feature: identifier tanya

CHANGELOG.md

@@ -1,9 +1,15 @@
 # go-utils
 
+<a name="v1.42.0"></a>
+## [v1.42.0] - 2025-10-22
+### New Features
+- identifier tanya
+
+
 <a name="v1.41.0"></a>
-## [v1.41.0] - 2025-09-11
+## [v1.41.0] - 2025-09-17
 ### New Features
-- reverse slice
+- reverse slice ([#71](https://github.com/kumparan/go-utils/issues/71))
 
 
 <a name="v1.40.2"></a>
@@ -207,6 +213,9 @@
 - fix marshal issue on gorm.DeletedAt empty value ([#32](https://github.com/kumparan/go-utils/issues/32))
 
 
+<a name="v.1.20.0"></a>
+## [v.1.20.0] - 2022-03-11
+
 <a name="v1.20.0"></a>
 ## [v1.20.0] - 2022-03-11
 ### New Features
@@ -306,11 +315,11 @@
 - add money formatter for multiple currencies ([#13](https://github.com/kumparan/go-utils/issues/13))
 
 
-<a name="v1.8.0"></a>
-## [v1.8.0] - 2020-12-10
-
 <a name="v1.7.1"></a>
 ## [v1.7.1] - 2020-12-10
+
+<a name="v1.8.0"></a>
+## [v1.8.0] - 2020-12-10
 ### New Features
 - add formatter for indonesian money and date
 
@@ -375,7 +384,8 @@
 - init go-utils
 
 
-[Unreleased]: https://github.com/kumparan/go-utils/compare/v1.41.0...HEAD
+[Unreleased]: https://github.com/kumparan/go-utils/compare/v1.42.0...HEAD
+[v1.42.0]: https://github.com/kumparan/go-utils/compare/v1.41.0...v1.42.0
 [v1.41.0]: https://github.com/kumparan/go-utils/compare/v1.40.2...v1.41.0
 [v1.40.2]: https://github.com/kumparan/go-utils/compare/v1.40.1...v1.40.2
 [v1.40.1]: https://github.com/kumparan/go-utils/compare/v1.40.0...v1.40.1
@@ -408,7 +418,8 @@
 [v1.23.0]: https://github.com/kumparan/go-utils/compare/v1.22.0...v1.23.0
 [v1.22.0]: https://github.com/kumparan/go-utils/compare/v1.21.0...v1.22.0
 [v1.21.0]: https://github.com/kumparan/go-utils/compare/v1.20.1...v1.21.0
-[v1.20.1]: https://github.com/kumparan/go-utils/compare/v1.20.0...v1.20.1
+[v1.20.1]: https://github.com/kumparan/go-utils/compare/v.1.20.0...v1.20.1
+[v.1.20.0]: https://github.com/kumparan/go-utils/compare/v1.20.0...v.1.20.0
 [v1.20.0]: https://github.com/kumparan/go-utils/compare/v1.19.3...v1.20.0
 [v1.19.3]: https://github.com/kumparan/go-utils/compare/v1.19.2...v1.19.3
 [v1.19.2]: https://github.com/kumparan/go-utils/compare/v1.19.1...v1.19.2
@@ -425,9 +436,9 @@
 [v1.12.0]: https://github.com/kumparan/go-utils/compare/v1.11.0...v1.12.0
 [v1.11.0]: https://github.com/kumparan/go-utils/compare/v1.10.0...v1.11.0
 [v1.10.0]: https://github.com/kumparan/go-utils/compare/v1.9.0...v1.10.0
-[v1.9.0]: https://github.com/kumparan/go-utils/compare/v1.8.0...v1.9.0
-[v1.8.0]: https://github.com/kumparan/go-utils/compare/v1.7.1...v1.8.0
-[v1.7.1]: https://github.com/kumparan/go-utils/compare/v1.7.0...v1.7.1
+[v1.9.0]: https://github.com/kumparan/go-utils/compare/v1.7.1...v1.9.0
+[v1.7.1]: https://github.com/kumparan/go-utils/compare/v1.8.0...v1.7.1
+[v1.8.0]: https://github.com/kumparan/go-utils/compare/v1.7.0...v1.8.0
 [v1.7.0]: https://github.com/kumparan/go-utils/compare/v1.6.0...v1.7.0
 [v1.6.0]: https://github.com/kumparan/go-utils/compare/v1.5.0...v1.6.0
 [v1.5.0]: https://github.com/kumparan/go-utils/compare/v1.4.0...v1.5.0

