50 @@ func FirstGraphemeClusterInString(str string, state int) (cluster, rest string,
 		}
 	}
 }
+
+const (
+	regionalIndicatorA = 0x1F1E6
+	regionalIndicatorZ = 0x1F1FF
+)
+
+// IsGraphemeClusterEmoji returns true if the given byte slice grapheme cluster
+// and width is an emoji according to the Unicode Standard Annex #51, Unicode
+// Emoji.
+func IsGraphemeClusterEmoji(cluster []byte, width int) bool {
+	return isGraphemeClusterEmoji(cluster, utf8.DecodeRune, width)
+}
+
+// IsGraphemeClusterInStringEmoji is like [IsGraphemeClusterEmoji] but its input
+// is a string.
+func IsGraphemeClusterInStringEmoji(cluster string, width int) bool {
+	return isGraphemeClusterEmoji(cluster, utf8.DecodeRuneInString, width)
+}
+
+func isGraphemeClusterEmoji[C []byte | string, F func(C) (rune, int)](cluster C, fn F, width int) bool {
+	if width != 2 {
+		return false
+	}
+
+	r, rw := fn(cluster)
+	if r == vs16 {
+		return true
+	}
+	if r >= regionalIndicatorA && r <= regionalIndicatorZ {
+		return true
+	}
+	if propertyGraphemes(r) == prExtendedPictographic &&
+		property(emojiPresentation, r) == prEmojiPresentation {
+		return true
+	}
+
+	cluster = cluster[rw:]
+	for len(cluster) > 0 {
+		r, rw := fn(cluster)
+		if r == vs16 {
+			return true
+		}
+		cluster = cluster[rw:]
+	}
+
+	return false
+}
diff --git a/grapheme_test.go b/grapheme_test.go
index 988aebe..aa4d065 100644
--- a/grapheme_test.go
+++ b/grapheme_test.go
@@ -510,6 +510,28 @@ func TestGraphemesFunctionString(t *testing.T) {
 	}
 }
 
+func TestIsGraphemeClusterEmoji(t *testing.T) {
+	testCases := []struct {
+		cluster  string
+		width    int
+		expected bool
+	}{
+		{"👋", 2, true},
+		{"a", 1, false},
+		{"咪", 2, false},
+		{"ض", 1, false},
+		{"🇩🇪", 2, true},
+		{"👨🏿‍🌾", 2, true},
+		{"🏳️‍🌈", 2, true},
+		{"☺️", 2, true}, // White smiling face (with variation selector 16 = emoji presentation)
+	}
+	for index, testCase := range testCases {
+		if result := IsGraphemeClusterEmoji([]byte(testCase.cluster), testCase.width); result != testCase.expected {
+			t.Errorf(`Test case %d %q failed: Expected %t, got %t`, index, testCase.cluster, testCase.expected, result)
+		}
+	}
+}
+
 // Benchmark the use of the Graphemes class.
 func BenchmarkGraphemesClass(b *testing.B) {
 	for i := 0; i < b.N; i++ {