将打印信息，改由日志输出

Ref_Audio_Selector/common/common.py

@@ -1,5 +1,5 @@
 from tools import my_utils
-import glob
+from Ref_Audio_Selector.config_param.log_config import logger
 import os
 
 
@@ -102,9 +102,9 @@ def write_text_to_file(text, output_file_path):
         with open(output_file_path, 'w', encoding='utf-8') as file:
             file.write(text)
     except IOError as e:
-        print(f"Error occurred while writing to the file: {e}")
+        logger.info(f"Error occurred while writing to the file: {e}")
     else:
-        print(f"Text successfully written to file: {output_file_path}")
+        logger.info(f"Text successfully written to file: {output_file_path}")
 
 
 if __name__ == '__main__':

Ref_Audio_Selector/config.ini

@@ -1,15 +1,21 @@
 # config.ini
 
 [Base]
-# 函数时间消耗日志打印类型 file 打印到文件; close 关闭
-time_log_print_type = file
-# 函数时间消耗日志保存目录路径
-time_log_print_dir = Ref_Audio_Selector/log/performance
 # 参考音频目录
 reference_audio_dir = refer_audio
 # 临时文件目录
 temp_dir = Ref_Audio_Selector/temp
 
+[Log]
+# 日志保存目录路径
+log_dir = Ref_Audio_Selector/log/general
+# 日志级别 CRITICAL、FATAL、ERROR、WARNING、WARN、INFO、DEBUG、NOTSET、
+log_level = INFO
+# 函数时间消耗日志打印类型 file 打印到文件; close 关闭
+time_log_print_type = file
+# 函数时间消耗日志保存目录路径
+time_log_print_dir = Ref_Audio_Selector/log/performance
+
 [AudioSample]
 # list转换待选参考音频目录
 list_to_convert_reference_audio_dir = refer_audio_all

Ref_Audio_Selector/config_param/config_manager.py

@@ -33,6 +33,9 @@ class ConfigManager:
     def get_base(self, key):
         return self.config.get('Base', key)
 
+    def get_log(self, key):
+        return self.config.get('Log', key)
+
     def get_audio_sample(self, key):
         return self.config.get('AudioSample', key)
 

Ref_Audio_Selector/config_param/config_params.py

@@ -3,15 +3,21 @@ import Ref_Audio_Selector.config_param.config_manager as config_manager
 config = config_manager.get_config()
 
 # [Base]
-# 函数时间消耗日志打印类型 file 打印到文件; close 关闭
-time_log_print_type = config.get_base('time_log_print_type')
-# 函数时间消耗日志保存目录路径
-time_log_print_dir = config.get_base('time_log_print_dir')
 # 参考音频目录
 reference_audio_dir = config.get_base('reference_audio_dir')
 # 临时文件目录
 temp_dir = config.get_base('temp_dir')
 
+# [Log]
+# 日志保存目录路径
+log_dir = config.get_log('log_dir')
+# 日志级别 CRITICAL、FATAL、ERROR、WARNING、WARN、INFO、DEBUG、NOTSET、
+log_level = config.get_log('log_level')
+# 函数时间消耗日志打印类型 file 打印到文件; close 关闭
+time_log_print_type = config.get_log('time_log_print_type')
+# 函数时间消耗日志保存目录路径
+time_log_print_dir = config.get_log('time_log_print_dir')
+
 # [AudioSample]
 # list转换待选参考音频目录
 list_to_convert_reference_audio_dir = config.get_audio_sample('list_to_convert_reference_audio_dir')

Ref_Audio_Selector/config_param/log_config.py

@@ -3,37 +3,56 @@ import datetime
 import Ref_Audio_Selector.config_param.config_params as params
 
 
-def setup_logging():
+def create_general_logger():
     # 获取当前日期，用于文件名和日志内容
     current_date = datetime.datetime.now().strftime('%Y-%m-%d')
 
+    # 创建一个用于控制台输出的处理器，并设置日志级别
+    console_handler = logging.StreamHandler()
+    # console_handler.setLevel(logging.INFO)
+    # 可以设置控制台输出的格式
+    console_formatter = logging.Formatter('%(asctime)s - %(levelname)s - %(message)s')
+    console_handler.setFormatter(console_formatter)
+
     # 创建一个用于常规日志的处理器
-    general_handler = logging.FileHandler('general.log', mode='a', encoding='utf-8')
+    general_handler = logging.FileHandler(f"{params.log_dir}/{current_date}.log", mode='a', encoding='utf-8')
-    general_handler.setLevel(logging.INFO)
+    # general_handler.setLevel(logging.INFO)
     general_formatter = logging.Formatter('%(asctime)s - %(levelname)s - %(message)s')
     general_handler.setFormatter(general_formatter)
 
