Allow Top-N controller to receive an initial list of exclusions

[joewandy]

For issue #89

The Top-N controller now accepts a parameter initial_exclusion_list. This is a list of precursor ions to be excluded during Top-N selection. The list can be obtained from e.g. a previous Top-N run.

Had to broaden the RT window when creating an exclusion item, so it could match the precursor ion in the next run https://github.com/sdrogers/vimms/compare/iss_89?expand=1#diff-c6299074495d125a6af6e704ba2434ddR170. Maybe this will help with RT drift too.

To test, please run test_TopN_controller_with_beer_chems_and_initial_exclusion_list

[joewandy]

It seems that changing the from_rt in

rt_lower = current_time
rt_upper = current_time + rt_tol
x = ExclusionItem(from_mz=mz_lower, to_mz=mz_upper, from_rt=rt_lower, to_rt=rt_upper)

to

rt_lower = current_time - rt_tol

broke the weight calculation in https://github.com/sdrogers/vimms/blob/iss_89/vimms/Controller/topN.py#L458-L459 (the assert failed).

weight = (rt - (self.exclusion_t_0 + x.from_rt)) / (self.rt_tol - self.exclusion_t_0)
assert weight <= 1, weight

Could you help to look into this? @sdrogers

[sdrogers]

Whoops - breaks excluding topN controller. Can we: Give an excluded item the time it was created as well? And then use that in place of the from time at the point where the excluding topN is failing?

[sdrogers]

More info: the problem is that the excluding top N needs the time at which the fragmentation happened, which is how things used to be.

Should be easy to fix: if we always store excluded items with a from time of the DEW value, then we can add it to any use of from time in the excluding controller

[sdrogers]

Fixes #89