NLZMarkov

Markov chain based on articles of the "Neue Luzerner Zeitung".

With two modes:

• one is a naive Markov chain that only depends on the previous words.
• the second is a part of speech enhanced variante, where the next word depends on the previous words and the type expected by the last given word.

What possible to make better:

• the gui is not robust, because it doesn't use multithreading
• split the methods better, such that a module doesn't consist of just one method.

What to do better next time:

• better names from the start.
• more methods.
• methods verbs, modules nouns.

Used libraries:

• PyQt5
• urllib3
• urllib
• nltk
• treetagger(wrapper) and the german language package
• standard python libraries (i think)