+    # 配置一个常规的logger
+    general_logger = logging.getLogger('general')
+    level = logging.getLevelName(params.log_level)
+    general_logger.setLevel(level)
+    general_logger.addHandler(console_handler)
+    general_logger.addHandler(general_handler)
+
+    # 配置根logger，以防万一
+    logging.basicConfig(level=logging.WARNING, handlers=[general_handler])
+
+    return general_logger
+
+
+def create_performance_logger():
+    # 获取当前日期，用于文件名和日志内容
+    current_date = datetime.datetime.now().strftime('%Y-%m-%d')
     # 创建一个专用于性能监控日志的处理器
     performance_handler = logging.FileHandler(
         f"{params.time_log_print_dir}/{current_date}.log", mode='a', encoding='utf-8')
-    performance_handler.setLevel(logging.INFO)
+    # performance_handler.setLevel(logging.INFO)
     performance_formatter = logging.Formatter('%(asctime)s - %(levelname)s - %(message)s')
     performance_handler.setFormatter(performance_formatter)
 
-    # 配置一个常规的logger
-    general_logger = logging.getLogger('general')
-    general_logger.setLevel(logging.INFO)
-    general_logger.addHandler(general_handler)
-
     # 配置一个专门用于性能监控的logger
     performance_logger = logging.getLogger('performance')
     performance_logger.setLevel(logging.INFO)
     performance_logger.addHandler(performance_handler)
 
-    # 配置根logger，以防万一
+    return performance_logger
-    logging.basicConfig(level=logging.WARNING, handlers=[general_handler])
-
-    return general_logger, performance_logger
 
 
-logger, p_logger = setup_logging()
+def setup_logging():
+    return create_general_logger(), create_performance_logger()
+
+
+logger, p_logger = setup_logging()

Ref_Audio_Selector/ref_audio_selector_webui.py

@@ -60,7 +60,7 @@ def convert_from_list(text_work_space_dir, text_role, text_list_input):
 
         # audio_similarity.convert_from_list(text_list_input, ref_audio_all)
     except Exception as e:
-        traceback.print_exc()
+        logger.error("发生异常: \n%s", traceback.format_exc())
         text_convert_from_list_info = f"发生异常：{e}"
         text_sample_dir = ''
     return i18n(text_convert_from_list_info), text_sample_dir
@@ -83,7 +83,7 @@ def start_similarity_analysis(work_space_dir, sample_dir, base_voice_path, need_
         cmd += f' -c "{sample_dir}"'
         cmd += f' -o {similarity_file}'
 
-        print(cmd)
+        logger.info(cmd)
         p_similarity = Popen(cmd, shell=True)
         p_similarity.wait()
 
@@ -136,7 +136,7 @@ def sample(text_work_space_dir, text_role, text_sample_dir, text_base_voice_path
         audio_similarity.sample(ref_audio_dir, similarity_list, int(text_subsection_num), int(text_sample_num))
 
     except Exception as e:
-        traceback.print_exc()
+        logger.error("发生异常: \n%s", traceback.format_exc())
         text_sample_info = f"发生异常：{e}"
         ref_audio_dir = ''
     text_model_inference_voice_dir = ref_audio_dir
@@ -191,7 +191,7 @@ def model_inference(text_work_space_dir, text_role, text_model_inference_voice_d
         # audio_inference.generate_audio_files(url_composer, text_list, ref_audio_manager.get_ref_audio_list(),
         #                                      inference_dir)
     except Exception as e:
-        traceback.print_exc()
+        logger.error("发生异常: \n%s", traceback.format_exc())
         text_model_inference_info = f"发生异常：{e}"
         text_asr_audio_dir = ''
     return i18n(text_model_inference_info), text_asr_audio_dir, text_asr_audio_dir
@@ -226,7 +226,7 @@ def asr(text_work_space_dir, text_role, text_asr_audio_dir, dropdown_asr_model,
         text_text_similarity_analysis_path = asr_file
         text_asr_info = f"耗时：{time_consuming:0.1f}秒；asr成功：生成文件{asr_file}"
     except Exception as e:
-        traceback.print_exc()
+        logger.error("发生异常: \n%s", traceback.format_exc())
         text_asr_info = f"发生异常：{e}"
         text_text_similarity_analysis_path = ''
     return i18n(text_asr_info), text_text_similarity_analysis_path
@@ -248,7 +248,7 @@ def open_asr(asr_inp_dir, asr_opt_dir, asr_model, asr_model_size, asr_lang):
         cmd += f' -l {asr_lang}'
         cmd += " -p %s" % ("float16" if is_half == True else "float32")
 
