new: language auto-detect - /api/guess-language, refresh button, guess on paste

guess.go

@@ -0,0 +1,74 @@
+package main
+
+import (
+	"encoding/json"
+	"net/http"
+	"regexp"
+	"strings"
+)
+
+// guessLang heuristically detects a language from source content.
+func guessLang(s string) string {
+	src := strings.TrimSpace(s)
+	if src == "" {
+		return ""
+	}
+
+	// JSON: must start with { or [ and parse
+	if src[0] == '{' || src[0] == '[' {
+		var v any
+		if json.Unmarshal([]byte(src), &v) == nil {
+			return "json"
+		}
+	}
+
+	rules := []struct {
+		lang string
+		re   *regexp.Regexp
+	}{
+		{"python", regexp.MustCompile(`(?m)^\s*(def |class |import |from \w+ import |@decorator)`)},
+		{"python", regexp.MustCompile(`(^|\n)\s*#!.*python`)},
+		{"go", regexp.MustCompile(`(?m)^\s*(package \w+|import \(|func (\w+|\() )`)},
+		{"rust", regexp.MustCompile(`(?m)(\bfn \w+|let mut \b|\bimpl \b|\bmatch \w+ \{|use std::)`)},
+		{"javascript", regexp.MustCompile(`(?m)(\bconst \w+ = |require\(|import \w+ from |=> \{|\bconsole\.log\()` )},
+		{"java", regexp.MustCompile(`(?m)(\bpublic (static |final |class )|\bSystem\.out\.print|import java\.)`)},
+		{"c", regexp.MustCompile(`(?m)(#include\s*<\w+\.h>|printf\(|\bint main\()` )},
+		{"cpp", regexp.MustCompile(`(?m)(#include\s*<(iostream|vector|string)>|std::|\bcout\s*<<)`)},
+		{"bash", regexp.MustCompile(`(?m)^(#!.*bash|#!.*sh|\w+\(\)\s*\{)` )},
+		{"sql", regexp.MustCompile(`(?i)\b(SELECT .+ FROM|INSERT INTO|CREATE TABLE|UPDATE \w+ SET)\b`)},
+		{"yaml", regexp.MustCompile(`(?m)^(\w[\w-]*:\s*(\||\S)|  \w[\w-]*: |---\s*$)`)},
+		{"markdown", regexp.MustCompile("(?m)^(#{1,6} \\S|\\|.*\\||-\\s\\[\\s?\\]|```)")},
+	}
+
+	scores := map[string]int{}
+	for _, r := range rules {
+		n := len(r.re.FindAllString(src, -1))
+		if n > 0 {
+			// later generic rules shouldn't drown out earlier specific ones
+			scores[r.lang] += n
+		}
+	}
+
+	best, bestN := "", 0
+	for lang, n := range scores {
+		if n > bestN {
+			best, bestN = lang, n
+		}
+	}
+	if bestN == 0 {
+		return "text"
+	}
+	return best
+}
+
+func (a *apiServer) handleGuessLang(w http.ResponseWriter, r *http.Request) {
+	var req struct {
+		Content string `json:"content"`
+	}
+	if err := json.NewDecoder(r.Body).Decode(&req); err != nil {
+		writeErr(w, http.StatusBadRequest, "invalid json body")
+		return
+	}
+	lang := guessLang(req.Content)
+	writeJSON(w, http.StatusOK, map[string]any{"language": lang})
+}