feat: enhance audio model handling and improve image URL resolution

- Updated `PollinationsAI` to exclude "gemini" model from `audio_models`
- Added logic in `PollinationsAI` to expand `audio_models` with voices from `default_audio_model`
- Appended voice names to `text_models` list in `PollinationsAI` if present in `default_audio_model`
- Modified `PollinationsAI._generate_text` to inject `audio` parameters when a voice model is used
- Updated `save_response_media` call to include voice name in model list
- Changed `OpenaiChat.get_generated_image` to support both `file-service://` and `sediment://` URLs using `conversation_id`
- Modified `OpenaiChat.create_messages` to optionally pass `prompt`
- Adjusted `OpenaiChat.run` to determine `prompt` explicitly and set messages accordingly
- Updated `OpenaiChat.iter_messages_line` to handle `None` in `fields.p` safely
- Passed `prompt` and `conversation_id` to `OpenaiChat.get_generated_image` inside image parsing loop
- Fixed redirect logic in `backend_api.py` to safely handle missing `skip` query param
- Enhanced `render` function in `website.py` to support live file serving with `live` query param
- Added new route `/dist/<path:name>` to serve static files from `DIST_DIR` in `website.py`
- Adjusted `render` to include `.live` suffix in cache filename when applicable
- Modified HTML replacements in `render` to preserve local `dist/` path if `add_origion` is True

g4f/Provider/PollinationsAI.py

@@ -105,9 +105,12 @@ class PollinationsAI(AsyncGeneratorProvider, ProviderModelMixin):
                 cls.audio_models = {
                     model.get("name"): model.get("voices")
                     for model in models
-                    if "output_modalities" in model and "audio" in model["output_modalities"]
+                    if "output_modalities" in model and "audio" in model["output_modalities"] and model.get("name") != "gemini"
                 }
 
+                if cls.default_audio_model in cls.audio_models:
+                    cls.audio_models = {**cls.audio_models, **{voice: {} for voice in cls.audio_models[cls.default_audio_model]}}
+
                 # Create a set of unique text models starting with default model
                 unique_text_models = cls.text_models.copy()
 
@@ -120,6 +123,9 @@ class PollinationsAI(AsyncGeneratorProvider, ProviderModelMixin):
                     if model_name and "input_modalities" in model and "text" in model["input_modalities"]:
                         unique_text_models.append(model_name)
 
+                if cls.default_audio_model in cls.audio_models:
+                    unique_text_models.extend([voice for voice in cls.audio_models[cls.default_audio_model]])
+
                 # Convert to list and update text_models
                 cls.text_models = list(dict.fromkeys(unique_text_models))
 
@@ -207,6 +213,11 @@ class PollinationsAI(AsyncGeneratorProvider, ProviderModelMixin):
                     "role": "user",
                     "content": prompt
                 }]
+            if model and model in cls.audio_models[cls.default_audio_model]:
+                kwargs["audio"] = {
+                    "voice": model,
+                }
+                model = cls.default_audio_model
             async for result in cls._generate_text(
                 model=model,
                 messages=messages,
@@ -359,6 +370,6 @@ class PollinationsAI(AsyncGeneratorProvider, ProviderModelMixin):
                     if finish_reason:
                         yield FinishReason(finish_reason)
                 else:
-                    async for chunk in save_response_media(response, format_image_prompt(messages), [model]):
+                    async for chunk in save_response_media(response, format_image_prompt(messages), [model, extra_parameters.get("audio", {}).get("voice")]):
                         yield chunk
                         return