feat(chat): scenario-based training — choice of channel/Situation + realistic per-turn evaluation

Backend:
- Channel/initiation now driven by a SCENARIO chosen at chat start, not baked into the
  persona: social (customer opens), f2f_call (seller must open, proactive), recontact
  (customer re-contacts after 1-3 months).
- /chat/start accepts {scenario}; session stores scenario + internal{turns,score}.
- persona_reply takes scenario + adapts tone; system-role transcript entries are fed to
  the persona as hidden scene notes.
- JUDGE updated for realism: good response can WIN even in hard/tough-text scenarios;
  long/no-close chats (turns >~12) lose; pushy/ignoring-need loses. Efficiency rewarded.

Frontend:
- Scenario picker before chat (choose Social / Face-to-face-call / Re-contact).
- Chat thread renders role=system as a centered time-lapse/scene note.
- Choose-scenario i18n (EN+TH).
Rebuilt dist.

frontend/src/api/index.js

@@ -47,7 +47,7 @@ export const api = {
   listPersonas: (gid) => request('GET', `/api/groups/${gid}/personas`),
   getPersona: (gid, pid) => request('GET', `/api/groups/${gid}/personas/${pid}`),
   updatePersona: (gid, pid, b) => request('PUT', `/api/groups/${gid}/personas/${pid}`, b),
-  chatStart: (gid, pid) => request('POST', `/api/chat/${gid}/personas/${pid}/chat/start`),
+  chatStart: (gid, pid, scenario) => request('POST', `/api/chat/${gid}/personas/${pid}/chat/start`, scenario || {}),
   chatSend: (gid, pid, text) => request('POST', `/api/chat/${gid}/personas/${pid}/chat/send`, { text }),
   chatFinish: (gid, pid) => request('POST', `/api/chat/${gid}/personas/${pid}/chat/finish`),
   mySessions: () => request('GET', '/api/chat/sessions'),