public static void main(String[] args) throws Exception {
ensureCorrectUsageOrExit(args);
final long startTime = System.currentTimeMillis();
final String[] ruleIds = args[0].split(",");
final File ruleFile = new File(args[1]);
final String languageCode = args[2];
final Language language = Language.getLanguageForShortName(languageCode);
final File indexDir = new File(args[3]);
if (args.length > 4 && "--no_limit".equals(args[4])) {
limitSearch = false;
}
final Searcher searcher = new Searcher(new SimpleFSDirectory(indexDir));
if (!limitSearch) {
searcher.setMaxHits(100000);
}
for (String ruleId : ruleIds) {
final long ruleStartTime = System.currentTimeMillis();
for (PatternRule rule : searcher.getRuleById(ruleId, ruleFile)) {
final SearcherResult searcherResult = searcher.findRuleMatchesOnIndex(rule, language);
int i = 1;
if (searcherResult.getMatchingSentences().size() == 0) {
System.out.println("[no matches]");
}
for (MatchingSentence ruleMatch : searcherResult.getMatchingSentences()) {
System.out.println(
i + ": " + ruleMatch.getSentence() + " (Source: " + ruleMatch.getSource() + ")");
i++;
}
System.out.println("Time: " + (System.currentTimeMillis() - ruleStartTime) + "ms");
System.out.println("==============================================================");
}
}
System.out.println("Total time: " + (System.currentTimeMillis() - startTime) + "ms");
}
/**
* Takes an XML file named <tt>rules-xx-language.xml</tt>, e.g. <tt>rules-de-German.xml</tt> and
* builds a Language object for that language.
*/
private static Language makeLanguage(final File file, final boolean isAdditional)
throws IllegalAccessException, InstantiationException {
Objects.requireNonNull(file, "file cannot be null");
if (!file.getName().endsWith(".xml")) {
throw new RuleFilenameException(file);
}
final String[] parts = file.getName().split("-");
final boolean startsWithRules = parts[0].equals("rules");
final boolean secondPartHasCorrectLength =
parts.length == 3
&& (parts[1].length() == "en".length()
|| parts[1].length() == "ast".length()
|| parts[1].length() == "en_US".length());
if (!startsWithRules || !secondPartHasCorrectLength) {
throw new RuleFilenameException(file);
}
// TODO: when the XML file is mergeable with
// other rules (check this in the XML Rule Loader by using rules[@integrate='add']?),
// subclass the existing language,
// and adjust the settings if any are set in the rule file default configuration set
Language newLanguage;
if (Language.isLanguageSupported(parts[1])) {
newLanguage = Language.getLanguageForShortName(parts[1]).getClass().newInstance();
newLanguage.addExternalRuleFile(file.getAbsolutePath());
newLanguage.setName(parts[2].replace(".xml", ""));
newLanguage.makeExternal();
} else {
newLanguage =
new Language() {
@Override
public Locale getLocale() {
return new Locale(getShortName());
}
@Override
public Contributor[] getMaintainers() {
return null;
}
@Override
public String getShortName() {
if (parts[1].length() == 2) {
return parts[1];
}
return parts[1].split("_")[0]; // en as in en_US
}
@Override
public String[] getCountries() {
if (parts[1].length() == 2) {
return new String[] {""};
}
return new String[] {parts[1].split("_")[1]}; // US as in en_US
}
@Override
public String getName() {
return parts[2].replace(".xml", "");
}
@Override
public void setName(final String name) {
// cannot be changed for this language
}
@Override
public List<Class<? extends Rule>> getRelevantRules() {
return Collections.emptyList();
}
@Override
public List<String> getRuleFileNames() {
final List<String> ruleFiles = new ArrayList<>();
ruleFiles.add(file.getAbsolutePath());
return ruleFiles;
}
@Override
public boolean isExternal() {
return isAdditional;
}
};
}
return newLanguage;
}
public static void main(String[] args) throws IOException, InterruptedException {
if (args.length != 2 && args.length != 3) {
System.out.println(
"Usage: "
+ AtomFeedCheckerCmd.class.getSimpleName()
+ " <atomFeedUrl> <sleepTime> [database.properties]");
System.out.println(" <atomFeedUrl> is a Wikipedia URL to the latest changes, for example:");
System.out.println(
" https://de.wikipedia.org/w/index.php?title=Spezial:Letzte_%C3%84nderungen&feed=atom&namespace=0");
System.out.println(
" <sleepTime> -1: don't loop at all (run once), 0: run in loop, other number: run in loop and");
System.out.println(" wait this many milliseconds between runs");
System.out.println(
" [database.properties] (optional) is a file that defines dbUrl, dbUser, and dbPassword,");
System.out.println(" used to write the results to a database via JDBC");
System.out.println("");
System.out.println(
" When the database.properties file is specified, this command will store all feed changes that");
System.out.println(
" cause LanguageTool rule matches to the database. If an error is then fixed later, this will");
System.out.println(
" usually also be detected and the rule match in the database will be marked as fixed. One case");
System.out.println(
" where this does not work is if the context of the error gets modified before the error is fixed.");
System.out.println("");
System.out.println(
" Run this command regularly so that you don't miss any changes from the feed.");
System.out.println(
" As the feed may contain only the latest 50 changes, running it more often than");
System.out.println(" once per minute may be needed for active Wikipedias.");
System.exit(1);
}
String url = args[0];
String langCode = url.substring(url.indexOf("//") + 2, url.indexOf("."));
System.out.println("Using URL: " + url);
System.out.println("Language code: " + langCode);
int sleepTimeMillis = Integer.parseInt(args[1]);
System.out.println("Sleep time: " + sleepTimeMillis + "ms (-1 = don't loop)");
DatabaseConfig databaseConfig = null;
if (args.length == 3) {
String propFile = args[2];
databaseConfig = new DatabaseConfig(propFile);
System.out.println("Writing results to database at: " + databaseConfig.getUrl());
}
Language language = Language.getLanguageForShortName(langCode);
AtomFeedChecker atomFeedChecker = new AtomFeedChecker(language, databaseConfig);
while (true) {
long startTime = System.currentTimeMillis();
try {
atomFeedChecker.runCheck(url);
System.out.println("Run time: " + (System.currentTimeMillis() - startTime) + "ms");
if (sleepTimeMillis == -1) {
// don't loop at all
break;
} else {
System.out.println("Sleeping " + sleepTimeMillis + "ms...");
Thread.sleep(sleepTimeMillis);
}
} catch (Exception e) {
// e.g. 50x HTTP errors, network problems
e.printStackTrace();
System.out.println("Sleeping " + sleepTimeMillis + "ms...");
Thread.sleep(sleepTimeMillis);
}
}
}
