feat: add local audio test mode

This commit is contained in:
2026-04-30 02:37:54 +09:00
parent 9dee708b64
commit cf6398f50a
12 changed files with 766 additions and 256 deletions

View File

@@ -1,237 +1,28 @@
import process from "node:process";
import {
GatewayIntentBits,
REST,
Routes,
SlashCommandBuilder,
type ChatInputCommandInteraction,
type Client,
type GuildMember,
type VoiceBasedChannel,
} from "discord.js";
import { Client as DiscordClient } from "discord.js";
import { GuildVoiceSession } from "./audio/guild-voice-session.js";
import { loadConfig } from "./config.js";
import { loadConfig, requireAssistantRuntimeConfig, requireDiscordRuntimeConfig } from "./config.js";
import { runDiscordBot } from "./discord-main.js";
import { Logger } from "./logger.js";
import { ElevenLabsSttService } from "./services/elevenlabs-stt.js";
import { ElevenLabsTtsService } from "./services/elevenlabs-tts.js";
import { OpenAiLlmService } from "./services/openai-llm.js";
import { printLocalAudioDevices, runLocalAssistant } from "./local-main.js";
const mode = process.argv[2] ?? "discord";
const config = loadConfig();
const logger = new Logger(config.LOG_LEVEL);
const commands = [
new SlashCommandBuilder().setName("join").setDescription("현재 들어가 있는 음성 채널에 봇을 입장시킵니다."),
new SlashCommandBuilder().setName("leave").setDescription("현재 음성 세션을 종료합니다."),
new SlashCommandBuilder().setName("status").setDescription("현재 음성 세션 상태를 확인합니다."),
new SlashCommandBuilder().setName("reset").setDescription("대화 문맥과 재생 큐를 초기화합니다."),
new SlashCommandBuilder()
.setName("say")
.setDescription("텍스트를 바로 음성으로 읽습니다.")
.addStringOption((option) =>
option.setName("text").setDescription("읽을 문장").setRequired(true).setMaxLength(400),
),
].map((command) => command.toJSON());
const client = new DiscordClient({
intents: [GatewayIntentBits.Guilds, GatewayIntentBits.GuildVoiceStates],
});
const stt = new ElevenLabsSttService(config);
const tts = new ElevenLabsTtsService(config);
const llm = new OpenAiLlmService(config);
const sessions = new Map<string, GuildVoiceSession>();
function getVoiceChannel(interaction: ChatInputCommandInteraction): VoiceBasedChannel | null {
const member = interaction.member as GuildMember | null;
return member?.voice.channel ?? null;
}
async function registerCommands(appClient: Client): Promise<void> {
const rest = new REST({ version: "10" }).setToken(config.DISCORD_BOT_TOKEN);
if (config.DISCORD_COMMAND_GUILD_ID) {
await rest.put(
Routes.applicationGuildCommands(config.DISCORD_APPLICATION_ID, config.DISCORD_COMMAND_GUILD_ID),
{
body: commands,
},
);
logger.info("Registered guild commands", config.DISCORD_COMMAND_GUILD_ID);
return;
}
await rest.put(Routes.applicationCommands(config.DISCORD_APPLICATION_ID), {
body: commands,
});
logger.info("Registered global commands");
}
async function createSession(interaction: ChatInputCommandInteraction): Promise<GuildVoiceSession> {
if (!interaction.guild) {
throw new Error("Guild interaction required");
}
const voiceChannel = getVoiceChannel(interaction);
if (!voiceChannel) {
throw new Error("먼저 음성 채널에 들어가 주세요.");
}
const existing = sessions.get(interaction.guild.id);
if (existing && existing.voiceChannelId === voiceChannel.id) {
existing.setTextChannel(interaction.channelId);
return existing;
}
if (existing) {
await existing.destroy();
sessions.delete(interaction.guild.id);
}
const session = await GuildVoiceSession.create({
client,
config,
logger,
guild: interaction.guild,
voiceChannel,
textChannelId: interaction.channelId,
stt,
tts,
llm,
});
sessions.set(interaction.guild.id, session);
return session;
}
async function handleJoin(interaction: ChatInputCommandInteraction): Promise<void> {
await interaction.deferReply({ ephemeral: true });
try {
const session = await createSession(interaction);
