lit-parse

a program that plots Vonnegut's shapes of stories with questionable accuracy

Essentially, this program takes a provided input text and does some simple natural language processing to determine the sentiment polarity of a chunk of text. The polarity versus time (as measured by the progression of sentences) is then plotted. A polarity of -1 indicates 'very bad', and a polarity of +1 indicates 'very good'.

Of course, a rather large flaw in this process is that it can't handling nonlinear/non-chronological texts. Also, the accuracy is sort of questionable.

usage

You can use the program simply by running python main.py. By default, it will read the input text from the provided input.txt file and will analyze each sentence separately. You can modify this behavior by specifying the input file and the step. When you specify a step of n, the program averages the polarity across every n sentences (probably not the best way of doing it, but oh well). For example, to analyze the provided cinderella.txt file with step 3, you could run

python main.py cinderella.txt 3

dependencies

This program uses TextBlob for natural language processing. You can install TextBlob with

pip install -U textblob

This program also uses matplotlib for grpahing. You can install matplotlib with

python -mpip install -U matplotlib