markovnews

#!/usr/bin/env python
# Use PyMarkovChain to generate markov chains from news articles.
import sys, os
from emissary.models import Article
from pymarkovchain import MarkovChain

mc = MarkovChain(os.getenv('HOME') + os.sep + '.markovdb')

if __name__ == "__main__":
        if len(sys.argv) < 3:
            print "Usage: %s <uid> [<uid>] <#lines>" % sys.argv[0]
            raise SystemExit
        num_lines = sys.argv[2]
        article = Article.query.filter(Article.uid == sys.argv[1]).first()
        if not article:
            print "Couldn't find an article with that UID."
            raise SystemExit
        if not article.jsonify()['content_available']:
            print "No content available for %s" % sys.argv[1]
            raise SystemExit
        content = article.jsonify(content=True)['content'].decode("utf-8","ignore")
        if num_lines.isalnum():
            num_lines = sys.argv[3]
            article = Article.query.filter(Article.uid == sys.argv[2]).first()
            if not article:
                print "Couldn't find an article with that UID."
                raise SystemExit
            if not article.jsonify()['content_available']:
                print "No content available for %s" % sys.argv[2]
                raise SystemExit
            content += article.jsonify(content=True)['content'].decode("utf-8","ignore")

        mc.generateDatabase(content)
        for i in range(int(num_lines)):
                print mc.generateString()