import gradio as gr from tools.i18n.i18n import I18nAuto i18n = I18nAuto() # 校验基础信息 def check_base_info(text_work_space_dir, text_character): if text_work_space_dir is None or text_work_space_dir == '': raise Exception(i18n("工作目录不能为空")) if text_character is None or text_character == '': raise Exception(i18n("角色名称不能为空")) # 从list文件，提取参考音频 def convert_from_list(text_work_space_dir, text_character, text_list_input): text_convert_from_list_info = "转换成功：生成目录XXX" text_sample_dir = "D：//tt" try: check_base_info(text_work_space_dir, text_character) if text_list_input is None or text_list_input == '': raise Exception(i18n("list文件路径不能为空")) pass except Exception as e: text_convert_from_list_info = f"发生异常：{e}" text_sample_dir = '' return [text_convert_from_list_info, text_sample_dir] # 基于一个基准音频，从参考音频目录中进行分段抽样 def sample(text_work_space_dir, text_character, text_sample_dir, text_base_voice_path, text_subsection_num, text_sample_num, checkbox_similarity_output): text_sample_info = "抽样成功：生成目录XXX" ref_audio_dir = "D：//tt" try: check_base_info(text_work_space_dir, text_character) if text_sample_dir is None or text_sample_dir == '': raise Exception(i18n("参考音频抽样目录不能为空，请先完成上一步操作")) if text_base_voice_path is None or text_base_voice_path == '': raise Exception(i18n("基准音频路径不能为空")) if text_subsection_num is None or text_subsection_num == '': raise Exception(i18n("分段数不能为空")) if text_sample_num is None or text_sample_num == '': raise Exception(i18n("每段随机抽样个数不能为空")) pass except Exception as e: text_sample_info = f"发生异常：{e}" ref_audio_dir = '' text_model_inference_voice_dir = ref_audio_dir text_sync_ref_audio_dir = ref_audio_dir text_sync_ref_audio_dir2 = ref_audio_dir return [text_sample_info, text_model_inference_voice_dir, text_sync_ref_audio_dir, text_sync_ref_audio_dir2] # 根据参考音频和测试文本，执行批量推理 def model_inference(text_work_space_dir, text_character, text_model_inference_voice_dir, text_url, text_text, text_ref_path, text_ref_text, text_emotion, text_test_content): text_model_inference_info = "推理成功：生成目录XXX" text_asr_audio_dir = "D：//tt" try: check_base_info(text_work_space_dir, text_character) if text_model_inference_voice_dir is None or text_model_inference_voice_dir == '': raise Exception(i18n("待推理的参考音频所在目录不能为空，请先完成上一步操作")) if text_url is None or text_url == '': raise Exception(i18n("推理服务请求地址不能为空")) if text_text is None or text_text == '': raise Exception(i18n("文本参数名不能为空")) if text_test_content is None or text_test_content == '': raise Exception(i18n("待推理文本路径不能为空")) if (text_ref_path is None or text_ref_path == '') and (text_ref_text is None or text_ref_text == '') and (text_emotion is None or text_emotion == ''): raise Exception(i18n("参考音频路径/文本和角色情绪二选一填写，不能全部为空")) pass except Exception as e: text_model_inference_info = f"发生异常：{e}" text_asr_audio_dir = '' return [text_model_inference_info, text_asr_audio_dir] # 对推理生成音频执行asr def asr(text_work_space_dir, text_character, text_asr_audio_dir, dropdown_asr_model, dropdown_asr_size, dropdown_asr_lang): text_asr_info = "asr成功：生成目录XXX" text_text_similarity_analysis_path = "D：//tt" try: check_base_info(text_work_space_dir, text_character) if text_asr_audio_dir is None or text_asr_audio_dir == '': raise Exception(i18n("待asr的音频所在目录不能为空，请先完成上一步操作")) if dropdown_asr_model