-        print(cmd)
+        logger.info(cmd)
         p_asr = Popen(cmd, shell=True)
         p_asr.wait()
         p_asr = None
@@ -284,7 +284,7 @@ def text_similarity_analysis(text_work_space_dir, text_role,
 
         # open_text_similarity_analysis(text_text_similarity_analysis_path, similarity_dir)
     except Exception as e:
-        traceback.print_exc()
+        logger.error("发生异常: \n%s", traceback.format_exc())
         text_text_similarity_analysis_info = f"发生异常：{e}"
     return i18n(text_text_similarity_analysis_info)
 
@@ -297,7 +297,7 @@ def open_text_similarity_analysis(asr_file_path, output_dir, similarity_enlarge_
         cmd += f' -o "{output_dir}"'
         cmd += f' -b {similarity_enlarge_boundary}'
 
-        print(cmd)
+        logger.info(cmd)
         p_text_similarity = Popen(cmd, shell=True)
         p_text_similarity.wait()
         p_text_similarity = None
@@ -335,7 +335,7 @@ def similarity_audio_output(text_work_space_dir, text_role, text_base_audio_path
         text_similarity_audio_output_info = f'耗时：{time_consuming:0.1f}秒；相似度分析成功：生成目录{similarity_file_dir}，文件{similarity_file}'
 
     except Exception as e:
-        traceback.print_exc()
+        logger.error("发生异常: \n%s", traceback.format_exc())
         text_similarity_audio_output_info = f"发生异常：{e}"
     return i18n(text_similarity_audio_output_info)
 
@@ -361,7 +361,7 @@ def sync_ref_audio(text_work_space_dir, text_role, text_sync_ref_audio_dir,
         text_sync_ref_audio_info = (f"耗时：{time_consuming:0.1f}秒；推理音频目录{text_sync_inference_audio_dir}下，"
                                     f"text目录删除了{delete_text_wav_num}个参考音频，emotion目录下，删除了{delete_emotion_dir_num}个目录")
     except Exception as e:
-        traceback.print_exc()
+        logger.error("发生异常: \n%s", traceback.format_exc())
         text_sync_ref_audio_info = f"发生异常：{e}"
     return i18n(text_sync_ref_audio_info)
 
@@ -392,7 +392,7 @@ def create_config(text_work_space_dir, text_role, text_template, text_sync_ref_a
         text_create_config_info = f"耗时：{time_consuming:0.1f}秒；配置生成成功：生成文件{config_file}"
 
     except Exception as e:
-        traceback.print_exc()
+        logger.error("发生异常: \n%s", traceback.format_exc())
         text_create_config_info = f"发生异常：{e}"
     return i18n(text_create_config_info)
 

Ref_Audio_Selector/tool/asr/funasr_asr_multi_level_dir.py

@@ -4,6 +4,7 @@ import argparse
 import os
 import traceback
 import Ref_Audio_Selector.config_param.config_params as params
+from Ref_Audio_Selector.config_param.log_config import logger
 from Ref_Audio_Selector.common.time_util import timeit_decorator
 from tqdm import tqdm
 from funasr import AutoModel
@@ -31,7 +32,7 @@ def only_asr(input_file):
         text = model.generate(input=input_file)[0]["text"]
     except:
         text = ''
-        print(traceback.format_exc())
+        logger.error(traceback.format_exc())
     return text
 
 
@@ -54,7 +55,7 @@ def execute_asr_multi_level_dir(input_folder, output_folder, model_size, languag
                     output.append(f"{input_file_path}|{original_text}|{language.upper()}|{asr_text}")
 
                 except:
-                    print(traceback.format_exc())
+                    logger.error(traceback.format_exc())
 
     # 创建或打开指定的输出目录
     output_folder = output_folder or "output/asr_opt"
@@ -67,7 +68,7 @@ def execute_asr_multi_level_dir(input_folder, output_folder, model_size, languag
     # 将输出写入文件
     with open(output_file_path, "w", encoding="utf-8") as f:
         f.write("\n".join(output))
-        print(f"ASR 任务完成->标注文件路径: {output_file_path}\n")
+        logger.info(f"ASR 任务完成->标注文件路径: {output_file_path}\n")
 
     return output_file_path
 

Ref_Audio_Selector/tool/audio_inference.py

@@ -4,6 +4,7 @@ import itertools
 import Ref_Audio_Selector.config_param.config_params as params
 from Ref_Audio_Selector.common.time_util import timeit_decorator
 from urllib.parse import urlparse, parse_qs, urlencode, urlunparse, quote
+from Ref_Audio_Selector.config_param.log_config import logger
 
 
 class URLComposer:
@@ -69,7 +70,7 @@ def safe_encode_query_params(original_url):
     new_parsed_url = parsed_url._replace(query=new_query_string)
     encoded_url = urlunparse(new_parsed_url)
 
