feat: voice message from bot

Dockerfile

@@ -9,6 +9,8 @@ FROM python:3.13-slim
 
 WORKDIR /app
 
+RUN apt-get update && apt-get install -y --no-install-recommends ffmpeg && rm -rf /var/lib/apt/lists/*
+
 COPY --from=builder /root/.local /root/.local
 ENV PATH=/root/.local/bin:$PATH
 

bot.py

@@ -2,8 +2,10 @@ import asyncio
 import io
 import logging
 import os
+import random
 import re
 import sqlite3
+import subprocess
 from dataclasses import dataclass
 from pathlib import Path
 
@@ -13,7 +15,7 @@ from aiogram.client.session.aiohttp import AiohttpSession
 from aiogram.enums import ChatAction, ParseMode
 from aiogram.exceptions import TelegramBadRequest
 from aiogram.filters import Command, CommandStart
-from aiogram.types import Message
+from aiogram.types import BufferedInputFile, Message
 from dotenv import load_dotenv
 from httpx import AsyncClient as HttpxAsyncClient
 from openai import APIConnectionError, APIError, APITimeoutError, AsyncOpenAI, NotFoundError, RateLimitError
@@ -33,6 +35,8 @@ class Settings:
     api_token: str
     model: str
     transcription_model: str
+    tts_model: str | None
+    tts_voice: str
     telegram_proxy: str | None
     whitelist_user_ids: frozenset[int]
     system_prompt: str
@@ -43,6 +47,33 @@ class Settings:
     database_path: Path
     max_context_messages: int
     message_batch_delay_seconds: float
+    voice_reply_enabled: bool
+    voice_reply_chance: float
+    voice_reply_min_chars: int
+
+
+def _bool_env(name: str, default: bool) -> bool:
+    value = os.getenv(name, "").strip().lower()
+    if not value:
+        return default
+    if value in {"1", "true", "yes", "on"}:
+        return True
+    if value in {"0", "false", "no", "off"}:
+        return False
+    raise RuntimeError(f"{name} must be a boolean")
+
+
+def _probability_env(name: str, default: float) -> float:
+    value = os.getenv(name, "").strip()
+    if not value:
+        return default
+    try:
+        parsed = float(value)
+    except ValueError as exc:
+        raise RuntimeError(f"{name} must be a number between 0 and 1") from exc
+    if parsed < 0 or parsed > 1:
+        raise RuntimeError(f"{name} must be between 0 and 1")
+    return parsed
 
 
 def _required_env(name: str) -> str:
@@ -126,6 +157,8 @@ def load_settings() -> Settings:
         api_token=_required_env("API_TOKEN"),
         model=_required_env("MODEL"),
         transcription_model=_required_env("TRANSCRIPTION_MODEL"),
+        tts_model=os.getenv("TTS_MODEL", "").strip() or None,
+        tts_voice=os.getenv("TTS_VOICE", "Kore").strip() or "Kore",
         telegram_proxy=os.getenv("TELEGRAM_PROXY", "").strip() or None,
         whitelist_user_ids=_whitelist_env(),
         system_prompt=_load_system_prompt(),
@@ -136,6 +169,9 @@ def load_settings() -> Settings:
         database_path=Path(os.getenv("DATABASE_PATH", "data/bot.sqlite3")).expanduser(),
         max_context_messages=_int_env("MAX_CONTEXT_MESSAGES", 20),
         message_batch_delay_seconds=_float_env("MESSAGE_BATCH_DELAY_SECONDS", 3),
+        voice_reply_enabled=_bool_env("VOICE_REPLY_ENABLED", False),
+        voice_reply_chance=_probability_env("VOICE_REPLY_CHANCE", 0.1),
+        voice_reply_min_chars=_int_env("VOICE_REPLY_MIN_CHARS", 250),
     )
 
 
@@ -245,6 +281,11 @@ def voice_transcript_text(transcript: str) -> str:
     return f"[Voice message transcript]\n{transcript.strip()}"
 
 
+def voice_reply_text(reply: str) -> str:
+    text = "\n\n".join(text for text, _ in split_reply(reply) if text).strip()
+    return text or "The model returned an empty response."
+
+
 async def keep_typing(bot: Bot, chat_id: int, stop: asyncio.Event) -> None:
     while not stop.is_set():
         try:
@@ -319,6 +360,66 @@ async def safe_answer(message: Message, text: str) -> None:
         await message.answer(text[: settings.max_response_chars])
 
 
+def should_send_voice_reply(reply: str) -> bool:
+    if not settings.voice_reply_enabled or not settings.tts_model:
+        return False
+    plain_text = voice_reply_text(reply)
+    return len(plain_text) >= settings.voice_reply_min_chars and random.random() < settings.voice_reply_chance
+
+
+async def synthesize_voice_reply(text: str) -> bytes:
+    response = await client.audio.speech.create(
+        model=settings.tts_model,
+        voice=settings.tts_voice,
+        input=text,
+        response_format="pcm",
+    )
+    audio_bytes = response.read()
+    if not audio_bytes:
+        raise RuntimeError("TTS model returned empty audio")
+    return audio_bytes
+
+
+async def convert_pcm_to_ogg_opus(audio_bytes: bytes) -> bytes:
+    def _convert() -> bytes:
+        process = subprocess.run(
+            [
+                "ffmpeg",
+                "-f",
+                "s16le",
+                "-ar",
+                "24000",
+                "-ac",
+                "1",
+                "-i",
+                "pipe:0",
+                "-c:a",
+                "libopus",
+                "-b:a",
+                "48k",
+                "-f",
+                "ogg",
+                "pipe:1",
+            ],
+            input=audio_bytes,
+            stdout=subprocess.PIPE,
+            stderr=subprocess.PIPE,
+            check=True,
+        )
+        if not process.stdout:
+            raise RuntimeError("ffmpeg returned empty output")
+        return process.stdout
+
+    return await asyncio.to_thread(_convert)
+
+
+async def send_voice_reply(message: Message, reply: str) -> None:
+    voice_text = voice_reply_text(reply)
+    pcm_audio = await synthesize_voice_reply(voice_text)
+    ogg_audio = await convert_pcm_to_ogg_opus(pcm_audio)
+    await message.answer_voice(BufferedInputFile(ogg_audio, filename="reply.ogg"))
+
+
 async def wait_between_reply_parts(bot: Bot, chat_id: int, timeout: int) -> None:
     if timeout <= 0:
         await bot.send_chat_action(chat_id=chat_id, action=ChatAction.TYPING)
@@ -398,6 +499,23 @@ async def process_pending_messages(message: Message, bot: Bot, user_id: int) ->
             if should_store_reply:
                 await history.append_turn(user_id, user_text, reply)
 
+        if should_send_voice_reply(reply):
+            try:
+                await send_voice_reply(message, reply)
+                return
+            except RateLimitError:
+                logger.warning("TTS API rate limit exceeded", exc_info=True)
+            except (APITimeoutError, APIConnectionError):
+                logger.warning("TTS API connection problem", exc_info=True)
+            except NotFoundError:
+                logger.exception("TTS endpoint or model was not found")
+            except APIError:
+                logger.exception("TTS API returned an error")
+            except subprocess.CalledProcessError:
+                logger.exception("ffmpeg failed to convert synthesized audio")
+            except Exception:
+                logger.exception("Unexpected voice reply error")
+
         await send_reply(message, bot, reply)
     except asyncio.CancelledError:
         raise