1
def process_audio_file(file_path, sample_rate=16000, n_mfcc=35):
2
    """
3
    音频文件处理函数
4

5
    参数:
6
        file_path (str): 音频文件路径
7
        sample_rate (int): 采样率，默认16000Hz
8
        n_mfcc (int): MFCC系数数量，默认35
9

10
    返回:
11
        np.ndarray: 提取的MFCC特征
12
    """
13
    try:
14
        y, sr = librosa.load(file_path, sr=sample_rate)
15
        mfccs = librosa.feature.mfcc(y=y, sr=sr, n_mfcc=n_mfcc)
16
        return np.mean(mfccs, axis=1)
17
    except Exception as e:
18
        print(f"音频处理失败: {e}")
19
        return None

1
class FeatureExtractor:
2
    """特征提取器基类"""
3

4
    def __init__(self, sample_rate=16000):
5
        self.sample_rate = sample_rate
6
        self.features = []
7

8
    def extract(self, file_path):
9
        """提取特征的抽象方法"""
10
        raise NotImplementedError("子类必须实现extract方法")
11

12
    def save_features(self, output_path):
13
        """保存提取的特征"""
14
        np.save(output_path, self.features)
15

16

17
class AudioFeatureExtractor(FeatureExtractor):
18
    """音频特征提取器"""
19

20
    def __init__(self, sample_rate=16000, n_mfcc=35):
21
        super().__init__(sample_rate)
22
        self.n_mfcc = n_mfcc
23

24
    def extract(self, file_path):
25
        """提取音频MFCC特征"""
26
        try:
27
            y, sr = librosa.load(file_path, sr=self.sample_rate)
28
            mfccs = librosa.feature.mfcc(y=y, sr=sr, n_mfcc=self.n_mfcc)
29
            feature = np.mean(mfccs, axis=1)
30
            self.features.append(feature)
31
            return feature
32
        except Exception as e:
33
            print(f"音频特征提取失败: {e}")
34
            return None

1
def safe_file_operation(file_path):
2
    """安全的文件操作示例"""
3
    try:
4
        with open(file_path, 'r', encoding='utf-8') as f:
5
            content = f.read()
6

7
        # 处理文件内容
8
        processed = content.upper()
9

10
        with open(file_path.replace('.txt', '_processed.txt'), 'w', encoding='utf-8') as f:
11
            f.write(processed)
12

13
        return True
14

15
    except FileNotFoundError:
16
        print(f"错误：文件 {file_path} 不存在")
17
        return False
18
    except PermissionError:
19
        print(f"错误：没有权限访问文件 {file_path}")
20
        return False
21
    except UnicodeDecodeError:
22
        print(f"错误：文件编码不是UTF-8")
23
        return False
24
    except Exception as e:
25
        print(f"发生未知错误：{e}")
26
        return False

1
import numpy as np
2

3
# 创建数组
4
arr = np.array([1, 2, 3, 4, 5])
5

6
# 数组运算
7
result = arr * 2 + 1  # 向量化运算，无需循环
8

9
# 多维数组操作
10
matrix = np.array([[1, 2], [3, 4]])
11
transposed = matrix.T  # 转置
12
flattened = matrix.flatten()  # 展平
13

14
# 广播机制
15
a = np.array([[1, 2, 3], [4, 5, 6]])
16
b = np.array([10, 20, 30])
17
result = a + b  # 自动广播

1
import pandas as pd
2

3
# 创建DataFrame
4
data = {
5
    'participant_id': [101, 102, 103],
6
    'phq_score': [12, 8, 15],
7
    'depression': [1, 0, 1]
8
}
9
df = pd.DataFrame(data)
10

11
# 数据筛选
12
depressed = df[df['phq_score'] >= 10]
13

14
# 数据转换
15
df['depression_level'] = pd.cut(
16
    df['phq_score'],
17
    bins=[0, 4, 9, 14, 27],
18
    labels=['正常', '轻度', '中度', '重度']
19
)
20

21
# 数据聚合
22
stats = df.groupby('depression').agg({
23
    'phq_score': ['mean', 'std', 'min', 'max']
24
})

1
import os
2
from pathlib import Path
3

4
# 传统路径操作
5
base_dir = os.path.join('data', 'processed')
6
os.makedirs(base_dir, exist_ok=True)
7
file_path = os.path.join(base_dir, 'features.npy')
8

9
# Pathlib现代路径操作（推荐）
10
base_path = Path('data') / 'processed'
11
base_path.mkdir(parents=True, exist_ok=True)
12
file_path = base_path / 'features.npy'
13

14
# 文件检查
15
if file_path.exists():
16
    print(f"文件大小: {file_path.stat().st_size / 1024:.2f} KB")
17

18
# 遍历目录
19
for file in base_path.glob('*.npy'):
20
    print(f"找到文件: {file.name}")

1
多模态抑郁检测系统/
2
├── 📁 本地测试系统/          # 推理系统和GUI界面
3
│   ├── gui_main.py         # GUI主程序入口
4
│   ├── inference.py        # 推理引擎核心
5
│   ├── feature_extractor.py # 实时特征提取
6
│   └── report_generator.py # 报告生成模块
7
├── 📁 模型文件/             # 训练好的模型
8
│   ├── audio_model.joblib  # 音频模型
9
│   ├── video_model.joblib  # 视频模型
10
│   └── fusion_model.joblib # 融合模型
11
├── 📁 配置文件/             # 系统配置
12
│   ├── config.py           # 全局配置
13
│   └── gui_config.json     # GUI配置
14
├── 📁 教程/                # 项目教程文档
15
├── 1.数据预处理.py          # 数据预处理脚本
16
├── 2.提取音视频特征.py       # 特征提取脚本
17
├── 3.训练模型.py           # 模型训练脚本
18
├── 4.本地测试.py           # 本地测试脚本
19
└── requirements.txt        # 依赖库列表

1
# config.py - 全局配置文件
2
import os
3
from pathlib import Path
4

5
# 项目根目录
6
PROJECT_ROOT = Path(__file__).parent
7

8
# 数据集配置
9
DATASET_ROOT = r"E:\DAIC-WOZ"
10
DATA_DIR = PROJECT_ROOT / "数据"
11
OUTPUT_DIR = PROJECT_ROOT / "输出"
12

13
# 外部工具配置
14
OPENFACE_PATH = r"D:\OpenFace_2.2.0_win_x64\FeatureExtraction.exe"
15

16
# 音频配置
17
SAMPLE_RATE = 16000
18
N_MFCC = 35
19

20
# 确保目录存在
21
os.makedirs(DATA_DIR, exist_ok=True)
22
os.makedirs(OUTPUT_DIR, exist_ok=True)

1
# 调试技巧示例
2
import logging
3

4
# 配置日志
5
logging.basicConfig(
6
    level=logging.DEBUG,
7
    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s'
8
)
9
logger = logging.getLogger(__name__)
10

11
def process_data(data):
12
    logger.debug(f"输入数据: {data}")
13

14
    try:
15
        result = data * 2
16
        logger.info(f"处理结果: {result}")
17
        return result
18
    except Exception as e:
19
        logger.error(f"处理失败: {e}", exc_info=True)
20
        return None