tanya/specs.go

@@ -0,0 +1,138 @@
+package tanya
+
+type (
+	// Intent is the intent of a query
+	Intent string
+	// MatchType is the type of match
+	MatchType string
+)
+
+// Intent and MatchType constants
+const (
+	IntentUpdate         Intent = "update"
+	IntentExplain        Intent = "explain"
+	IntentHowTo          Intent = "how_to"
+	IntentDefinition     Intent = "definition"
+	IntentComparison     Intent = "comparison"
+	IntentRecommendation Intent = "recommendation"
+	IntentTroubleshoot   Intent = "troubleshoot"
+	IntentLocation       Intent = "location"
+	IntentTime           Intent = "time"
+	IntentPrice          Intent = "price"
+	IntentContact        Intent = "contact"
+	IntentQuestion       Intent = "question" // general fallback
+	IntentOther          Intent = "other"
+
+	MatchTypeContains    MatchType = "contains"
+	MatchTypeStarts      MatchType = "starts"
+	MatchTypeEnds        MatchType = "ends"
+	MatchTypeTokenSuffix MatchType = "token_suffix" // nolint:gosec
+)
+
+type (
+	// Rule is a rule for matching a query to intent
+	Rule struct {
+		Terms       []string
+		Weight      int
+		MatchType   MatchType
+		MinTokenLen int // optional: for token_suffix; <=0 => default 4
+	}
+
+	// IntentSpec is a specification for intent
+	IntentSpec struct {
+		Intent   Intent
+		Priority int
+		Rules    []Rule
+	}
+)
+
+func terms(ss ...string) []string { return ss }
+
+var intentTable = []IntentSpec{
+	{IntentUpdate, 95, []Rule{
+		{terms("update", "perkembangan", "terbaru", "terkini", "progress", "lanjutan", "pembaruan"), 3, MatchTypeContains, 0},
+		{terms("hari ini", "sekarang", "terkini banget"), 1, MatchTypeContains, 0},
+	}},
+	{IntentExplain, 90, []Rule{
+		{terms("jelaskan", "jelasin", "penjelasan", "uraikan", "explain"), 3, MatchTypeContains, 0},
+		{terms("arti", "artinya", "maksud", "makna", "definisi"), 2, MatchTypeContains, 0},
+	}},
+	{IntentHowTo, 80, []Rule{
+		{terms("bagaimana cara ", "gimana cara "), 3, MatchTypeStarts, 0},
+		{terms("cara "), 3, MatchTypeStarts, 0},
+		{terms(" cara ", " langkah ", " step "), 1, MatchTypeContains, 0},
+		{terms("resep "), 3, MatchTypeStarts, 0},
+		{terms(" panduan ", "panduan "), 2, MatchTypeContains, 0},
+		{terms(" tutorial ", "tutorial "), 2, MatchTypeContains, 0},
+	}},
+	{IntentDefinition, 75, []Rule{
+		{terms("apa itu "), 3, MatchTypeStarts, 0},
+		{terms("apa arti", "apa maksud"), 2, MatchTypeContains, 0},
+	}},
+	{IntentComparison, 70, []Rule{
+		{terms(" vs ", " versus "), 2, MatchTypeContains, 0},
+		{terms("perbedaan ", "beda "), 2, MatchTypeContains, 0},
