Merge fails with TypeError

[ThomDietrich]

Hey there, I want to report the following bug.
In case I am not correct and this is wrong use on my end, I wonder if documentation needs to be updated.
All the best!

---

Python 3.8.5, traces 0.6.0

import traces
ts = [traces.TimeSeries(), traces.TimeSeries()]
ts[1][2] = 3
traces.TimeSeries.merge(ts, operation=sum)

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/lib/python3.8/site-packages/traces/timeseries.py", line 778, in merge
    default = operation(default)
TypeError: unsupported operand type(s) for +: 'int' and 'NoneType'

[tasq-inc]

I found a work around for now. It would seem that NoneType is produced probably when data is not present for the time interval being merged and aggregated. Passing in a custom aggregate function that forces NoneType to 0 worked for me:

import traces

def custom_func(x):
    x_fillna = [i if i else 0 for i in x]
    return sum(x_fillna)

ts = [traces.TimeSeries(), traces.TimeSeries()]
ts[1][2] = 3
traces.TimeSeries.merge(ts, operation=custom_func)

[stringertheory]

@ThomDietrich thanks for opening this issue! It's a great idea to improve the documentation about this.

(The Exception is the desired behavior in this case. The first time series has the default value None for all times, and the second is None for all times before 2, and then 3 afterwards. At, for example, t=4, merge with operation=sum is summing None + 3, and does not automatically assume that Nones should be ignored).

Thanks @tasq-inc for posting a solution. This solution or something similar is exactly what I'd suggest using if you want to ignore Nones.