寻找时间序列的周期

Question

I have a time series in a dataframe and i would like to find cases when signal "S" is significantly high and then low again.

So i want to find 3 time periods A,B,C for which A: (s = average) -> B: (s = high -> C: (df[s] = average again). I have tried the following which finds n largest B moments for "S" and then tries to find periods before and after that B moment that satisfies the constraints:

B = df.nlargest(3, 'S').index
threshold = 1
A = []
C = []
for b in B:
    A.append(min([i for i in df.index if (i < b) & (df['S'][i] <= df['S'][b] - threshold)], key=lambda x: abs(x - b)))
    C.append(min([i for i in df.index if (i > b) & (df['S'][i] <= df['S'][b] - threshold)], key=lambda x: abs(x - b)))

The problem is that get a ValueError: min() arg is an empty sequence error since there are not always A and C moments found. How can i solve this? Can i use some kind of error handling and just print "No moments found"? Or maybe a more clever method? I would appreciate a lot some help

Thank you

example for df:

In [2]: df['S']
Out[2]: Date   Values
2019-01-06    194.021278
2019-02-03    196.034816
2019-03-03    193.071251
2019-03-31    190.647765
2019-04-28    190.791410
2019-05-26    190.465997
2019-06-23    192.097875
2019-07-21    192.704626
2019-08-18    193.757577
2019-09-15    193.480539
2019-10-13    193.800638
2019-11-10    192.289041
2019-12-08    195.755910
2020-01-05    198.654722
2020-02-02    199.804535
2020-03-01    198.191558