+		{terms("bagusan mana", "lebih bagus mana", "pilih mana"), 2, MatchTypeContains, 0},
+	}},
+	{IntentRecommendation, 65, []Rule{
+		{terms("rekomendasi", "rekom", "saran"), 2, MatchTypeContains, 0},
+		{terms("bagusan mana", "pilih mana", "cocok yang mana"), 2, MatchTypeContains, 0},
+		{terms("menu "), 2, MatchTypeStarts, 0},
+		{terms(" ide ", "ide "), 1, MatchTypeContains, 0},
+	}},
+	{IntentTroubleshoot, 60, []Rule{
+		{terms("kenapa", "mengapa"), 2, MatchTypeContains, 0},
+		{terms("kok "), 2, MatchTypeStarts, 0},
+		{terms("error", "gagal", "bug", "crash", "macet", "hang"), 2, MatchTypeContains, 0},
+		{terms("solusi ", "fix ", "gimana sih", "kenapa sih"), 1, MatchTypeContains, 0},
+	}},
+	{IntentLocation, 55, []Rule{
+		{terms("dimana", "di mana", "kemana", "ke mana", "lokasi", "alamat"), 2, MatchTypeContains, 0},
+		{terms(" kemana", " dimana", "di mana", " alamat", " lokasi"), 2, MatchTypeEnds, 0},
+	}},
+	{IntentTime, 55, []Rule{
+		{terms("kapan", "jadwal", "jam berapa", "pukul berapa"), 2, MatchTypeContains, 0},
+		{terms("hari ini", "minggu ini", "sekarang", "besok", "nanti sore", "malam ini"), 1, MatchTypeContains, 0},
+	}},
+	{IntentPrice, 50, []Rule{
+		{terms("harga", "biaya", "tarif", "fee", "ongkir"), 2, MatchTypeContains, 0},
+	}},
+	{IntentContact, 50, []Rule{
+		{terms("kontak", "contact", "telepon", "telp", "nomor", "email", "whatsapp", "wa"), 2, MatchTypeContains, 0},
+	}},
+	// fallback tanya umum
+	{IntentQuestion, 10, []Rule{
+		{terms("apa", "apakah", "bagaimana", "gimana", "kapan", "siapa", "dimana", "di mana", "kemana", "ke mana", "berapa"), 2, MatchTypeContains, 0},
+		{terms(" vs ", " versus "), 1, MatchTypeContains, 0},
+		{terms(" yang mana "), 2, MatchTypeContains, 0},
+		{terms("yang mana "), 2, MatchTypeStarts, 0},
+		{terms(" mana"), 2, MatchTypeEnds, 0},
+		{terms("kah"), 1, MatchTypeTokenSuffix, 5},
+		{terms("ya ga sih", "ya gak sih", "ya nggak sih", "ya kan", "apa sih", "gimana sih", "kenapa sih"), 2, MatchTypeContains, 0},
+		{terms(" kok "), 2, MatchTypeContains, 0},
+		{terms("?"), 3, MatchTypeContains, 0},
+	}},
+}
+
+var abbrevMap = map[string]string{
+	"gmn":   "gimana",
+	"gmna":  "gimana",
+	"bgmn":  "bagaimana",
+	"knp":   "kenapa",
+	"knpa":  "kenapa",
+	"dmn":   "di mana",
+	"dmna":  "di mana",
+	"dimn":  "di mana",
+	"kmn":   "ke mana",
+	"kmna":  "ke mana",
+	"brp":   "berapa",
+	"brpa":  "berapa",
+	"kpn":   "kapan",
+	"kpan":  "kapan",
+	"sapa":  "siapa",
+	"sp":    "siapa",
+	"syp":   "siapa",
+	"sypa":  "siapa",
+	"apkh":  "apakah",
+	"apakh": "apakah",
+}

