ASR auto-submit

This commit is contained in:
Dita Aji Pratama 2026-10-08 21:10:02 +07:00
parent 911e44fdd5
commit d3d16e85c7
3 changed files with 36 additions and 4 deletions

View File

@ -28,6 +28,8 @@ asr: # Speech-to-Text (push-to-talk TUI). Model dikelola via moop tipe 'asr'.
max_seconds : 60 # auto-stop pengaman
transcribe_timeout: 60 # detik timeout request ASR
keep_recordings : true # false = hapus file setelah ditranskrip
auto_submit : false # true = hasil transkrip langsung dikirim
# false = hanya mengisi form (Ctrl+Enter)
session: # Chat session
db_path: "~/.config/hendrik/sessions.json"

View File

@ -61,6 +61,10 @@ ASR_KEEP_RECORDINGS = _as_bool(_yaml_get("asr", "keep_recordings", default=True)
# Tombol push-to-talk: "f1".."f12" atau "ctrl-<huruf>". Default F9.
ASR_RECORD_KEY = str(_yaml_get("asr", "record_key", default="f9")).strip().lower()
# true = hasil transkripsi langsung dikirim sebagai pesan,
# false = hasil hanya disisipkan ke form input (user kirim sendiri via Ctrl+Enter).
ASR_AUTO_SUBMIT = _as_bool(_yaml_get("asr", "auto_submit", default=False), default=False)
XMPP_USERNAME = _yaml_get("xmpp", "username", default="")
XMPP_PASSWORD = _yaml_get("xmpp", "password", default="")

View File

@ -2,9 +2,13 @@
#
# Alur:
# * F9 (configurable) → mulai rekam (ffmpeg subprocess, non-blocking).
# * F9 lagi / Enter → stop rekam → transkrip (thread) → hasil DISISIPKAN
# ke form input pada posisi kursor, TIDAK submit.
# * F9 lagi / Enter → stop rekam → transkrip (thread) → hasil
# disisipkan ke form input pada posisi kursor,
# TIDAK submit (default).
# * Backspace / Esc → batal: rekaman dibuang / hasil transkrip diabaikan.
# * config.asr.auto_submit: true → hasil transkrip langsung dikirim sebagai
# pesan (setara Ctrl+Enter). Jika form sudah terisi
# saat transkrip selesai, hasil tetap disisipkan.
#
# State disimpan di app:
# app.talk_recorder — instance Recorder saat merekam, else None
@ -18,7 +22,7 @@ import threading
import config
from tools import asr as asr_mod
from .agent import log
from .agent import log, submit
def _shortcut_label():
@ -209,7 +213,8 @@ def shutdown(app):
def finish_talk(app, stdscr):
"""Dipanggil main loop saat app.talk_done ter-set: sisipkan hasil."""
"""Dipanggil main loop saat app.talk_done ter-set: sisipkan hasil (atau
langsung submit bila config.asr.auto_submit aktif)."""
res = getattr(app, "_talk_result", None)
app._talk_result = None
if getattr(app, "talk_cancelled", False):
@ -222,6 +227,27 @@ def finish_talk(app, stdscr):
if not text:
log(app, "system", " Tidak ada ucapan terdeteksi.")
return
if config.ASR_AUTO_SUBMIT:
# Auto-submit: taruh teks di buffer lalu kirim (submit() yang
# membersihkan buffer). Jika user sudah mengetik sesuatu sebelum
# transkrip selesai, teks tidak ditimpa — disisipkan saja.
if not "\n".join(app.input_buffer).strip():
app.input_buffer = text.split("\n")
app.input_line = len(app.input_buffer) - 1
app.input_col = len(app.input_buffer[-1])
submit(app, stdscr)
log(app, "system",
f" \u266a [{res.get('provider', '?')} / {res.get('model', '?')}] "
"teks otomatis dikirim")
else:
insert_text(app, text)
log(app, "system",
f" \u266a [{res.get('provider', '?')} / {res.get('model', '?')}] "
"form sedang terisi \u2014 hasil disisipkan, tidak auto-submit "
"(Ctrl+Enter untuk kirim)")
if app.scroll_follow:
app.scroll = 999999
return
if insert_text(app, text):
log(app, "system",
f" \u266a [{res.get('provider', '?')} / {res.get('model', '?')}] "