fix(audio): use captions as the default YuE2 style (#12703)

* fix(audio): use captions as the default YuE2 style

Studio sends a caption, but YuE2 requires a style request option.
Default audio-cpp style from caption while preserving explicit options
and the caller's parameter map. Document the Studio fields.

Assisted-by: Codex:gpt-6
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>

* fix(audio-cpp): map captions to style in backend

Move the YuE2 caption fallback into the C++ request builder as requested.
Preserve explicit style options, including empty values, and move the
regression cases from the Go layer into the backend suite.

Assisted-by: Codex:GPT-6

---------

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Co-authored-by: localai-org-maint-bot <306269227+localai-org-maint-bot@users.noreply.github.com>

backend/cpp/audio-cpp/generation_request.cpp

@@ -232,9 +232,12 @@ build_sound_generation_request(const backend::SoundGenerationRequest &request,
         task.options["think"] = bool_option(request.think());
     }
     if (request.has_caption()) {
-        // Read only by irodori_tts, which advertises Tts/VoiceCloning/VoiceDesign
-        // and not AudioGeneration, so it is unreachable from this RPC: DEAD here.
         task.options["caption"] = request.caption();
+        // YuE2 calls the musical caption "style". Preserve explicit options,
+        // including an empty style, while accepting Studio's caption field.
+        if (!request.caption().empty()) {
+            task.options.emplace("style", request.caption());
+        }
     }
     if (request.has_lyrics()) {
         // Read by ace_step and heartmula.

backend/cpp/audio-cpp/generation_request_ctest.cpp

@@ -426,6 +426,42 @@ static void test_sound_generation_minimal() {
     check(!task.voice.has_value(), "sound: no voice condition");
 }
 
+static void test_sound_generation_caption_style() {
+    backend::SoundGenerationRequest request;
+    request.set_caption("slow acoustic folk");
+    request.set_lyrics("first line\nsecond line");
+
+    auto task = build_sound_generation_request(request, std::nullopt);
+    check(option_or(task.options, "style", "") == "slow acoustic folk",
+          "sound: caption supplies the default YuE2 style");
+    check(option_or(task.options, "caption", "") == "slow acoustic folk",
+          "sound: caption remains available under its own name");
+    check(option_or(task.options, "lyrics", "") == "first line\nsecond line",
+          "sound: style fallback preserves lyrics");
+    check(request.params().empty(), "sound: style fallback does not mutate params");
+
+    (*request.mutable_params())["style"] = "fast jazz";
+    task = build_sound_generation_request(request, std::nullopt);
+    check(option_or(task.options, "style", "") == "fast jazz",
+          "sound: explicit style overrides caption");
+
+    (*request.mutable_params())["style"] = "";
+    task = build_sound_generation_request(request, std::nullopt);
+    check(has_key(task.options, "style") && task.options.at("style").empty(),
+          "sound: explicit empty style overrides caption");
+
+    request.mutable_params()->clear();
+    request.set_caption("");
+    task = build_sound_generation_request(request, std::nullopt);
+    check(!has_key(task.options, "style"), "sound: empty caption supplies no style");
+
+    request.clear_caption();
+    (*request.mutable_params())["style"] = "instrumental jazz";
+    task = build_sound_generation_request(request, std::nullopt);
+    check(option_or(task.options, "style", "") == "instrumental jazz",
+          "sound: explicit style works without a caption");
+}
+
 static void test_sound_generation_full() {
     backend::SoundGenerationRequest request;
     request.set_text("a slow blues in E");
@@ -616,6 +652,7 @@ int main() {
     test_tts_language_and_params();
     test_tts_reference_transcript();
     test_sound_generation_minimal();
+    test_sound_generation_caption_style();
     test_sound_generation_full();
     test_transform_text_absent();
     test_transform_text_canonical_key();