feat: label code fences so Vicinae can colour them (#10)

src/lib/code.ts

@@ -0,0 +1,197 @@
+/**
+ * Labels fenced code blocks the model left unlabelled.
+ *
+ * Vicinae renders fences through `MdCodeBlock.qml`, which gives a dark panel, a
+ * header showing the language, a Copy button, and syntax colouring resolved by
+ * KSyntaxHighlighting. That last part is why the label matters beyond
+ * cosmetics: `syntax::highlight` looks the language up with
+ * `definitionForName(language)` and then `definitionForFileName("file." + language)`,
+ * so `ts` resolves through `file.ts` and `bash` through its definition name,
+ * while an unrecognised string such as `terminal` still renders as a label but
+ * produces no highlighting at all. A wrong label is therefore worse than none,
+ * and this stays deliberately conservative: it only labels a block when there is
+ * a clear signal, and leaves anything ambiguous alone.
+ */
+
+/** First words that mark a block as shell rather than a programming language. */
+const SHELL_COMMANDS = new Set([
+  "apt",
+  "brew",
+  "bun",
+  "cat",
+  "cd",
+  "chmod",
+  "chown",
+  "code",
+  "cp",
+  "curl",
+  "df",
+  "docker",
+  "du",
+  "echo",
+  "export",
+  "find",
+  "gh",
+  "git",
+  "grep",
+  "htop",
+  "jq",
+  "kill",
+  "kubectl",
+  "ln",
+  "ls",
+  "make",
+  "man",
+  "mkdir",
+  "mv",
+  "nano",
+  "node",
+  "npm",
+  "npx",
+  "open",
+  "pacman",
+  "pip",
+  "pnpm",
+  "ps",
+  "psql",
+  "pwd",
+  "python",
+  "rm",
+  "rsync",
+  "scp",
+  "sed",
+  "source",
+  "ssh",
+  "sudo",
+  "systemctl",
+  "tar",
+  "top",
+  "touch",
+  "unzip",
+  "vim",
+  "wget",
+  "which",
+  "yarn",
+  "zip",
+]);
+
+function significantLines(code: string): string[] {
+  return code
+    .split("\n")
+    .map((line) => line.trim())
+    .filter((line) => line.length > 0 && !line.startsWith("#") && !line.startsWith("//"));
+}
+
+function looksLikeShell(code: string): boolean {
+  const lines = significantLines(code);
+  if (lines.length === 0) return false;
+
+  // An explicit prompt marker, or a leading `~` path, is about as clear as it gets.
+  if (lines.some((line) => /^[$>] /.test(line) || line.startsWith("~/"))) return true;
+
+  // A shell command as the first word of the first meaningful line.
+  const [firstWord] = lines[0].split(/\s+/);
+  if (SHELL_COMMANDS.has(firstWord)) return true;
+
+  // Failing that, a pipeline or chained command between two known commands.
+  if (/[|&]{1,2}/.test(code)) {
+    return lines.some((line) => {
+      const [word] = line.split(/\s+/);
+      return SHELL_COMMANDS.has(word.replace(/[|&].*$/, ""));
+    });
+  }
+
+  return false;
+}
+
+function looksLikeJson(code: string): boolean {
+  const trimmed = code.trim();
+  if (!trimmed.startsWith("{") && !trimmed.startsWith("[")) return false;
+
+  try {
+    const parsed: unknown = JSON.parse(trimmed);
+    return typeof parsed === "object" && parsed !== null;
+  } catch {
+    return false;
+  }
+}
+
+/** TypeScript-specific syntax. Plain JS is left unlabelled rather than guessed at. */
+function looksLikeTypeScript(code: string): boolean {
+  return (
+    /\binterface\s+\w+\s*\{/.test(code) ||
+    /\btype\s+\w+\s*=/.test(code) ||
+    /\bas\s+const\b/.test(code) ||
+    /:\s*(string|number|boolean|void|unknown|never)\b/.test(code) ||
+    /<\w+>(\s*\(|\s*=|\s*=>)/.test(code) ||
+    /\b(implements|readonly)\s+/.test(code)
+  );
+}
+
+function looksLikePython(code: string): boolean {
+  return (
+    /^\s*def\s+\w+\s*\(/m.test(code) ||
+    /^\s*from\s+[\w.]+\s+import\s+/m.test(code) ||
+    /^\s*import\s+\w+$/m.test(code) ||
+    /\bself\./.test(code) ||
+    /^\s*(async\s+)?def\s+/m.test(code)
+  );
+}
+
+/**
+ * Best-effort language tag, or `undefined` when nothing is clear enough.
+ * Order matters: a shell pipeline can contain what look like code fragments.
+ */
+export function inferLanguage(code: string): string | undefined {
+  if (code.trim().length === 0) return undefined;
+
+  if (looksLikeShell(code)) return "bash";
+  if (looksLikeJson(code)) return "json";
+  if (looksLikeTypeScript(code)) return "ts";
+  if (looksLikePython(code)) return "python";
+
+  return undefined;
+}
+
+const OPENING_FENCE = /^(\s*)(`{3,}|~{3,})\s*$/;
+
+/**
+ * Adds a language to fences that have none, leaving labelled fences untouched.
+ *
+ * Only bare opening fences are rewritten, so the model's own choice always wins.
+ */
+export function labelCodeFences(markdown: string): string {
+  const lines = markdown.split("\n");
+  const out: string[] = [];
+  let index = 0;
+
+  while (index < lines.length) {
+    const openingLine = lines[index];
+    const opener = OPENING_FENCE.exec(openingLine);
+
+    if (!opener) {
+      out.push(openingLine);
+      index += 1;
+      continue;
+    }
+
+    const [, indent, fence] = opener;
+    const body: string[] = [];
+    index += 1;
+
+    while (index < lines.length && !lines[index].trimStart().startsWith(fence)) {
+      body.push(lines[index]);
+      index += 1;
+    }
+
+    const closing = index < lines.length ? lines[index] : "";
+    if (closing) index += 1;
+
+    const language = inferLanguage(body.join("\n"));
+    out.push(language ? `${indent}${fence}${language}` : openingLine);
+    out.push(...body);
+    if (closing) out.push(closing);
+  }
+
+  return out.join("\n");
+}