is None or dropdown_asr_model == '': raise Exception(i18n("asr模型不能为空")) if dropdown_asr_size is None or dropdown_asr_size == '': raise Exception(i18n("asr模型大小不能为空")) if dropdown_asr_lang is None or dropdown_asr_lang == '': raise Exception(i18n("asr语言不能为空")) pass except Exception as e: text_asr_info = f"发生异常：{e}" text_text_similarity_analysis_path = '' return [text_asr_info, text_text_similarity_analysis_path] # 对asr生成的文件，与原本的文本内容，进行相似度分析 def text_similarity_analysis(text_work_space_dir, text_character, text_text_similarity_analysis_path): text_text_similarity_analysis_info = "相似度分析成功：生成目录XXX" try: check_base_info(text_work_space_dir, text_character) if text_text_similarity_analysis_path is None or text_text_similarity_analysis_path == '': raise Exception(i18n("asr生成的文件路径不能为空，请先完成上一步操作")) pass except Exception as e: text_text_similarity_analysis_info = f"发生异常：{e}" return text_text_similarity_analysis_info # 根据一个参考音频，对指定目录下的音频进行相似度分析，并输出到另一个目录 def similarity_audio_output(text_work_space_dir, text_character, text_base_audio_path, text_compare_audio_dir): text_similarity_audio_output_info = "相似度分析成功：生成目录XXX" try: check_base_info(text_work_space_dir, text_character) if text_base_audio_path is None or text_base_audio_path == '': raise Exception(i18n("基准音频路径不能为空")) if text_compare_audio_dir is None or text_compare_audio_dir == '': raise Exception(i18n("待分析的音频所在目录不能为空")) pass except Exception as e: text_similarity_audio_output_info = f"发生异常：{e}" return text_similarity_audio_output_info # 根据参考音频目录的删除情况，将其同步到推理生成的音频目录中，即参考音频目录下，删除了几个参考音频，就在推理目录下，将这些参考音频生成的音频文件移除 def sync_ref_audio(text_work_space_dir, text_character, text_sync_ref_audio_dir, text_sync_inference_audio_dir): text_sync_ref_audio_info = "同步参考音频成功：生成目录XXX" try: check_base_info(text_work_space_dir, text_character) if text_sync_ref_audio_dir is None or text_sync_ref_audio_dir == '': raise Exception(i18n("参考音频目录不能为空")) if text_sync_inference_audio_dir is None or text_sync_inference_audio_dir == '': raise Exception(i18n("推理生成的音频目录不能为空")) pass except Exception as e: text_sync_ref_audio_info = f"发生异常：{e}" return text_sync_ref_audio_info # 根据模板和参考音频目录，生成参考音频配置内容 def create_config(text_work_space_dir, text_character, text_template, text_sync_ref_audio_dir2): text_create_config_info = "配置生成成功：生成目录XXX" try: check_base_info(text_work_space_dir, text_character) if text_template is None or text_template == '': raise Exception(i18n("参考音频抽样目录不能为空")) if text_sync_ref_audio_dir2 is None or text_sync_ref_audio_dir2 == '': raise Exception(i18n("参考音频目录不能为空")) pass except Exception as e: text_create_config_info = f"发生异常：{e}" return text_create_config_info # 基于请求路径和参数，合成完整的请求路径 def whole_url(text_url, text_text, text_ref_path, text_ref_text, text_emotion): text_whole_url = f'{text_url}?