Compile with ghc randomizer.hs and run with ./randomizer <input file> <sentence count>. If you're on Windows, converting these commands to Windows command prompt commands shouldn't be too hard.

The program reads the text file you give as its argument and compiles a statistic of how often words follow after other words (aka counts word pairs). It then builds sentences not unlike a Markov chain, using a weighted choice algorithm to choose the following words in the sentence. The first word of every sentence is the only word that is chosen without considering weight. A sentence ends when the generator picks a word that ends with a '.', '!' or a '?'. You control how many sentences it generates with the sentence count parameter.

In this repository, you'll find a couple of test files you can run the program with to see what it does. The following files are present:

• hgttg.txt - a chapter from the Hitchhiker's Guide to the Galaxy
• trump.txt - a speech Donald Trump made in September 2017 during his first United Nations address
• cuni.txt - a brief excerpt from the history of the Charles University
• simple.txt - two simple sentences, there are only a couple of sentences that can actually be generated
• empty.txt - in case you want to try to break my program and/or don't know how to make an empty file

made as part of my non-procedural programming lab at MFF CUNI