ixt / 605204

Simple Haiku converter written in Python + Project to convert the vital wikipedia articles to Haiku

60//5\204

A sequel to 60//5\203, now with automation!

in this repo there is:

• an article list
• a python scrypt with no none core dependencies
• a shell script that goes through the list, grabbing the articles putting the output through the python script

Using lynx to get the articles is good most of the time this is demonstrated in the included script which is meant for pulling the 1000 articles that are important to have in every wikipedia & convert them into haikus. The wikipedia chosen was the Simple English one. Only one article fails right now (2016/09/05) which is Trimurti due to too little words, but if one does the process outlined in the python script it should actually find a haiku. This is due to the counting algorithm, which can probably be improved greatly (many examples can be found of it counting too little syllables)

Utilities I use in the example script:

• Lynx
• cat
• sed
• echo
• python (imports sys, os, getopt)
• ascii2uni && uni2ascii

605204's Issues

Some entries fail to get through sumy

Need to find a better way to summarize/parse text from articles, sumy sucks for not getting the titles to parse and sometimes sumy does too good a job at summarising so the script doesnt have enough words to chance by 5 - 7 - 5

Some haikus are filled with dates or paren parts

This can probably be solved with more sed manipulation. For some reason the bracket removing sed i have implemented in it doesnt seem to get rid of them all.