From b252c39517b9fce181d1a0fe4761404c7d981397 Mon Sep 17 00:00:00 2001
From: salmonik <salmonik@tuta.io>
Date: Tue, 18 Aug 2026 23:18:37 +0200
Subject: feat: local TTS via piper subprocess with Google fallback

- New src/tts.rs: centralize TTS in synthesize(config, http_client, text)
  Piper subprocess primary, Google Translate fallback on error/missing
- TTS_MODEL_PATH/TTS_VOICE env config (defaults to pl_PL-bass-high)
- Fix Unknown interaction errors: spawn play_tts_in_channel in background
  so voice-state TTS no longer blocks the shard event loop (3s Discord limit)
- Best-effort confirmation replies in /tts to prevent error cascade
- Custom on_error handler: graceful interaction-expired error handling
- Comprehensive tracing logging (engine, model, duration, stderr)
---
 src/commands/tts.rs | 59 ++++++++++++++++++-----------------------------------
 1 file changed, 20 insertions(+), 39 deletions(-)

(limited to 'src/commands/tts.rs')

diff --git a/src/commands/tts.rs b/src/commands/tts.rs
index 4a99539..8d20179 100644
--- a/src/commands/tts.rs
+++ b/src/commands/tts.rs
@@ -3,6 +3,7 @@ use poise::serenity_prelude as serenity;
 use crate::Data;
 use crate::error::{Error, Result};
 use crate::lang;
+use tracing::warn;
 
 type Context<'a> = poise::Context<'a, Data, Error>;
 
@@ -14,7 +15,7 @@ pub enum OutputMode {
     Vc,
 }
 
-/// Generate Google TTS (Polish) and output to file or voice channel
+/// Generate TTS (Piper locally, Google as fallback) and output to file or voice channel
 #[poise::command(slash_command)]
 pub async fn tts(
     ctx: Context<'_>,
@@ -22,23 +23,10 @@ pub async fn tts(
     #[description = "Output destination (file or vc)"] output: Option<OutputMode>,
 ) -> Result<()> {
     ctx.defer().await?;
-    let lang = "pl";
-    let tts_url = format!(
-        "https://translate.google.com/translate_tts?ie=UTF-8&q={}&tl={}&client=tw-ob",
-        urlencoding(&text),
-        lang
-    );
 
     let output = output.unwrap_or(OutputMode::Vc);
-
-    let audio_data = ctx
-        .data()
-        .http_client
-        .get(&tts_url)
-        .send()
-        .await?
-        .bytes()
-        .await?;
+    let (audio_data, ext) =
+        crate::tts::synthesize(&ctx.data().config, &ctx.data().http_client, &text).await?;
 
     match output {
         OutputMode::File => {
@@ -46,17 +34,22 @@ pub async fn tts(
                 .await
                 .ok();
             let path = format!(
-                "{}/tts_{}.mp3",
+                "{}/tts_{}.{}",
                 crate::config::TEMP_PATH,
-                uuid::Uuid::new_v4()
+                uuid::Uuid::new_v4(),
+                ext
             );
             tokio::fs::write(&path, &audio_data).await?;
-            ctx.send(
-                poise::CreateReply::default()
-                    .content(lang::L_TTS_SAVED)
-                    .attachment(serenity::CreateAttachment::path(&path).await.unwrap()),
-            )
-            .await?;
+            if let Err(e) = ctx
+                .send(
+                    poise::CreateReply::default()
+                        .content(lang::L_TTS_SAVED)
+                        .attachment(serenity::CreateAttachment::path(&path).await.unwrap()),
+                )
+                .await
+            {
+                warn!("Failed to send /tts file reply: {e}");
+            }
         }
         OutputMode::Vc => {
             let guild_id = ctx.guild_id().ok_or_else(lang::err_not_in_guild)?;
@@ -81,24 +74,12 @@ pub async fn tts(
             if call_lock.current_channel().is_some() {
                 call_lock.play_input(input);
             }
-            ctx.say(lang::L_TTS_PLAYING).await?;
-        }
-    }
-    Ok(())
-}
-
-pub fn urlencoding(s: &str) -> String {
-    let mut out = String::with_capacity(s.len());
-    for &byte in s.as_bytes() {
-        match byte {
-            b'A'..=b'Z' | b'a'..=b'z' | b'0'..=b'9' | b'_' | b'-' | b'.' | b'~' => {
-                out.push(byte as char)
+            if let Err(e) = ctx.say(lang::L_TTS_PLAYING).await {
+                warn!("Failed to send /tts voice reply: {e}");
             }
-            b' ' => out.push('+'),
-            _ => out.push_str(&format!("%{:02X}", byte)),
         }
     }
-    out
+    Ok(())
 }
 
 pub fn commands() -> Vec<poise::Command<Data, Error>> {
-- 
cgit v1.2.3