{text_text}=文本内容&{text_ref_path}=参考音频路径&{text_ref_text}=参考文本&{text_emotion}=情绪类型' return [text_whole_url] with gr.Blocks() as app: gr.Markdown(value=i18n("基本介绍：这是一个从训练素材中，批量提取参考音频，并进行效果评估与配置生成的工具")) with gr.Row(): text_work_space_dir = gr.Text(label=i18n("工作目录，后续操作所生成文件都会保存在此目录下"), value="") text_character = gr.Text(label=i18n("请输入角色名称"), value="") with gr.Accordion(label=i18n("第一步：基于训练素材，生成待选参考音频列表"), open=False): gr.Markdown(value=i18n("1.1：选择list文件，并提取3-10秒的素材作为参考候选")) text_list_input = gr.Text(label=i18n("请输入list文件路径"), value="") with gr.Row(): button_convert_from_list = gr.Button(i18n("开始生成待参考列表"), variant="primary") text_convert_from_list_info = gr.Text(label=i18n("参考列表生成结果"), value="", interactive=False) gr.Markdown(value=i18n("1.2：选择基准音频，执行相似度匹配，并分段随机抽样")) text_sample_dir = gr.Text(label=i18n("参考音频抽样目录"), value="", interactive=False) button_convert_from_list.click(convert_from_list, [text_work_space_dir, text_character, text_list_input], [text_convert_from_list_info, text_sample_dir]) with gr.Row(): text_base_voice_path = gr.Text(label=i18n("请输入基准音频路径"), value="") text_subsection_num = gr.Text(label=i18n("请输入分段数"), value="10") text_sample_num = gr.Text(label=i18n("请输入每段随机抽样个数"), value="4") checkbox_similarity_output = gr.Checkbox(label=i18n("是否将相似度匹配结果输出到临时目录？"), show_label=True) with gr.Row(): button_sample = gr.Button(i18n("开始分段随机抽样"), variant="primary") text_sample_info = gr.Text(label=i18n("分段随机抽样结果"), value="", interactive=False) with gr.Accordion(label=i18n("第二步：基于参考音频和测试文本，执行批量推理"), open=False): gr.Markdown(value=i18n("2.1：配置推理服务参数信息，参考音频路径/文本和角色情绪二选一，如果是角色情绪，需要先执行第四步，" "将参考音频打包配置到推理服务下，在推理前，请确认完整请求地址是否与正常使用时的一致，包括角色名称，尤其是文本分隔符是否正确")) text_model_inference_voice_dir = gr.Text(label=i18n("待推理的参考音频所在目录"), value="", interactive=False) text_url = gr.Text(label=i18n("请输入推理服务请求地址与参数"), value="") with gr.Row(): text_text = gr.Text(label=i18n("请输入文本参数名"), value="text") text_ref_path = gr.Text(label=i18n("请输入参考音频路径参数名"), value="text") text_ref_text = gr.Text(label=i18n("请输入参考音频文本参数名"), value="text") text_emotion = gr.Text(label=i18n("请输入角色情绪参数名"), value="text") text_whole_url = gr.Text(label=i18n("完整地址"), value="5555555555555555", interactive=False) text_url.input(whole_url, [text_url, text_text, text_ref_path, text_ref_text, text_emotion], [text_whole_url]) text_text.input(whole_url, [text_url, text_text, text_ref_path, text_ref_text, text_emotion], [text_whole_url]) text_ref_path.input(whole_url, [text_url, text_text, text_ref_path, text_ref_text, text_emotion], [text_whole_url]) text_ref_text.input(whole_url, [text_url, text_text, text_ref_path, text_ref_text, text_emotion], [text_whole_url]) text_emotion.input(whole_url, [text_url, text_text, text_ref_path, text_ref_text, text_emotion], [text_whole_url]) gr.Markdown(value=i18n("2.2：配置待推理文本，一句一行，不要太多，10条即可")) text_test_content = gr.Text(label=i18n("请输入待推理文本路径"), value="text") gr.Markdown(value=i18n("2.3：启动推理服务，如果还没启动的话")) gr.Markdown(value=i18n("2.4：开始批量推理，这个过程比较耗时，可以去干点别的")) with gr.Row(): button_model_inference = gr.Button(i18n("开启批量推理"), variant="primary") text_model_inference_info = gr.Text(label=i18n("批量推理结果"), value="", interactive=False) with gr.Accordion(label=i18n("第三步：进行参考音频效果校验与筛选"), open=False): gr.Markdown(value=i18n("3.1：启动asr，获取推理音频文本")) text_asr_audio_dir = gr.Text(label=i18n("待asr的音频所在目录"), value="", interactive=False) button_model_inference.click(model_inference, [text_work_space_dir, text_character, text_model_inference_voice_dir, text_url, text_text, text_ref_path, text_ref_text, text_emotion, text_test_content], [text_model_inference_info, text_asr_audio_dir]) with gr.Row(): dropdown_asr_model = gr.Dropdown( label=i18n("ASR 模型"), choices=[], interactive=True, value="达摩 