Made a small comparison of the outputs and made sure that the strategy (algorithm) that is to be used can be chosen here.

It is integrated with the existing framework

line is important such that it overrides the find_alignment method in Alignment_Strategy

Put a general class that is supposed to be an interface to resemble the strategy pattern. The now existing algorithms implement the find_alignment method

Fixed some bugs and rewrote the code in such a way that the pointer class is no longer needed. Since this class did not have a lot of functionality.

I modified several things about the Needleman-Wunsch implementation. These changes made the code more readable and also shorter. It also allowed me to several functions and parameters from the Pointer class. Since those were no longer needed. One primary change is that I have changed the pointer-frame from a list structure to a numpy array. Which allowed for easier operations and allowed me to shorten several lines.

I have included a dummy accuracy score function. The current version is based on the shadowed attribute of source words. This also means that it will only function properly after this attribute has been assigned to its correct value for every source word.

Finalised the algorithm. Only have to update the documentation and format the code (aka make smaller functions).

Added the new version. Will clean up the code later. Tried testing with other files. Seems to work alright, though there are things going wrong for file 9.

Added get_difference function in Word classes and started on a new version of my comparison algorithm. However, I will restart next time. I have the idea to work with anchors: words that only occur once in both files will serve as anchors, and the remaining words will only be matched between two sets of anchors.

I made sure my implementation of the sequence alignment algorithm works with...

See merge request !25

I made sure my implementation of the sequence alignment algorithm works with the recently changed strategy of the program. I also improved the implementation, such that it now gives the correct results. Note that while the program does give the correct results, it does not do this in an optimal way yet.

Analysed the code more and tried to solve the problem of it skipping a word while it should not. Isolated the problem, to be fixed later. Also fixed the shadow_index that is returned by search_locally.

Added self.save_pref initialisation to init to fix 'AttributeError: 'ModelClass' object has no attribute 'save_pref'

Added self.save_pref initialisation to init to fix 'AttributeError: 'ModelClass' object has no attribute 'save_pref'

Improved code quality of search_locally and added functionality to compare(), both in data_comparer.py. Now the compare() method looks back in time. I manually compared it with the so called true comparison. It goes wrong at two places, which I have noted down and will change another day.

Sc 75 #bulk load

See merge request !21

Sc 80 #full saving options

See merge request !20

Sc 80.1

See merge request !19

Sc 76

See merge request !17

Added mistake counter functionality and applied changes to report it in the saved documents.

See merge request !16

Created a new solution concept working with kernels. Did a singllar test and interpreted it to understand what still needs to be adapted.