-    print(encoded_url)
+    logger.info(encoded_url)
     return encoded_url
 
 
@@ -107,7 +108,7 @@ def generate_audio_files(url_composer, text_list, emotion_list, output_dir_path)
         # 检查是否已经存在对应的音频文件，如果存在则跳过
         if os.path.exists(text_subdir_text_file_path) and os.path.exists(emotion_subdir_emotion_file_path):
             has_generated_count += 1
-            print(f"进度: {has_generated_count}/{all_count}")
+            logger.info(f"进度: {has_generated_count}/{all_count}")
             continue
 
         if url_composer.is_emotion():
@@ -124,7 +125,7 @@ def generate_audio_files(url_composer, text_list, emotion_list, output_dir_path)
             f.write(audio_bytes)
 
         has_generated_count += 1
-        print(f"进度: {has_generated_count}/{all_count}")
+        logger.info(f"进度: {has_generated_count}/{all_count}")
 
 
 def inference_audio_from_api(url):

Ref_Audio_Selector/tool/audio_similarity.py

@@ -2,6 +2,7 @@ import os
 import shutil
 import random
 import librosa
+from Ref_Audio_Selector.config_param.log_config import logger
 
 
 def check_audio_duration(path, min_duration=3, max_duration=10):
@@ -17,7 +18,7 @@ def check_audio_duration(path, min_duration=3, max_duration=10):
             return False
 
     except Exception as e:
-        print(f"无法打开或处理音频文件：{e}")
+        logger.error(f"无法打开或处理音频文件：{e}")
         return None
 
 
@@ -33,7 +34,7 @@ def convert_from_list(list_file, output_dir):
     for line in lines:
         parts = line.strip().split('|')
         if len(parts) != 4:
-            print(f"Line format incorrect: {line}")
+            logger.error(f"Line format incorrect: {line}")
             continue
 
         audio_path, _, _, transcription = parts
@@ -46,27 +47,27 @@ def convert_from_list(list_file, output_dir):
 
         # 如果目标文件已存在，不要覆盖
         if os.path.exists(new_path):
-            print(f"File already exists: {new_path}")
+            logger.info(f"File already exists: {new_path}")
             continue
 
         try:
             # 检查音频文件是否存在
             if not os.path.exists(audio_path):
-                print(f"Audio file does not exist: {audio_path}")
+                logger.info(f"Audio file does not exist: {audio_path}")
                 continue
 
             if check_audio_duration(audio_path):
                 # 复制音频文件到output目录并重命名
                 shutil.copy2(audio_path, new_path)
-                print(f"File copied and renamed to: {new_path}")
+                logger.info(f"File copied and renamed to: {new_path}")
             else:
-                print(f"File skipped due to duration: {audio_path}")
+                logger.info(f"File skipped due to duration: {audio_path}")
 
         except Exception as e:
-            print(f"An error occurred while processing: {audio_path}")
+            logger.error(f"An error occurred while processing: {audio_path}")
-            print(e)
+            logger.error(e)
 
-    print("Processing complete.")
+    logger.info("Processing complete.")
 
 
 def sample(output_audio_dir, similarity_list, subsection_num, sample_num):
@@ -101,7 +102,7 @@ def sample(output_audio_dir, similarity_list, subsection_num, sample_num):
             dst_path = os.path.join(subdir_path, os.path.basename(src_path))
             shutil.copyfile(src_path, dst_path)
 
-    print("Sampling completed.")
+    logger.info("Sampling completed.")
 
 
 def parse_similarity_file(file_path):
@@ -150,7 +151,7 @@ def copy_and_move(output_audio_directory, similarity_scores):
         # 复制文件到新目录
         shutil.copyfile(item['wav_path'], new_path)
 
-    print("已完成复制和重命名操作。")
+    logger.info("已完成复制和重命名操作。")
 
 
 if __name__ == '__main__':