ASR (中文)" ) dropdown_asr_size = gr.Dropdown( label=i18n("ASR 模型尺寸"), choices=["large"], interactive=True, value="large" ) dropdown_asr_lang = gr.Dropdown( label=i18n("ASR 语言设置"), choices=["zh"], interactive=True, value="zh" ) with gr.Row(): button_asr = gr.Button(i18n("启动asr"), variant="primary") text_asr_info = gr.Text(label=i18n("asr结果"), value="", interactive=False) gr.Markdown(value=i18n("3.2：启动文本相似度分析")) text_text_similarity_analysis_path = gr.Text(label=i18n("待分析的文件路径"), value="", interactive=False) button_asr.click(asr, [text_work_space_dir, text_character, text_asr_audio_dir, dropdown_asr_model, dropdown_asr_size, dropdown_asr_lang], [text_asr_info, text_text_similarity_analysis_path]) with gr.Row(): button_text_similarity_analysis = gr.Button(i18n("启动文本相似度分析"), variant="primary") text_text_similarity_analysis_info = gr.Text(label=i18n("文本相似度分析结果"), value="", interactive=False) button_text_similarity_analysis.click(text_similarity_analysis, [text_work_space_dir, text_character, text_text_similarity_analysis_path], [text_text_similarity_analysis_info]) gr.Markdown(value=i18n("3.3：根据相似度分析结果，重点检查最后几条是否存在复读等问题")) gr.Markdown(value=i18n("3.4：对结果按音频相似度排序，筛选低音质音频")) with gr.Row(): text_base_audio_path = gr.Text(label=i18n("请输入基准音频"), value="text") text_compare_audio_dir = gr.Text(label=i18n("请输入待比较的音频文件目录"), value="text") with gr.Row(): button_similarity_audio_output = gr.Button(i18n("输出相似度-参考音频到临时目录"), variant="primary") text_similarity_audio_output_info = gr.Text(label=i18n("输出结果"), value="", interactive=False) button_similarity_audio_output.click(similarity_audio_output, [text_work_space_dir, text_character, text_base_audio_path, text_compare_audio_dir], [text_similarity_audio_output_info]) with gr.Row(): text_sync_ref_audio_dir = gr.Text(label=i18n("参考音频路径"), value="", interactive=False) text_sync_inference_audio_dir = gr.Text(label=i18n("被同步的推理音频路径"), value="", interactive=False) with gr.Row(): button_sync_ref_audio = gr.Button(i18n("将参考音频的删除情况，同步到推理音频目录"), variant="primary") text_sync_ref_info = gr.Text(label=i18n("同步结果"), value="", interactive=False) button_sync_ref_audio.click(sync_ref_audio, [text_work_space_dir, text_character, text_sync_ref_audio_dir, text_sync_inference_audio_dir], [text_sync_ref_info]) with gr.Accordion("第四步：生成参考音频配置文本", open=False): gr.Markdown(value=i18n("4.1：编辑模板")) text_template_path = gr.Text(label=i18n("模板文件路径"), value="", interactive=False) text_template = gr.Text(label=i18n("模板内容"), value="text", lines=10) gr.Markdown(value=i18n("4.2：生成配置")) text_sync_ref_audio_dir2 = gr.Text(label=i18n("参考音频路径"), value="", interactive=False) with gr.Row(): button_create_config = gr.Button(i18n("生成配置"), variant="primary") text_create_config_info = gr.Text(label=i18n("生成结果"), value="", interactive=False) button_create_config.click(create_config, [text_work_space_dir, text_character, text_template, text_sync_ref_audio_dir2], [text_create_config_info]) button_sample.click(sample, [text_work_space_dir, text_character, text_sample_dir, text_base_voice_path, text_subsection_num, text_sample_num, checkbox_similarity_output], [text_sample_info, text_model_inference_voice_dir, text_sync_ref_audio_dir, text_sync_ref_audio_dir2]) app.launch( server_port=9423, quiet=True, )