文案模式增加了一组文案配置，也就是说有2块文案，会进行左右切换的播放，都播放完毕后会进行新的一轮播放

hengeldev · Jul 20, 2023 · b53d11f · b53d11f
1 parent 7c5ca16
commit b53d11f
Show file tree

Hide file tree

Showing 12 changed files with 466 additions and 268 deletions.
diff --git a/README.md b/README.md
@@ -716,6 +716,15 @@ out_audio, out_sr = svc_model.infer(spk, tran, raw_path)
 out_audio, out_sr, n_frames = svc_model.infer(spk, tran, raw_path)
 ```
 
+#### 5.RuntimeWarning: Couldn't find ffmpeg or avconv - defaulting to ffmpeg, but may not work warn("Couldn't find ffmpeg or avconv - defaulting to ffmpeg, but may not work", RuntimeWarning)
+问题：pydub依赖ffmpeg，但环境不存在ffmpeg  
+解决方案：安装ffmpeg并配置环境变量  
+ffmpeg安装教程：[ffmpeg安装教程](https://blog.csdn.net/HYEHYEHYE/article/details/122000352)  
+ffmpeg官方下载：[http://ffmpeg.org/download.html](http://ffmpeg.org/download.html)  
+ffmpeg本仓库内下载：[releases](https://github.com/Ikaros-521/AI-Vtuber/releases/tag/v0.13.0)  
+
+简述就是下载，解压，配置ffmpeg的bin到系统环境变量path中，完事。  
+
 </details>
 
 
@@ -919,6 +928,10 @@ cmd输入命令即可：`doctoc /path/to/file`
 - [x] 监听1秒2弹幕的直播间，复读机配合vits-fast做语音合成，持续工作30分钟，正常监听并合成音频。
 
 ## 🖹待办事项
+- [ ] chatglm拥有记忆功能
+- [ ] 文案字幕显示
+- [ ] 礼物答谢combo情况下的重复答谢问题
+- [ ] LLM输出后，后置的过滤（双重过滤）
 - [ ] 懒人包优化
 - [ ] 接入更多的LLM（科大讯飞、new bing等）
 - [ ] 接入更多的TTS方案
@@ -931,7 +944,7 @@ cmd输入命令即可：`doctoc /path/to/file`
 - [ ] 本机语音直接交流模式（本地版本whisper，其他方式）
 - [ ] 本地的用户拉黑机制
 - [ ] 礼物互动机制（跳舞，唱歌）
-- [ ] 用户入场、投食感谢（自定义 文案）
+- [x] 用户入场、投食感谢（自定义 文案）
 - [ ] 弹幕跳过机制
 - [ ] 点歌后的弹幕触发问题（如：只取最新的几个）
 - [ ] 弹幕丢失bug修复（很可能是线程sleep造成的）
@@ -1097,6 +1110,8 @@ cmd输入命令即可：`doctoc /path/to/file`
 - 优化audio中vits接口的设计&优化代码减少冗余
 - 新增本地问答音频库，机制类似点歌，通过编辑音频文件名做为触发匹配关键词，优先级第二高，匹配命中后直接会进行匹配音频的播放。
 - 新增自定义上下限音频随机变速，对普通音频播放和文案音频播放可以进行随机变速配置。(GUI未适配)
+- 增加ffmpeg缺失的解决方案
+- 文案模式增加了一组文案配置，也就是说有2块文案，会进行左右切换的播放，都播放完毕后会进行新的一轮播放。
 
 </details>
 

diff --git a/UI_main.py b/UI_main.py
diff --git a/config.json b/config.json
@@ -170,19 +170,16 @@
   },
   "copywriting": {
     "file_path": "data/copywriting/",
+    "file_path2": "data/copywriting2/",
     "audio_path": "out/copywriting/",
+    "audio_path2": "out/copywriting2/",
     "play_list": [
       "伊卡日语介绍.wav",
       "吐槽.wav"
     ],
-    "audio_interval": 5.0,
-    "switching_interval": 5.0,
-    "random_play": true
-  },
-  "copywriting2": {
-    "file_path": "data/copywriting2/",
-    "audio_path": "out/copywriting2/",
-    "play_list": [],
+    "play_list2": [
+      "test.wav"
+    ],
     "audio_interval": 5.0,
     "switching_interval": 5.0,
     "random_play": true

diff --git a/config.json.bak b/config.json.bak
@@ -170,19 +170,16 @@
   },
   "copywriting": {
     "file_path": "data/copywriting/",
+    "file_path2": "data/copywriting2/",
     "audio_path": "out/copywriting/",
+    "audio_path2": "out/copywriting2/",
     "play_list": [
       "伊卡日语介绍.wav",
       "吐槽.wav"
     ],
-    "audio_interval": 5.0,
-    "switching_interval": 5.0,
-    "random_play": true
-  },
-  "copywriting2": {
-    "file_path": "data/copywriting2/",
-    "audio_path": "out/copywriting2/",
-    "play_list": [],
+    "play_list2": [
+      "test.wav"
+    ],
     "audio_interval": 5.0,
     "switching_interval": 5.0,
     "random_play": true

diff --git a/data/copywriting2/test.txt b/data/copywriting2/test.txt
@@ -0,0 +1,2 @@
+随便测一个看看。
+第二句话
diff --git a/data/copywriting2/test2.txt b/data/copywriting2/test2.txt
@@ -0,0 +1,2 @@
+随便合成几句。
+测试用
diff --git a/main.py b/main.py
@@ -391,15 +391,19 @@ def init_config(self):
 
             # 文案
             self.ui.label_copywriting_file_path.setToolTip("文案文件存储路径，默认不可编辑。不建议更改。")
+            self.ui.label_copywriting_file_path2.setToolTip("文案文件存储路径2，默认不可编辑。不建议更改。")
