update: 增加相关配置

main/manager-api/src/main/resources/db/changelog/202508200911.sql

@@ -1,33 +0,0 @@
--- 添加 MinimaxHTTPStream 流式 TTS 供应器
-delete from `ai_model_provider` where id = 'SYSTEM_TTS_MinimaxStreamTTS';
-INSERT INTO `ai_model_provider` (`id`, `model_type`, `provider_code`, `name`, `fields`, `sort`, `creator`, `create_date`, `updater`, `update_date`) VALUES
-('SYSTEM_TTS_MinimaxStreamTTS', 'TTS', 'minimax_stream', 'Minimax流式语音合成', '[{"key":"group_id","label":"组ID","type":"string"},{"key":"api_key","label":"API密钥","type":"string"},{"key":"model","label":"模型","type":"string"},{"key":"voice_id","label":"音色ID","type":"string"},{"key":"output_dir","label":"输出目录","type":"string"},{"key":"voice_setting","label":"音色设置","type":"dict","dict_name":"voice_setting"},{"key":"pronunciation_dict","label":"发音字典","type":"dict","dict_name":"pronunciation_dict"},{"key":"audio_setting","label":"音频设置","type":"dict","dict_name":"audio_setting"},{"key":"timber_weights","label":"音色权重","type":"string"}]', 18, 1, NOW(), 1, NOW());
-
--- 添加Minimax流式TTS模型配置
-delete from `ai_model_config` where id = 'TTS_MinimaxStreamTTS';
-INSERT INTO `ai_model_config` VALUES ('TTS_MinimaxStreamTTS', 'TTS', 'MinimaxStreamTTS', 'Minimax流式语音合成', 0, 1, '{"type": "minimax_stream", "group_id": "", "api_key": "", "model": "speech-01", "voice_id": "female-shaonv", "output_dir": "tmp/", "voice_setting": {"speed": 1, "vol": 1, "pitch": 0, "emotion": "happy"}, "pronunciation_dict": {"tone": ["处理/(chu3)(li3)", "危险/dangerous"]}, "audio_setting": {"sample_rate": 24000, "bitrate": 128000, "format": "pcm", "channel": 1}}', NULL, NULL, 21, NULL, NULL, NULL, NULL);
-
--- 更新Minimax流式TTS配置说明
-UPDATE `ai_model_config` SET 
-`doc_link` = 'https://platform.minimaxi.com/',
-`remark` = 'Minimax流式TTS配置说明：
-1. 需要先申请Minimax API Key
-2. 需要填写Group ID
-3. 支持多种音色设置和音频参数调整
-4. 支持实时流式合成，具有较低的延迟
-5. 支持自定义发音字典和音色权重
-6. 隐藏参数配置：声音设定(voice_setting)、发音字典(pronunciation_dict)、音色权重(timber_weights)
-   - 语速(speed): 范围[0.5,2]，默认1.0，取值越大语速越快
-   - 音量(vol): 范围(0,10]，默认1.0，取值越大音量越高
-   - 音调(pitch): 范围[-12,12]，默认0，取值需为整数
-   - timbre_weights与voice_id二选一必填
-   - voice_id(请求的音色id，须和weight参数同步填写)
-   - weight(权重，最多支持4种音色混合。范围[1,100])
-' WHERE `id` = 'TTS_MinimaxStreamTTS';
-
--- 添加Minimax流式TTS音色
-delete from `ai_tts_voice` where tts_model_id = 'TTS_MinimaxStreamTTS';
--- 默认音色
-INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0001', 'TTS_MinimaxStreamTTS', '少女音', 'female-shaonv', '中文', NULL, NULL, NULL, NULL, 1, NULL, NULL, NULL, NULL);
-INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0002', 'TTS_MinimaxStreamTTS', '成熟女声', 'female-chengshu', '中文', NULL, NULL, NULL, NULL, 2, NULL, NULL, NULL, NULL);
-INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0003', 'TTS_MinimaxStreamTTS', '少年音', 'male-shaonian', '中文', NULL, NULL, NULL, NULL, 3, NULL, NULL, NULL, NULL);

