@koda-sl/baker-bridge 0.138.0-dev.43453bfcd → 0.139.0-dev.a1467d0a8

package/dist/review/reply-language.js

@@ -0,0 +1,81 @@
+// Which language a chat message is written in, read by eld's n-gram model:
+// offline and deterministic, so the same Session always digests the same way.
+// franc's trigrams read short English lines ("I'll look at what Alex sent.",
+// "publish it please") as German or French; eld reads them as English. Only the
+// languages below compete, which keeps a short Spanish line from being read as
+// Galician or Asturian.
+import { eld } from "eld/small";
+const LANGUAGES = {
+    en: "English",
+    es: "Spanish",
+    pt: "Portuguese",
+    ca: "Catalan",
+    fr: "French",
+    it: "Italian",
+    de: "German",
+};
+const detector = eld.newInstance();
+detector.setLanguageSubset(Object.keys(LANGUAGES));
+// Fenced and inline code, quoted text and paths are the marketer's own words or
+// our names played back — a page title, a funnel step, a file, a code comment —
+// and keep their language whatever the sentence around them is written in.
+const NOT_PROSE = /```[\s\S]*?```|"[^"\n]*"|“[^”\n]*”|«[^»\n]*»|`[^`\n]*`|https?:\/\/\S+|(?:^|\s)[\w.-]*\/[\w./-]+/g;
+// Fewer letters than this is too short to tell: "ok", "go", "Listo." Such a
+// message is never flagged, and never sets the marketer's language.
+const MIN_LETTERS = 15;
+const SPEAKER_LABEL = /\[Current speaker: [^\]\n]*\]\n+/g;
+const ATTACHED_FILES = /=== ATTACHED [\s\S]*?=== END ATTACHED FILES ===/g;
+// A workspace skill the marketer typed (`/landing Crea una página…`) reaches the
+// SDK verbatim, without a speaker label, so its arguments are theirs. The CLI
+// records it either as typed or as `<command-args>`.
+const TYPED_COMMAND = /^\/[^\s/]+\s+([\s\S]+)$/;
+const COMMAND_ARGS = /<command-args>([\s\S]*?)<\/command-args>/;
+const letters = (text) => (text.match(/\p{L}/gu) ?? []).length;
+function read(text) {
+    return letters(text) < MIN_LETTERS ? undefined : LANGUAGES[detector.detect(text).language];
+}
+/** The language a message is written in, or undefined when it is too short to tell. */
+export function messageLanguage(text) {
+    return read(text.replace(NOT_PROSE, " "));
+}
+/**
+ * What the marketer typed in a turn's opening line, or undefined when nobody
+ * typed it. The bridge labels every message a person sent with
+ * `[Current speaker: …]`, and sends a typed skill command as it was typed; a
+ * skill body, a task notification or hook feedback arrives in the same seat
+ * with neither, and must not set the language the replies are read against.
+ * What follows the label is theirs, less the attachment block the bridge
+ * appends in English.
+ */
+export function marketerWords(userLine) {
+    const labels = [...userLine.matchAll(SPEAKER_LABEL)];
+    const last = labels[labels.length - 1];
+    if (last?.index === undefined) {
+        const command = COMMAND_ARGS.exec(userLine) ?? TYPED_COMMAND.exec(userLine.trim());
+        return command?.[1]?.replace(ATTACHED_FILES, " ");
+    }
+    return userLine.slice(last.index + last[0].length).replace(ATTACHED_FILES, " ");
+}
+/**
+ * A reply in another language than the marketer's, or undefined. Only English
+ * beside another language counts: Spanish, Catalan and Portuguese are too close
+ * to tell apart on one line, and the drift that reaches marketers is into the
+ * English of the tools, or out of English into an older Spanish context. The
+ * reply is read whole and sentence by sentence, so one English line inside a
+ * Spanish message still shows.
+ */
+export function languageMismatch(reply, marketer) {
+    if (marketer === undefined) {
+        return undefined;
+    }
+    const words = reply.replace(NOT_PROSE, " ");
+    const parts = [words, ...words.split(/(?<=[.!?:;])\s+|\n+/)];
+    for (const part of parts) {
+        const written = read(part);
+        if (written !== undefined && written !== marketer && (written === "English" || marketer === "English")) {
+            return { reply: written, marketer };
+        }
+    }
+    return undefined;
+}
+//# sourceMappingURL=reply-language.js.map