Ref_Audio_Selector/tool/delete_inference_with_ref.py

@@ -2,6 +2,7 @@ import os
 import shutil
 import Ref_Audio_Selector.common.common as common
 import Ref_Audio_Selector.config_param.config_params as params
+from Ref_Audio_Selector.config_param.log_config import logger
 
 
 def remove_matching_audio_files_in_text_dir(text_dir, emotions_list):
@@ -12,12 +13,12 @@ def remove_matching_audio_files_in_text_dir(text_dir, emotions_list):
             wav_file_name = f"{emotion_tag}.wav"
             file_path = os.path.join(root, wav_file_name)
             if not os.path.exists(file_path):
-                print(f"Deleting file: {file_path}")
+                logger.info(f"Deleting file: {file_path}")
                 try:
                     os.remove(file_path)
                     count += 1
                 except Exception as e:
-                    print(f"Error deleting file {file_path}: {e}")
+                    logger.error(f"Error deleting file {file_path}: {e}")
     return count
 
 
@@ -39,13 +40,13 @@ def delete_emotion_subdirectories(emotion_dir, emotions_list):
 
         # 检查emotion子目录是否存在
         if not os.path.isdir(folder_path):
-            print(f"Deleting directory: {folder_path}")
+            logger.info(f"Deleting directory: {folder_path}")
             try:
                 # 使用shutil.rmtree删除整个子目录及其内容
                 shutil.rmtree(folder_path)
                 count += 1
             except Exception as e:
-                print(f"Error deleting directory {folder_path}: {e}")
+                logger.error(f"Error deleting directory {folder_path}: {e}")
     return count
 
 

Ref_Audio_Selector/tool/speaker_verification/voice_similarity.py

@@ -5,6 +5,7 @@ import torchaudio.transforms as T
 import platform
 import Ref_Audio_Selector.config_param.config_params as params
 from Ref_Audio_Selector.common.time_util import timeit_decorator
+from Ref_Audio_Selector.config_param.log_config import logger
 
 from modelscope.pipelines import pipeline
 
@@ -32,6 +33,8 @@ def compare_audio_and_generate_report(reference_audio_path, comparison_dir_path,
         reference_audio_16k = reference_audio_path
 
     # Step 2: 用参考音频依次比较音频目录下的每个音频，获取相似度分数及对应路径
+    all_count = len(comparison_audio_paths)
+    has_processed_count = 0
     similarity_scores = []
     for audio_path in comparison_audio_paths:
         score = sv_pipeline([reference_audio_16k, audio_path])['score']
@@ -39,7 +42,8 @@ def compare_audio_and_generate_report(reference_audio_path, comparison_dir_path,
             'score': score,
             'path': audio_path
         })
-        print(f'similarity score: {score}, path: {audio_path}')
+        has_processed_count += 1
+        logger.info(f'进度：{has_processed_count}/{all_count}')
 
     # Step 3: 根据相似度分数降序排列
     similarity_scores.sort(key=lambda x: x['score'], reverse=True)

Ref_Audio_Selector/tool/text_comparison/asr_text_process.py

@@ -6,6 +6,7 @@ from Ref_Audio_Selector.common.time_util import timeit_decorator
 import Ref_Audio_Selector.tool.text_comparison.text_comparison as text_comparison
 import Ref_Audio_Selector.config_param.config_params as params
 import Ref_Audio_Selector.common.common as common
+from Ref_Audio_Selector.config_param.log_config import logger
 
 
 def parse_asr_file(file_path):
@@ -42,7 +43,7 @@ def calculate_similarity_and_append_to_list(input_list, boundary):
         item['similarity_score'] = similarity_score
         item['original_score'] = original_score
         has_been_processed_count += 1
-        print(f'进度：{has_been_processed_count}/{all_count}')
+        logger.info(f'进度：{has_been_processed_count}/{all_count}')
 
     return input_list
 
@@ -135,7 +136,7 @@ def process(asr_file_path, output_dir, similarity_enlarge_boundary):
     original_text_detail_file = os.path.join(output_dir, f'{params.text_similarity_by_text_detail_filename}.txt')
     format_list_to_emotion(original_text_detail_list, original_text_detail_file)
 
-    print('文本相似度分析完成。')
+    logger.info('文本相似度分析完成。')
 
 
 def parse_arguments():

Ref_Audio_Selector/tool/text_comparison/text_comparison.py

@@ -2,7 +2,7 @@ import os
 import torch
 from transformers import AutoTokenizer, AutoModel
 from scipy.spatial.distance import cosine
-import math
+from Ref_Audio_Selector.config_param.log_config import logger
 
 bert_path = os.environ.get(
     "bert_path", "GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large"
@@ -11,7 +11,7 @@ bert_path = os.environ.get(
 # Set device to GPU if available, else CPU
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
-print(f'使用计算设备: {device}')
+logger.info(f'使用计算设备: {device}')
 
 tokenizer = AutoTokenizer.from_pretrained(bert_path)
 model = AutoModel.from_pretrained(bert_path).to(device)