main/manager-api/src/main/resources/db/changelog/202509051745.sql

@@ -0,0 +1,45 @@
+-- 添加 MinimaxHTTPStream 流式 TTS 供应器
+delete from `ai_model_provider` where id = 'SYSTEM_TTS_MinimaxStreamTTS';
+INSERT INTO `ai_model_provider` (`id`, `model_type`, `provider_code`, `name`, `fields`, `sort`, `creator`, `create_date`, `updater`, `update_date`) VALUES
+('SYSTEM_TTS_MinimaxStreamTTS', 'TTS', 'minimax_httpstream', 'Minimax流式语音合成', '[{"key":"group_id","label":"组ID","type":"string"},{"key":"api_key","label":"API密钥","type":"string"},{"key":"model","label":"模型","type":"string"},{"key":"voice_id","label":"音色ID","type":"string"},{"key":"output_dir","label":"输出目录","type":"string"},{"key":"voice_setting","label":"音色设置","type":"dict","dict_name":"voice_setting"},{"key":"pronunciation_dict","label":"发音字典","type":"dict","dict_name":"pronunciation_dict"},{"key":"audio_setting","label":"音频设置","type":"dict","dict_name":"audio_setting"},{"key":"timber_weights","label":"音色权重","type":"string"}]', 18, 1, NOW(), 1, NOW());
+
+-- 添加Minimax流式TTS模型配置
+delete from `ai_model_config` where id = 'TTS_MinimaxStreamTTS';
+INSERT INTO `ai_model_config` VALUES ('TTS_MinimaxStreamTTS', 'TTS', 'MinimaxStreamTTS', 'Minimax流式语音合成', 0, 1, '{"type": "minimax_httpstream", "group_id": "", "api_key": "", "model": "speech-01-turbo", "voice_id": "female-shaonv", "output_dir": "tmp/", "voice_setting": {"speed": 1, "vol": 1, "pitch": 0, "emotion": "happy"}, "pronunciation_dict": {"tone": ["处理/(chu3)(li3)", "危险/dangerous"]}, "audio_setting": {"sample_rate": 24000, "bitrate": 128000, "format": "pcm", "channel": 1}}', NULL, NULL, 21, NULL, NULL, NULL, NULL);
+
+-- 更新Minimax流式TTS配置说明
+UPDATE `ai_model_config` SET 
+`doc_link` = 'https://platform.minimaxi.com/',
+`remark` = 'Minimax流式TTS配置说明：
+1. 需要先申请Minimax API Key
+2. 需要填写Group ID
+3. 支持多种音色设置和音频参数调整
+4. 支持实时流式合成，具有较低的延迟
+5. 支持自定义发音字典和音色权重
+6. 隐藏参数配置：声音设定(voice_setting)、发音字典(pronunciation_dict)、音色权重(timber_weights)
+   - 语速(speed): 范围[0.5,2]，默认1.0，取值越大语速越快
+   - 音量(vol): 范围(0,10]，默认1.0，取值越大音量越高
+   - 音调(pitch): 范围[-12,12]，默认0，取值需为整数
+   - 情绪(emotion): 控制合成语音的情绪，支持7种值：["happy", "sad", "angry", "fearful", "disgusted", "surprised", "calm"]，该参数仅对 speech-2.5-hd-preview、speech-2.5-turbo-preview、speech-02-hd、speech-02-turbo、speech-01-turbo、speech-01-hd 生效
+   - timbre_weights与voice_id二选一必填
+   - voice_id(请求的音色id，须和weight参数同步填写)
