python - 从日志文件的特定行中选取值并计算平均值

Question

各位，我有一个如下所示的日志文件

Input image filename:(416, 416, 3)
Found 3 boxes for img
Inference time : 3.100685347104464 
traffic light 0.85 
traffic light 0.96 
traffic light 0.98 
Input image filename:(416, 416, 3)
Found 3 boxes for img
Inference time : 2.0816197767817197 
traffic light 0.90 
traffic light 0.92 
Input image filename:(416, 416, 3)
Found 3 boxes for img
Inference time : 2.0610929683485253 
traffic light 0.82 
traffic light 0.96 
traffic light 0.99 

我想编写一个循环，将“推理时间..”和“输入图像文件名..”行之间的交通灯值相加并计算平均值。
例如:
第一部分:(0.85+0.96+0.98)/3 = 0.93
第二部分:(0.90+0.92)/2 = 0.91
第三部分:(0.82+0.96+0.99)/3 = 0.93

返回值:(0.93, 0.91, 0.93)

到目前为止，我只能读取找到“交通灯”一词的行中的所有值。

import os 
log_path  = "data/test.log"

accuracy_lines = []
accuracy = [] 

for line_acc in open(log): 
    if 'traffic light ' in line_acc: 
        accuracy_lines.append(line_acc)

for n in range(len(accuracy_lines)):
    lineParts = accuracy_lines[n].split(',')
    accuracy.append(float(lineParts[0].split()[2]))
print(accuracy)

不幸的是，我不知道如何继续。如果你能帮助我，我会很高兴。

许多问候

Answer 1

使用以下方法:

log_path  = "data/test.log"

with open(log_path) as f:
    numbers, res = [], []
    for line in f:
        if line.startswith('traffic light'):
            numbers.append(float(line.split()[-1]))
        elif line.startswith('Input image filename') and numbers:
            res.append(sum(numbers) / (len(numbers)))
            numbers.clear()
    if numbers: res.append(sum(numbers) / len(numbers))  # capturing trailing numbers
    print(res)

输出:

[0.93, 0.91, 0.9233333333333332]