quicksandzn
66 lines
2.6 KiB
Python
66 lines
2.6 KiB
Python
from collections.abc import Generator
|
|
from typing import Any
|
|
from dify_plugin import Tool
|
|
from dify_plugin.entities.tool import ToolInvokeMessage
|
|
from tools.base import MiniMaxBaseTool
|
|
|
|
|
|
class MiniMaxMusicGenerationTool(Tool):
|
|
def _invoke(self, tool_parameters: dict[str, Any]) -> Generator[ToolInvokeMessage]:
|
|
api_key = self.runtime.credentials.get("api_key")
|
|
group_id = self.runtime.credentials.get("group_id")
|
|
minimax = MiniMaxBaseTool(api_key=api_key, group_id=group_id)
|
|
|
|
model = tool_parameters.get("model")
|
|
song_file = tool_parameters.get("song")
|
|
refer_vocal = tool_parameters.get("refer_vocal")
|
|
lyrics = tool_parameters.get("lyrics")
|
|
|
|
upload_response = minimax.music_upload(
|
|
file_name=song_file.filename,
|
|
file_blob=song_file.blob,
|
|
mime_type=song_file.mime_type,
|
|
)
|
|
if upload_response.status_code != 200:
|
|
yield self.create_text_message(
|
|
f"Music generation upload failed {upload_response.status_code} {upload_response.text}"
|
|
)
|
|
return
|
|
upload_data = upload_response.json()
|
|
voice_id = upload_data.get("voice_id")
|
|
instrumental_id = upload_data.get("instrumental_id")
|
|
if not voice_id or not instrumental_id:
|
|
yield self.create_text_message(
|
|
f"Music generation upload failed {upload_response.text}"
|
|
)
|
|
return
|
|
gen_response = minimax.music_generation(
|
|
model=model,
|
|
refer_voice=voice_id,
|
|
refer_instrumental=instrumental_id,
|
|
refer_vocal=None if not refer_vocal else refer_vocal,
|
|
lyrics=lyrics,
|
|
)
|
|
if gen_response.status_code != 200:
|
|
yield self.create_text_message(
|
|
f"Music generation failed {gen_response.status_code} {gen_response.text}"
|
|
)
|
|
return
|
|
status_code = gen_response.json().get("base_resp", {}).get("status_code", -1)
|
|
if status_code != 0:
|
|
yield self.create_text_message(
|
|
f"Music generation failed {gen_response.text}"
|
|
)
|
|
return
|
|
audio_hex = gen_response.json().get("data", {}).get("audio")
|
|
|
|
if not audio_hex:
|
|
yield self.create_text_message(
|
|
f"Music generation failed {gen_response.text}"
|
|
)
|
|
return
|
|
(self.create_text_message("Audio generated successfully"),)
|
|
yield self.create_blob_message(
|
|
blob=bytes.fromhex(audio_hex), meta={"mime_type": "audio/mpeg"}
|
|
)
|