tanya/tanya.go

@@ -0,0 +1,173 @@
+package tanya
+
+import (
+	"sort"
+	"strings"
+	"unicode"
+	"unicode/utf8"
+)
+
+// IsQuestion returns true if a query is a question
+func IsQuestion(q string) bool {
+	intent := ClassifyIntent(q)
+	switch intent { // nolint:exhaustive
+	case IntentPrice, IntentContact, IntentOther:
+		return false
+	default:
+		return true
+	}
+}
+
+// ClassifyIntent returns the most likely intent for the given query
+func ClassifyIntent(q string) Intent {
+	q = normalize(q)
+	if q == "" {
+		return IntentOther
+	}
+	type scored struct {
+		intent      Intent
+		score, prio int
+	}
+	var candidates []scored
+
+	for _, spec := range intentTable {
+		score := 0
+		for _, r := range spec.Rules {
+			if matchByType(q, r) {
+				score += r.Weight
+			}
+		}
+		if score != 0 {
+			candidates = append(candidates, scored{spec.Intent, score, spec.Priority})
+		}
+	}
+	if len(candidates) == 0 {
+		return IntentOther
+	}
+
+	sort.Slice(candidates, func(i, j int) bool {
+		if candidates[i].score == candidates[j].score {
+			return candidates[i].prio > candidates[j].prio
+		}
+		return candidates[i].score > candidates[j].score
+	})
+	return candidates[0].intent
+}
+
+func matchByType(q string, r Rule) bool {
+	switch r.MatchType {
+	case MatchTypeContains:
+		for _, t := range r.Terms {
+			if strings.Contains(q, t) {
+				return true
+			}
+		}
+	case MatchTypeStarts:
+		for _, t := range r.Terms {
+			if strings.HasPrefix(q, t) {
+				return true
+			}
+		}
+	case MatchTypeEnds:
+		for _, t := range r.Terms {
+			if strings.HasSuffix(q, t) {
+				return true
+			}
+		}
+	case MatchTypeTokenSuffix:
+		minLen := r.MinTokenLen
+		if minLen <= 0 {
+			minLen = 4
+		}
+		for _, tok := range tokenize(q) {
+			if len(tok) < minLen {
+				continue
+			}
+			for _, suf := range r.Terms {
+				if strings.HasSuffix(tok, suf) {
+					return true
+				}
+			}
+		}
+	}
+
+	return false
+}
+
+func normalize(s string) string {
+	s = strings.ToLower(strings.TrimSpace(collapseSpaces(s)))
+	s = " " + s + " "
+	s = expandAbbreviations(s)
+	return strings.TrimSpace(collapseSpaces(s))
+}
+
+func collapseSpaces(s string) string {
+	var b strings.Builder
+	sp := false
+	for _, r := range s {
+		if unicode.IsSpace(r) {
+			if !sp {
+				b.WriteByte(' ')
+				sp = true
+			}
+		} else {
+			b.WriteRune(r)
+			sp = false
+		}
+	}
+	return strings.TrimSpace(b.String())
+}
+
+// normalize abbreviations anywhere (start/mid/end)
+func expandAbbreviations(s string) string {
+	words := strings.Fields(s)
+	for i, w := range words {
+		if repl, ok := abbrevMap[w]; ok {
+			words[i] = repl
+			continue
+		}
+		// handle punctuation like "knp?" or "dmn," etc.
+		base := strings.TrimRight(w, "?.!,")
+		suffix := w[len(base):]
+		if repl, ok := abbrevMap[base]; ok {
+			words[i] = repl + suffix
+		}
+	}
+	return strings.Join(words, " ")
+}
+
+// tokenize splits on whitespace and trims leading/trailing non-letters/digits per token.
+// keeps tokens simple & fast (no regex).
+func tokenize(s string) []string {
+	raw := strings.Fields(s)
+	out := make([]string, 0, len(raw))
+	for _, t := range raw {
+		t = trimNonAlphaNum(t)
+		if t != "" {
+			out = append(out, t)
+		}
+	}
+	return out
+}
+
+func trimNonAlphaNum(s string) string {
+	start, end := 0, len(s)
+	for start < end {
+		r := rune(s[start])
+		if isAlphaNum(r) {
+			break
+		}
+		_, w := utf8.DecodeRuneInString(s[start:])
+		start += w
+	}
+	for end > start {
+		r, w := utf8.DecodeLastRuneInString(s[:end])
+		if isAlphaNum(r) {
+			break
+		}
+		end -= w
+	}
+	return s[start:end]
+}
+
+func isAlphaNum(r rune) bool { return unicode.IsLetter(r) || unicode.IsDigit(r) }