             self.ui.label_copywriting_audio_path.setToolTip("文案音频文件存储路径，默认不可编辑。不建议更改。")
+            self.ui.label_copywriting_audio_path2.setToolTip("文案音频文件存储路径2，默认不可编辑。不建议更改。")
             self.ui.label_copywriting_audio_interval.setToolTip("文案音频播放之间的间隔时间。就是前一个文案播放完成后，到后一个文案开始播放之间的间隔时间。")
             self.ui.label_copywriting_switching_interval.setToolTip("文案音频切换到弹幕音频的切换间隔时间（反之一样）。\n就是在播放文案时，有弹幕触发并合成完毕，此时会暂停文案播放，然后等待这个间隔时间后，再播放弹幕回复音频。")
             self.ui.label_copywriting_switching_random_play.setToolTip("文案随机播放，就是不根据播放音频文件列表的顺序播放，而是随机打乱顺序进行播放。")
             self.ui.label_copywriting_list.setToolTip("加载配置文件中配置的文案路径（data/copywriting/）下的所有文件，请勿放入其他非文案文件")
             self.ui.label_copywriting_play_list.setToolTip("此处填写需要播放的音频文件全名，填写完毕后点击 保存配置。文件全名从音频列表中复制，换行分隔，请勿随意填写")
-            self.ui.label_copywriting_audio_list.setToolTip("加载配置文件中配置的音频路径（data/audio_path/）下的所有文件，请勿放入其他非音频文件")
+            self.ui.label_copywriting_play_list2.setToolTip("此处填写需要播放的音频文件全名，填写完毕后点击 保存配置。文件全名从音频列表2中复制，换行分隔，请勿随意填写")
+            self.ui.label_copywriting_audio_list.setToolTip("加载配置文件中配置的音频路径（out/copywriting/）下的所有文件，请勿放入其他非音频文件")
+            self.ui.label_copywriting_audio_list2.setToolTip("加载配置文件中配置的音频路径2（out/copywriting2/）下的所有文件，请勿放入其他非音频文件")
             self.ui.label_copywriting_select.setToolTip("输入要加载的文案文件全名，文件全名从文案列表中复制。如果文件不存在，则会自动创建")
-            self.ui.pushButton_copywriting_select.setToolTip("加载 左侧输入框中的文件内容输出到下方编辑框内。如果文件不存在，则会自动创建")
+            self.ui.pushButton_copywriting_select.setToolTip("加载 左侧输入框中的文件相对/绝对路径的文件内容，输出到下方编辑框内。如果文件不存在，则会自动创建")
             self.ui.pushButton_copywriting_refresh_list.setToolTip("刷新 文案列表、音频列表中的内容，用于加载新数据")
             self.ui.label_copywriting_edit.setToolTip("此处由上方 选择的文案通过加载读取文件内容，在此可以修改文案内容，方便重新合成")
             self.ui.pushButton_copywriting_save.setToolTip("保存上方 文案编辑框中的内容到文案文件中")
@@ -732,7 +736,9 @@ def init_config(self):
 
             # 文案 刷新列表
             self.ui.lineEdit_copywriting_file_path.setText(self.copywriting_config['file_path'])
+            self.ui.lineEdit_copywriting_file_path2.setText(self.copywriting_config['file_path2'])
             self.ui.lineEdit_copywriting_audio_path.setText(self.copywriting_config['audio_path'])
+            self.ui.lineEdit_copywriting_audio_path2.setText(self.copywriting_config['audio_path2'])
             self.ui.lineEdit_copywriting_audio_interval.setText(str(self.copywriting_config['audio_interval']))
             self.ui.lineEdit_copywriting_switching_interval.setText(str(self.copywriting_config['switching_interval']))
             if self.copywriting_config['random_play']:
@@ -742,6 +748,10 @@ def init_config(self):
             for tmp in self.copywriting_config['play_list']:
                 tmp_str = tmp_str + tmp + "\n"
             self.ui.textEdit_copywriting_play_list.setText(tmp_str)
+            tmp_str = ""
+            for tmp in self.copywriting_config['play_list2']:
+                tmp_str = tmp_str + tmp + "\n"
+            self.ui.textEdit_copywriting_play_list2.setText(tmp_str)
 
             # 聊天
             self.ui.lineEdit_talk_username.setText(self.talk_config['username'])
@@ -1181,6 +1191,11 @@ def save(self):
             if 0 != len(play_list):
                 play_list = play_list[1:]
             config_data["copywriting"]["play_list"] = play_list
+            copywriting_play_list = self.ui.textEdit_copywriting_play_list2.toPlainText()
+            play_list = [token.strip() for separator in separators for part in copywriting_play_list.split(separator) if (token := part.strip())]
+            if 0 != len(play_list):
+                play_list = play_list[1:]
