refactor: enhance text-to-speech processing by splitting content into chunks and merging audio segments

Author: wxg0103

apps/application/flow/step_node/text_to_speech_step_node/impl/base_text_to_speech_node.py

@@ -6,9 +6,11 @@
 
 from application.flow.i_step_node import NodeResult
 from application.flow.step_node.text_to_speech_step_node.i_text_to_speech_node import ITextToSpeechNode
+from common.utils.common import _remove_empty_lines
 from knowledge.models import FileSourceType
 from models_provider.tools import get_model_instance_by_model_workspace_id
 from oss.serializers.file import FileSerializer
+from pydub import AudioSegment
 
 
 def bytes_to_uploaded_file(file_bytes, file_name="generated_audio.mp3"):
@@ -41,32 +43,72 @@ def save_context(self, details, workflow_manage):
 
     def execute(self, tts_model_id, chat_id,
                 content, model_params_setting=None,
-                **kwargs) -> NodeResult:
-        self.context['content'] = content
-        workspace_id = self.workflow_manage.get_body().get('workspace_id')
-        model = get_model_instance_by_model_workspace_id(tts_model_id, workspace_id,
-                                                         **model_params_setting)
-        audio_byte = model.text_to_speech(content)
-        # 需要把这个音频文件存储到数据库中
-        file_name = 'generated_audio.mp3'
-        file = bytes_to_uploaded_file(audio_byte, file_name)
+                max_length=1024, **kwargs) -> NodeResult:
+        # 分割文本为合理片段
+        content = _remove_empty_lines(content)
+        content_chunks = [content[i:i + max_length]
+                          for i in range(0, len(content), max_length)]
+
+        # 生成并收集所有音频片段
+        audio_segments = []
+        temp_files = []
+
+        for i, chunk in enumerate(content_chunks):
+            self.context['content'] = chunk
+            workspace_id = self.workflow_manage.get_body().get('workspace_id')
+            model = get_model_instance_by_model_workspace_id(
+                tts_model_id, workspace_id, **model_params_setting)
+
+            audio_byte = model.text_to_speech(chunk)
+
+            # 保存为临时音频文件用于合并
+            temp_file = io.BytesIO(audio_byte)
+            audio_segment = AudioSegment.from_file(temp_file)
+            audio_segments.append(audio_segment)
+            temp_files.append(temp_file)
+
+        # 合并所有音频片段
+        combined_audio = AudioSegment.empty()
+        for segment in audio_segments:
+            combined_audio += segment
+
+        # 将合并后的音频转为字节流
+        output_buffer = io.BytesIO()
+        combined_audio.export(output_buffer, format="mp3")
+        combined_bytes = output_buffer.getvalue()
+
+        # 存储合并后的音频文件
+        file_name = 'combined_audio.mp3'
+        file = bytes_to_uploaded_file(combined_bytes, file_name)
+
         application = self.workflow_manage.work_flow_post_handler.chat_info.application
         meta = {
             'debug': False if application.id else True,
             'chat_id': chat_id,
             'application_id': str(application.id) if application.id else None,
         }
+
         file_url = FileSerializer(data={
             'file': file,
             'meta': meta,
             'source_id': meta['application_id'],
             'source_type': FileSourceType.APPLICATION.value
         }).upload()
-        # 拼接一个audio标签的src属性
-        audio_label = f'<audio src="{file_url}" controls style = "width: 300px; height: 43px"></audio>'
+
+        # 生成音频标签
+        audio_label = f'<audio src="{file_url}" controls style="width: 300px; height: 43px"></audio>'
         file_id = file_url.split('/')[-1]
         audio_list = [{'file_id': file_id, 'file_name': file_name, 'url': file_url}]
-        return NodeResult({'answer': audio_label, 'result': audio_list}, {})
+
+        # 关闭所有临时文件
+        for temp_file in temp_files:
+            temp_file.close()
+        output_buffer.close()
+
+        return NodeResult({
+            'answer': audio_label,
+            'result': audio_list
+        }, {})
 
     def get_details(self, index: int, **kwargs):
         return {

apps/models_provider/impl/volcanic_engine_model_provider/model/tts.py

@@ -14,6 +14,8 @@
 import json
 import re
 import ssl
+
+import requests
 import uuid_utils.compat as uuid
 from typing import Dict