Architecture issues

[ilev-sergey]

Now, with the addition of interactive plotting (39af056), it became obvious that some features that I want to implement require both raw data and results obtained after data processing. The currently implemented pipeline for data: reading -> processing -> fitting -> processing of results doesn't allow to implement these features conveniently, because to access the data in the end, it should be read again, which takes time. Although, such a pipeline was chosen to reduce memory consumption and to simplify working with data, it seems that it is no longer an option. I'm going to try to rewrite the functionality in OOP style, although it seems that there may be some issues with complication of logic and increase in peak memory consumption that need to be considered.