+            config_data["copywriting"]["play_list2"] = play_list
 
             config_data["header"]["userAgent"] = self.ui.lineEdit_header_useragent.text()
 
@@ -1331,8 +1346,9 @@ def copywriting_select(self):
             self.show_message_box("警告", "请输入 文案路径喵~", QMessageBox.Critical, 3000)
             return
 
-        logging.info(f"准备加载 {self.copywriting_config['file_path']} 路径下的 [{select_file_path}]")
-        new_file_path = os.path.join(self.copywriting_config['file_path'], select_file_path)
+        # 传入完整文件路径 绝对或相对
+        logging.info(f"准备加载 文件：[{select_file_path}]")
+        new_file_path = os.path.join(select_file_path)
 
         content = common.read_file_return_content(new_file_path)
         if content is None:
@@ -1362,13 +1378,23 @@ def copywriting_refresh_list(self):
             for tmp in copywriting_file_names:
                 tmp_str = tmp_str + tmp + "\n"
             self.ui.textEdit_copywriting_list.setText(tmp_str)
+            tmp_str = ""
+            copywriting_file_names = self.get_dir_txt_filename(self.copywriting_config['file_path2'])
+            for tmp in copywriting_file_names:
+                tmp_str = tmp_str + tmp + "\n"
+            self.ui.textEdit_copywriting_list2.setText(tmp_str)
 
             # 文案音频数据回显到UI
             tmp_str = ""
             copywriting_audio_file_names = self.get_dir_audio_filename(self.copywriting_config['audio_path'])
             for tmp in copywriting_audio_file_names:
                 tmp_str = tmp_str + tmp + "\n"
             self.ui.textEdit_copywriting_audio_list.setText(tmp_str)
+            tmp_str = ""
+            copywriting_audio_file_names = self.get_dir_audio_filename(self.copywriting_config['audio_path2'])
+            for tmp in copywriting_audio_file_names:
+                tmp_str = tmp_str + tmp + "\n"
+            self.ui.textEdit_copywriting_audio_list2.setText(tmp_str)
 
             logging.info("刷新文件列表")
         except Exception as e:
@@ -1389,7 +1415,7 @@ def copywriting_save(self):
             logging.warning(f"请输入 文案路径喵~")
             return
 
-        new_file_path = os.path.join(self.copywriting_config['file_path'], select_file_path)
+        new_file_path = os.path.join(select_file_path)
         if True == common.write_content_to_file(new_file_path, content):
             self.show_message_box("提示", "保存成功~", QMessageBox.Information, 3000)
         else:
@@ -1404,7 +1430,20 @@ def on_pushButton_copywriting_save_clicked(self):
     # 合成音频 正经版
     async def copywriting_synthetic_audio_main(self):
         select_file_path = self.ui.lineEdit_copywriting_select.text()
-        ret = await audio.copywriting_synthesis_audio(select_file_path)
+
+        # 将一个文件路径的字符串切分成路径和文件名
+        folder_path, file_name = common.split_path_and_filename(select_file_path)
+
+        # 判断文件是哪个路径的
+        if folder_path == config.get("copywriting", "file_path"):
+            folder_path = config.get("copywriting", "audio_path")
+        elif folder_path == config.get("copywriting", "file_path2"):
+            folder_path = config.get("copywriting", "audio_path2")
+        else:
+            logging.error(f"文件路径与配置不匹配，将默认输出到同路径")
+            self.show_message_box("提示", f"文件路径与配置不匹配，将默认输出到同路径", QMessageBox.Information, 3000)
+
+        ret = await audio.copywriting_synthesis_audio(select_file_path, folder_path)
         if ret is None:
             logging.error(f"合成失败！请排查原因")
             self.show_message_box("错误", "合成失败！请排查原因", QMessageBox.Critical)

diff --git a/out/copywriting2/test.wav b/out/copywriting2/test.wav
diff --git a/out/copywriting2/test2.wav b/out/copywriting2/test2.wav