Added the functionality that was necessary to create a data frame of the delays. Will currently only work for NW because of has_shadow property.

I cleaned up the code in the FormChecker class mainly. I also commented out some print statements that were not necessary for me. The cleaning up mainly included making code more readable, for which I also made one additional function.

Added print statements such that we can more easily see the words that the algorithms flag, and thus the differences.

Added documentation and an accuracy function able to calculate the accuracy of a given aligned shadowing task.

Added the form mistake type to the mistake enum for good measure. Added a mistake_counter class that gets the aligned source and shadow words and outputs the number of each mistake made in the shadow task. For now also prints these values to the console, for testing purposes. Added instance of this class to the statistics class, for analysis after alignment.

It works, so I'll put in the merge request. Intended to make prints disable-able, but too much effort right now. Will try later

I started fixing the various errors in the current master branch. Still some errors are left. I am already pushing this code so that Tanja and Romeo can also look at it.

I relocated the initialization of the Wn_grid_parser object, in order to make the statistics class less messy. I also fixed a bug that was in the master branch at the moment I pulled the code.

Tested the algorithm on the test data and tried implementing another rule. Lastly inspected the results.

This reverts merge request !51

Mistake finder under 'Emma' did not get parser argument for sementic mistake recognition. It now uses the MistakeFinder object initialised in the constructor.

Now the comparing of the algorithms is fair. First, anchor algorithm used the list that was already modified by the SAA algorithm.

Fixed mistake assignment in words.py and made the code runnable (aka commented out some things in mistake_finder.py)

Increased modularity in MistakeFinder, changed Statistics accordingly and made MistakeDefinitions.txt more readable.

Added the mistake_finder.py (forgot in last commit). Used for filtering out the mistakes. Skeleton now. Integrated its call in statistics.py for the Anchor Algorithm (commented out now). Small PEP8 change in anchor_algorithm.py.

Solved python -m django --version and refactored menu. Started on adding file selection functionality.

anchor_algorithm called source and shadow as parameter or as 
self._source/shadow interchangably. This is now unified as self._source 
and self._shadow everywhere.

Removed the functions to calculate the latencies and mistakes as they relied on the old frame which is deprecated.

Refactored all namings such that the functions using the functions in Word which were named differently after the update work again.