await interaction.editReply(`음성 비서를 시작했습니다. 채널: ${session.statusSummary().split("\n")[1]?.replace("음성 채널: ", "") ?? "알 수 없음"}`);
} catch (error) {
const message = error instanceof Error ? error.message : "세션 생성에 실패했습니다.";
await interaction.editReply(message);
}
}
async function handleLeave(interaction: ChatInputCommandInteraction): Promise<void> {
const session = interaction.guild ? sessions.get(interaction.guild.id) : undefined;
if (!session) {
await interaction.reply({ content: "현재 활성화된 음성 세션이 없습니다.", ephemeral: true });
return;
}
await session.destroy();
sessions.delete(interaction.guildId!);
await interaction.reply({ content: "음성 세션을 종료했습니다.", ephemeral: true });
}
async function handleStatus(interaction: ChatInputCommandInteraction): Promise<void> {
const session = interaction.guild ? sessions.get(interaction.guild.id) : undefined;
if (!session) {
await interaction.reply({ content: "현재 활성화된 음성 세션이 없습니다.", ephemeral: true });
return;
}
await interaction.reply({
content: session.statusSummary(),
ephemeral: true,
});
}
async function handleReset(interaction: ChatInputCommandInteraction): Promise<void> {
const session = interaction.guild ? sessions.get(interaction.guild.id) : undefined;
if (!session) {
await interaction.reply({ content: "현재 활성화된 음성 세션이 없습니다.", ephemeral: true });
return;
}
session.clearConversation();
await interaction.reply({ content: "대화 문맥과 재생 큐를 초기화했습니다.", ephemeral: true });
}
async function handleSay(interaction: ChatInputCommandInteraction): Promise<void> {
await interaction.deferReply({ ephemeral: true });
const session = interaction.guild ? sessions.get(interaction.guild.id) : undefined;
if (!session) {
await interaction.editReply("먼저 `/join` 으로 음성 세션을 시작해 주세요.");
return;
}
const text = interaction.options.getString("text", true).trim();
await session.speakText(text);
await interaction.editReply("읽기 요청을 대기열에 추가했습니다.");
}
async function shutdown(exitCode = 0): Promise<void> {
logger.info("Shutting down");
for (const session of sessions.values()) {
await session.destroy().catch((error) => {
logger.warn("Session shutdown failed", error);
});
}
sessions.clear();
await client.destroy();
process.exit(exitCode);
}
client.once("ready", async () => {
logger.info("Discord client ready", client.user?.tag ?? "unknown");
try {
await registerCommands(client);
} catch (error) {
logger.error("Command registration failed", error);
}
});
client.on("interactionCreate", async (interaction) => {
if (!interaction.isChatInputCommand()) {
return;
}
try {
switch (interaction.commandName) {
case "join":
await handleJoin(interaction);
return;
case "leave":
await handleLeave(interaction);
return;
case "status":
await handleStatus(interaction);
return;
case "reset":
await handleReset(interaction);
return;
case "say":
await handleSay(interaction);
return;
default:
await interaction.reply({ content: "알 수 없는 명령입니다.", ephemeral: true });
}
} catch (error) {
logger.error("Interaction handler failed", error);
if (interaction.deferred || interaction.replied) {
await interaction.editReply("명령 처리 중 오류가 발생했습니다.").catch(() => null);
return;
}
await interaction.reply({ content: "명령 처리 중 오류가 발생했습니다.", ephemeral: true }).catch(() => null);
}
});
process.on("SIGINT", () => {
void shutdown(0);
});
process.on("SIGTERM", () => {
void shutdown(0);
});
async function main(): Promise<void> {
await client.login(config.DISCORD_BOT_TOKEN);
switch (mode) {
case "discord":
await runDiscordBot(requireDiscordRuntimeConfig(config), logger);
return;
case "local":
await runLocalAssistant(requireAssistantRuntimeConfig(config), logger);
return;
case "local-devices":
await printLocalAudioDevices();
return;
default:
throw new Error(`알 수 없는 실행 모드입니다: ${mode}. 사용 가능: discord, local, local-devices`);
}
}
void main().catch((error) => {