update quickstart with the tile locator, add locator for all scorer and fix negative levels

[nicolebussola]

Resolve issue 207

[coveralls]

Coverage remained the same at 100.0% when pulling e481035 on quick_start into 0a5c448 on master.

[ernestoarbitrio]

please do the same adjustment suggested in the readme also in the quickstart.rst. Furthermore i'd like to mention the possibility to use the locate tiles even in the ScoreTiler, justo to inform that this method is available for each scorer in histolab

[ernestoarbitrio]

LGTM

[alessiamarcolini]

I can't see why to generate the tiles and the scores twice (if we save the report) by getting the scores and the scaled scores separately. I guess we are duplicating the execution time because we are calling self._tiles_generator(slide) twice.
We could @lru_cache the method _tiles_generator so with different slide parameter the result is recalculated.... but I'm not sure, @ernestoarbitrio what do you think?

[ernestoarbitrio]

I can't see why to generate the tiles and the scores twice (if we save the report) by getting the scores and the scaled scores separately. I guess we are duplicating the execution time because we are calling self._tiles_generator(slide) twice.
We could @lru_cache the method _tiles_generator so with different slide parameter the result is recalculated.... but I'm not sure, @ernestoarbitrio what do you think?

The problem here is that for the locate tiles (scorer) we can't have a method that returns 2 values, otherwise we cannot generalize that method. The idea of just parametrizing with the scored(bool) was just a POC for having some workable and usable thing. Myabe this can be refactored in a separate PR ... i can take care of it ... lmk what you think. If you agree just merge this and re-open a new issue and assign it to me :D

[alessiamarcolini]

@ernestoarbitrio alright, approving this and opening a new issue