From 293031fe38cfaa6ebf5aa85090910e64fd32dc1d Mon Sep 17 00:00:00 2001
From: Branden J Brown <zephyrtronium@gmail.com>
Date: Sun, 6 Apr 2025 14:20:47 -0400
Subject: [PATCH] emote: add emote parsing engine

---
 emote/emote.go      |  65 ++++++++++++++++++++++++++
 emote/emote_test.go | 111 ++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 176 insertions(+)
 create mode 100644 emote/emote.go
 create mode 100644 emote/emote_test.go

diff --git a/emote/emote.go b/emote/emote.go
new file mode 100644
index 0000000..2fb229b
--- /dev/null
+++ b/emote/emote.go
@@ -0,0 +1,65 @@
+package emote
+
+import (
+	"strings"
+	"unicode"
+)
+
+// Emote is the information Kaiyan needs about an emote.
+type Emote struct {
+	// ID is the emote ID per the source.
+	ID string
+	// Name is the text of the emote as would be parsed from message text.
+	Name string
+	// Source is the name of the emote source, e.g. "7TV", "Twitch:cirno_tv", &c.
+	Source string
+	// Link is a hyperlink to manage the emote.
+	Link string
+	// Image is a hyperlink to the emote image of any size.
+	Image string
+}
+
+// Parser finds emotes in a message.
+//
+// Parser assumes that emotes are bound by whitespace.
+type Parser struct {
+	m map[string]string
+	// TODO(branden): more efficient data structure; trie?
+}
+
+// NewParser creates a Parser for the given list of emotes.
+func NewParser(emotes ...Emote) Parser {
+	m := make(map[string]string, len(emotes))
+	for _, e := range emotes {
+		m[e.Name] = e.ID
+	}
+	return Parser{m}
+}
+
+// Next parses the next emote instance from the message and returns the
+// remainder of the message text following it.
+// If there is no emote in the message the returned emote is the empty string.
+func (p Parser) Next(text string) (name, id, following string) {
+	for text != "" {
+		// First trim any existing space.
+		text = strings.TrimSpace(text)
+		// Then look for the next space.
+		// If there is none, this is the last word of the message; we still
+		// need to look it up.
+		k := strings.IndexFunc(text, unicode.IsSpace)
+		word := text
+		if k >= 0 {
+			word = text[:k]
+			following = text[k:]
+		} else {
+			following = ""
+		}
+		id = p.m[word]
+		if id != "" {
+			return word, id, following
+		}
+		text = following
+	}
+	// No emote found.
+	return "", "", ""
+}
diff --git a/emote/emote_test.go b/emote/emote_test.go
new file mode 100644
index 0000000..4fd4077
--- /dev/null
+++ b/emote/emote_test.go
@@ -0,0 +1,111 @@
+package emote_test
+
+import (
+	"strings"
+	"testing"
+
+	"git.sunturtle.xyz/zephyr/kaiyan/emote"
+	"github.com/google/go-cmp/cmp"
+)
+
+func TestParser(t *testing.T) {
+	cases := []struct {
+		name   string
+		emotes []string
+		text   string
+		want   []string
+	}{
+		{
+			name:   "empty",
+			emotes: strings.Fields("bocchi ryō nijika kita"),
+			text:   "",
+			want:   nil,
+		},
+		{
+			name:   "none",
+			emotes: nil,
+			text:   "bocchi ryō nijika kita",
+			want:   nil,
+		},
+		{
+			name:   "all",
+			emotes: strings.Fields("bocchi ryō nijika kita"),
+			text:   "bocchi ryō nijika kita",
+			want:   strings.Fields("bocchi ryō nijika kita"),
+		},
+		{
+			name:   "some",
+			emotes: strings.Fields("bocchi nijika"),
+			text:   "bocchi ryō nijika kita",
+			want:   strings.Fields("bocchi nijika"),
+		},
+		{
+			name:   "others",
+			emotes: strings.Fields("ryō kita"),
+			text:   "bocchi ryō nijika kita",
+			want:   strings.Fields("ryō kita"),
+		},
+	}
+	for _, c := range cases {
+		t.Run(c.name, func(t *testing.T) {
+			// Convert the emote names to emotes proper.
+			e := make([]emote.Emote, len(c.emotes))
+			for i, w := range c.emotes {
+				e[i] = emote.Emote{ID: w, Name: w, Source: c.name}
+			}
+			p := emote.NewParser(e...)
+			var got []string
+			text := c.text
+			for {
+				name, id, rest := p.Next(text)
+				if id == "" {
+					break
+				}
+				if name != id {
+					// Not normally the case, but this test is constructed
+					// so that it is.
+					t.Errorf("wrong id %q for name %q", id, name)
+				}
+				got = append(got, name)
+				text = rest
+			}
+			if diff := cmp.Diff(c.want, got); diff != "" {
+				t.Errorf("wrong emotes (-want/+got):\n%s", diff)
+			}
+		})
+	}
+}
+
+func BenchmarkParser(b *testing.B) {
+	e := []emote.Emote{
+		{ID: "bocchi", Name: "bocchi"},
+		{ID: "ryō", Name: "ryō"},
+		{ID: "nijika", Name: "nijika"},
+		{ID: "kita", Name: "kita"},
+		{ID: "seika", Name: "seika"},
+		{ID: "kikuri", Name: "kikuri"},
+		{ID: "eliza", Name: "eliza"},
+		{ID: "shima", Name: "shima"},
+	}
+	p := emote.NewParser(e...)
+	texts := []struct {
+		name string
+		text string
+	}{
+		{"none-short", "none"},
+		{"none-long", strings.TrimSpace(strings.Repeat("none ", 100))},
+		{"single", "bocchi"},
+		{"natural", strings.TrimSpace(strings.Repeat("kessoku band is bocchi ryō nijika kita starry is seika PA-san sickhack is kikuri eliza shima ", 5))},
+	}
+	for _, c := range texts {
+		b.Run(c.name, func(b *testing.B) {
+			b.SetBytes(int64(len(c.text)))
+			for b.Loop() {
+				text := c.text
+				for text != "" {
+					_, _, text = p.Next(text)
+				}
+			}
+		})
+	}
+}