private List<Word> ranking(SWCDocument document, CommandLineArguments cmd) {
RankingAlgo algo = RankingAlgorithmRegistry.getById(cmd.getRankAlgorithm());
document.weightFilter(cmd.getMaxWords(), algo);
List<Word> words = document.getWords();
if (words.size() < 10)
throw new RuntimeException("The input text is too short (" + words.size() + " words)");
return words;
}
private SWCDocument readDocument(CommandLineArguments cmd) throws FileNotFoundException {
Scanner scanner =
cmd.getInputFile() != null
? new Scanner(new File(cmd.getInputFile()))
: new Scanner(System.in);
StringBuilder sb = new StringBuilder();
while (scanner.hasNextLine()) {
sb.append(scanner.nextLine() + "\n");
}
scanner.close();
SWCDocument doc = new SWCDocument(sb.toString());
doc.parse(cmd.getParseOptions());
return doc;
}