package/dist/review/reply-language.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"reply-language.js","sourceRoot":"","sources":["../../src/review/reply-language.ts"],"names":[],"mappings":"AAAA,2EAA2E;AAC3E,8EAA8E;AAC9E,6EAA6E;AAC7E,gFAAgF;AAChF,+EAA+E;AAC/E,wBAAwB;AACxB,OAAO,EAAE,GAAG,EAAE,MAAM,WAAW,CAAC;AAEhC,MAAM,SAAS,GAA2B;IACxC,EAAE,EAAE,SAAS;IACb,EAAE,EAAE,SAAS;IACb,EAAE,EAAE,YAAY;IAChB,EAAE,EAAE,SAAS;IACb,EAAE,EAAE,QAAQ;IACZ,EAAE,EAAE,SAAS;IACb,EAAE,EAAE,QAAQ;CACb,CAAC;AAEF,MAAM,QAAQ,GAAG,GAAG,CAAC,WAAW,EAAE,CAAC;AACnC,QAAQ,CAAC,iBAAiB,CAAC,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,CAAC;AAEnD,gFAAgF;AAChF,gFAAgF;AAChF,2EAA2E;AAC3E,MAAM,SAAS,GAAG,kGAAkG,CAAC;AAErH,4EAA4E;AAC5E,oEAAoE;AACpE,MAAM,WAAW,GAAG,EAAE,CAAC;AAEvB,MAAM,aAAa,GAAG,mCAAmC,CAAC;AAC1D,MAAM,cAAc,GAAG,kDAAkD,CAAC;AAC1E,iFAAiF;AACjF,8EAA8E;AAC9E,qDAAqD;AACrD,MAAM,aAAa,GAAG,yBAAyB,CAAC;AAChD,MAAM,YAAY,GAAG,0CAA0C,CAAC;AAEhE,MAAM,OAAO,GAAG,CAAC,IAAY,EAAE,EAAE,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,CAAC,MAAM,CAAC;AAEvE,SAAS,IAAI,CAAC,IAAY;IACxB,OAAO,OAAO,CAAC,IAAI,CAAC,GAAG,WAAW,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,SAAS,CAAC,QAAQ,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC,QAAQ,CAAC,CAAC;AAC7F,CAAC;AAED,uFAAuF;AACvF,MAAM,UAAU,eAAe,CAAC,IAAY;IAC1C,OAAO,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC,CAAC,CAAC;AAC5C,CAAC;AAED;;;;;;;;GAQG;AACH,MAAM,UAAU,aAAa,CAAC,QAAgB;IAC5C,MAAM,MAAM,GAAG,CAAC,GAAG,QAAQ,CAAC,QAAQ,CAAC,aAAa,CAAC,CAAC,CAAC;IACrD,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC;IACvC,IAAI,IAAI,EAAE,KAAK,KAAK,SAAS,EAAE,CAAC;QAC9B,MAAM,OAAO,GAAG,YAAY,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,aAAa,CAAC,IAAI,CAAC,QAAQ,CAAC,IAAI,EAAE,CAAC,CAAC;QACnF,OAAO,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,OAAO,CAAC,cAAc,EAAE,GAAG,CAAC,CAAC;IACpD,CAAC;IACD,OAAO,QAAQ,CAAC,KAAK,CAAC,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,OAAO,CAAC,cAAc,EAAE,GAAG,CAAC,CAAC;AAClF,CAAC;AAED;;;;;;;GAOG;AACH,MAAM,UAAU,gBAAgB,CAC9B,KAAa,EACb,QAA4B;IAE5B,IAAI,QAAQ,KAAK,SAAS,EAAE,CAAC;QAC3B,OAAO,SAAS,CAAC;IACnB,CAAC;IACD,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,CAAC,SAAS,EAAE,GAAG,CAAC,CAAC;IAC5C,MAAM,KAAK,GAAG,CAAC,KAAK,EAAE,GAAG,KAAK,CAAC,KAAK,CAAC,qBAAqB,CAAC,CAAC,CAAC;IAC7D,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,OAAO,GAAG,IAAI,CAAC,IAAI,CAAC,CAAC;QAC3B,IAAI,OAAO,KAAK,SAAS,IAAI,OAAO,KAAK,QAAQ,IAAI,CAAC,OAAO,KAAK,SAAS,IAAI,QAAQ,KAAK,SAAS,CAAC,EAAE,CAAC;YACvG,OAAO,EAAE,KAAK,EAAE,OAAO,EAAE,QAAQ,EAAE,CAAC;QACtC,CAAC;IACH,CAAC;IACD,OAAO,SAAS,CAAC;AACnB,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@koda-sl/baker-bridge",
-  "version": "0.138.0-dev.43453bfcd",
+  "version": "0.139.0-dev.a1467d0a8",
   "description": "HTTP server wrapping the Claude Agent SDK. Persists chat events to Convex; the dashboard subscribes via Convex realtime",
   "license": "MIT",
   "type": "module",
@@ -44,6 +44,7 @@
     "@anthropic-ai/claude-agent-sdk": "^0.3.295",
     "@hono/node-server": "^1.19.14",
     "@t3-oss/env-core": "^0.13.10",
+    "eld": "^2.1.0",
     "hono": "^4.12.25",
     "image-size": "^2.0.2",
     "mockttp": "4.6.3",