Resampler in strategies can misalign indices

[kirby561]

There's sort of an implicit assumption in the way that triggers work that the "current index" is the same timestamp for both the original dataset and the filtered dataset. If a resampler is run though, it can change the number of samples in the filtered dataset even if it covers the same timespan. This can cause weird effects and potentially crashes when running strategies.

I can think of a few solutions:

1. A quick solution would just be to enforce equal spacing to run strategies and not allow filters that change the shape of the data within the strategy itself. That would mean the user would need to resample it before running the strategy which is probably preferred anyway.

2. Another potential solution would be to allow filters that change shape to be run first and treat the end of shape-changing filter section as the original dataset. While this would work, I think this may be confusing and expecting the shape not to change mid-run is probably fine.

3, If the datasets were indexed using timestamps instead of indices, this problem would go away too. this would be fancy and may need to allow interpolation to be specified somewhere too.