脸书求real-time avg time spent for each step

先谢谢大家，求问一道脸书的python题。 写code 把stream 改成 real-time data 实时更新每步的平均时间。

input：[session, step, time_start] | data like [[a,1,110], [a,2,120], [a,3,150], [b,1,110],[b,2,190],[b,3,220]]
output: {1:45, 2:35}

我的理解是，点在于如何随时有新的[session, step, time_start] 进行实时更新output。使用extend() 新的input吗？求各位大神赐教

万分感谢

"抱歉，您不能对匿名帖评分"

如果希望大米还请不要匿名发帖。无法给米。

Here is my first thought here, let me know if there is any doubts.

1. outputDict = {} #{1:45, 2:35}
   
2. sessionDict = {} #{'a':[1,110], 'b':[1,110]} keep the latest step of each session
   
3. stepCountDict = {} #{1:2, 2:2}
   
4. 
   
5. #input [a,1,110]
   
6. def step_avg_time(session, step, time_start):
   
7.         last_session = sessionDict.get(session)
   
8.         #for new session, no change
   
9.         if not last_session:
   
10.                 sessionDict[session] = [step, time_start]
    
11.                 return outputDict
    
12. 
    
13.         time_spend = time_start - last_session[1]
    
14.         last_step = last_session[0]
    
15.         step_avg = outputDict.get(last_step, 0.0)
    
16.         #update the average
    
17.         step_avg = step_avg + (time_spend - step_avg) / (stepCountDict.get(last_step,0) + 1)
    
18.         outputDict[last_step] = step_avg
    
19.         #update the sessionDict
    
20.         sessionDict[session] = [step, time_start]
    
21.         #update stepCountDict
    
22.         stepCountDict[last_step] = stepCountDict[last_step] + 1
    
23.         return outputDict
    
24. 
    
25. #call the function
    
26. #inputList: [[a,1,110], [a,2,120], [a,3,150], [b,1,110],[b,2,190],[b,3,220]]
    
27. for i in inputList:
    
28.         print(step_avg_time(*i))
    

复制代码