+   - weight(权重，最多支持4种音色混合。范围[1,100])
+' WHERE `id` = 'TTS_MinimaxStreamTTS';
+
+-- 添加Minimax流式TTS音色
+delete from `ai_tts_voice` where tts_model_id = 'TTS_MinimaxStreamTTS';
+
+-- 默认音色
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0001', 'TTS_MinimaxStreamTTS', '少女音', 'female-shaonv', '中文', NULL, NULL, NULL, NULL, 1, NULL, NULL, NULL, NULL);
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0002', 'TTS_MinimaxStreamTTS', '成熟女声', 'female-chengshu', '中文', NULL, NULL, NULL, NULL, 2, NULL, NULL, NULL, NULL);
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0003', 'TTS_MinimaxStreamTTS', '霸道少爷', 'badao_shaoye', '中文', NULL, NULL, NULL, NULL, 3, NULL, NULL, NULL, NULL);
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0004', 'TTS_MinimaxStreamTTS', '病娇弟弟', 'bingjiao_didi', '中文', NULL, NULL, NULL, NULL, 4, NULL, NULL, NULL, NULL);
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0005', 'TTS_MinimaxStreamTTS', '纯真学弟', 'chunzhen_xuedi', '中文', NULL, NULL, NULL, NULL, 5, NULL, NULL, NULL, NULL);
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0006', 'TTS_MinimaxStreamTTS', '冷淡学长', 'lengdan_xiongzhang', '中文', NULL, NULL, NULL, NULL, 6, NULL, NULL, NULL, NULL);
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0007', 'TTS_MinimaxStreamTTS', '甜美小玲', 'tianxin_xiaoling', '中文', NULL, NULL, NULL, NULL, 7, NULL, NULL, NULL, NULL);
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0008', 'TTS_MinimaxStreamTTS', '俏皮萌妹', 'qiaopi_mengmei', '中文', NULL, NULL, NULL, NULL, 8, NULL, NULL, NULL, NULL);
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0009', 'TTS_MinimaxStreamTTS', '妩媚御姐', 'wumei_yujie', '中文', NULL, NULL, NULL, NULL, 9, NULL, NULL, NULL, NULL);
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0010', 'TTS_MinimaxStreamTTS', '嗲嗲学妹', 'diadia_xuemei', '中文', NULL, NULL, NULL, NULL, 7, NULL, NULL, NULL, NULL);
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0011', 'TTS_MinimaxStreamTTS', '淡雅学姐', 'danya_xuejie', '中文', NULL, NULL, NULL, NULL, 8, NULL, NULL, NULL, NULL);
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0012', 'TTS_MinimaxStreamTTS', 'Santa Claus', 'Santa_Claus', '中文', NULL, NULL, NULL, NULL, 9, NULL, NULL, NULL, NULL);
+INSERT INTO `ai_tts_voice` VALUES ('TTS_MinimaxStreamTTS_0013', 'TTS_MinimaxStreamTTS', 'Grinch', 'Grinch', '中文', NULL, NULL, NULL, NULL, 10, NULL, NULL, NULL, NULL);

main/manager-api/src/main/resources/db/changelog/db.changelog-master.yaml

@@ -304,9 +304,9 @@ databaseChangeLog:
             encoding: utf8
             path: classpath:db/changelog/202508131557.sql
   - changeSet:
-      id: 202508200911
+      id: 202509051745
       author: RanChen
       changes:
         - sqlFile:
             encoding: utf8
-            path: classpath:db/changelog/202508200911.sql
+            path: classpath:db/changelog/202509051745.sql

main/xiaozhi-server/core/providers/tts/minimax_httpstream.py

@@ -1,12 +1,14 @@
 import os
 import json
+import time
 import queue
 import asyncio
-import traceback
 import aiohttp
-from core.utils.util import parse_string_to_list
+import requests
+import traceback
 from config.logger import setup_logging
 from core.utils.tts import MarkdownCleaner
+from core.utils.util import parse_string_to_list
 from core.providers.tts.base import TTSProviderBase
 from core.utils import opus_encoder_utils, textUtils
 from core.providers.tts.dto.dto import SentenceType, ContentType
@@ -80,7 +82,6 @@ class TTSProvider(TTSProviderBase):
                     self.processed_chars = 0
                     self.tts_text_buff = []
                     self.before_stop_play_files.clear()
-                    self.reset_flow_controller()
                 elif ContentType.TEXT == message.content_type:
                     self.tts_text_buff.append(message.content_detail)
                     segment_text = self._get_segment_text()
@@ -93,13 +94,7 @@ class TTSProvider(TTSProviderBase):
                     )
                     if message.content_file and os.path.exists(message.content_file):
                         # 先处理文件音频数据
-                        self._process_audio_file_stream(
+                        self._process_audio_file_stream(message.content_file, callback=lambda audio_data: self.handle_audio_file(audio_data, message.content_detail))
-                            message.content_file,
-                            callback=lambda audio_data: self.handle_audio_file(
-                                audio_data, message.content_detail
-                            ),
-                        )
-
                 if message.sentence_type == SentenceType.LAST:
                     # 处理剩余的文本
                     self._process_remaining_text_stream(True)
@@ -259,3 +254,88 @@ class TTSProvider(TTSProviderBase):
         await super().close()
         if hasattr(self, "opus_encoder"):
             self.opus_encoder.close()
+
+    def to_tts(self, text: str) -> list:
+        """非流式TTS处理，用于测试及保存音频文件的场景
+        Args:
+            text: 要转换的文本
+        Returns:
+            list: 返回opus编码后的音频数据列表
+        """
+        start_time = time.time()
+        text = MarkdownCleaner.clean_markdown(text)
+
+        payload = {
+            "model": self.model,
+            "text": text,
+            "stream": True,
+            "voice_setting": self.voice_setting,
+            "pronunciation_dict": self.pronunciation_dict,
+            "audio_setting": self.audio_setting,
+        }
+
+        if type(self.timber_weights) is list and len(self.timber_weights) > 0:
+            payload["timber_weights"] = self.timber_weights
+            payload["voice_setting"]["voice_id"] = ""
+
+        headers = {
+            "Content-Type": "application/json",
+            "Authorization": f"Bearer {self.api_key}",
+        }
+
+        try:
+            with requests.post(
+                self.api_url, data=json.dumps(payload), headers=headers, timeout=5
+            ) as response:
+                if response.status_code != 200:
+                    logger.bind(tag=TAG).error(
+                        f"TTS请求失败: {response.status_code}, {response.text}"
+                    )
+                    return []
+
+                logger.info(f"TTS请求成功: {text}, 耗时: {time.time() - start_time}秒")
+
+                # 使用opus编码器处理PCM数据
+                opus_datas = []
+                full_content = response.content.decode('utf-8')
+                pcm_data = bytearray()
+                for data_block in full_content.split('\n\n'):
+                    if not data_block.startswith('data: '):
+                        continue
+
+                    try:
+                        json_str = data_block[6:]  # 去除'data: '前缀
+                        data = json.loads(json_str)
+                        if data.get('data', {}).get('status') == 1:
+                            audio_hex = data['data']['audio']
+                            pcm_data.extend(bytes.fromhex(audio_hex))
+                    except (json.JSONDecodeError, KeyError) as e:
+                        logger.bind(tag=TAG).warning(f"无效数据块: {e}")
+                        continue
+
+                # 计算每帧的字节数
+                frame_bytes = int(
+                    self.opus_encoder.sample_rate
+                    * self.opus_encoder.channels
+                    * self.opus_encoder.frame_size_ms
+                    / 1000
+                    * 2
+                )
+
+                # 分帧处理合并后的PCM数据
+                for i in range(0, len(pcm_data), frame_bytes):
+                    frame = bytes(pcm_data[i:i+frame_bytes])
+                    if len(frame) < frame_bytes:
+                        frame += b"\x00" * (frame_bytes - len(frame))
+ 
+                    self.opus_encoder.encode_pcm_to_opus_stream(
+                        frame,
+                        end_of_stream=(i + frame_bytes >= len(pcm_data)),
+                        callback=lambda opus: opus_datas.append(opus)
+                    )
+
+                return opus_datas
+
+        except Exception as e:
+            logger.bind(tag=TAG).error(f"TTS请求异常: {e}")
+            return []