private void assertClasspathDelegation(IResourceLoader rl)
{
//var englishStopText = System.IO.Path.Combine(analysisCommonFolder.FullName, @"Analysis\Snowball\english_stop.txt");
// LUCENENET specific - rather than being completely dependent on the location of the file
// in the file system, we use the embedded resource to write the file to a known location
// before passing it to our resource loader.
string englishStopFile = "english_stop.txt";
var file = CreateTempFile(System.IO.Path.GetFileNameWithoutExtension(englishStopFile), System.IO.Path.GetExtension(englishStopFile));
using (var stream = typeof(Snowball.SnowballFilter).GetTypeInfo().Assembly.FindAndGetManifestResourceStream(typeof(Snowball.SnowballFilter), englishStopFile))
{
using (var outputStream = new FileStream(file.FullName, FileMode.OpenOrCreate, FileAccess.Write))
{
stream.CopyTo(outputStream);
}
}
// try a stopwords file from classpath
CharArraySet set = WordlistLoader.GetSnowballWordSet(new System.IO.StreamReader(rl.OpenResource(file.FullName), Encoding.UTF8), TEST_VERSION_CURRENT);
assertTrue(set.contains("you"));
// try to load a class; we use string comparison because classloader may be different...
assertEquals("Lucene.Net.Analysis.Util.RollingCharBuffer", rl.NewInstance <object>("Lucene.Net.Analysis.Util.RollingCharBuffer").ToString());
// theoretically classes should also be loadable:
//IOUtils.CloseWhileHandlingException(rl.OpenResource("java/lang/String.class")); // LUCENENET TODO: Not sure what the equivalent to this is (or if there is one).
}
private void assertClasspathDelegation(IResourceLoader rl)
{
// try a stopwords file from classpath
CharArraySet set = WordlistLoader.GetSnowballWordSet(new System.IO.StreamReader(rl.OpenResource(System.IO.Path.GetFullPath(@"..\..\..\Lucene.Net.Analysis.Common\Analysis\Snowball\english_stop.txt")), Encoding.UTF8), TEST_VERSION_CURRENT);
assertTrue(set.contains("you"));
// try to load a class; we use string comparison because classloader may be different...
assertEquals("Lucene.Net.Analysis.Util.RollingCharBuffer", rl.NewInstance <object>("Lucene.Net.Analysis.Util.RollingCharBuffer").ToString());
// theoretically classes should also be loadable:
//IOUtils.CloseWhileHandlingException(rl.OpenResource("java/lang/String.class")); // LUCENENET TODO: Not sure what the equivalent to this is (or if there is one).
}
public virtual void TestSnowballListLoading()
{
string s = "|comment\n" + " |comment\n" + "\n" + " \t\n" + " |comment | comment\n" + "ONE\n" + " two \n" + " three four five \n" + "six seven | comment\n"; //multiple stopwords + comment - multiple stopwords - stopword with leading/trailing space - stopword, in uppercase - commented line with comment - line with only whitespace - blank line - commented line with leading whitespace - commented line
CharArraySet wordset = WordlistLoader.GetSnowballWordSet(new StringReader(s), TEST_VERSION_CURRENT);
assertEquals(7, wordset.size());
assertTrue(wordset.contains("ONE"));
assertTrue(wordset.contains("two"));
assertTrue(wordset.contains("three"));
assertTrue(wordset.contains("four"));
assertTrue(wordset.contains("five"));
assertTrue(wordset.contains("six"));
assertTrue(wordset.contains("seven"));
}
/// <summary>
/// Same as <see cref="GetWordSet(IResourceLoader, string, bool)"/>,
/// except the input is in snowball format.
/// </summary>
protected CharArraySet GetSnowballWordSet(IResourceLoader loader, string wordFiles, bool ignoreCase)
{
AssureMatchVersion();
IList <string> files = SplitFileNames(wordFiles);
CharArraySet words = null;
if (files.Count > 0)
{
// default stopwords list has 35 or so words, but maybe don't make it that
// big to start
words = new CharArraySet(m_luceneMatchVersion, files.Count * 10, ignoreCase);
foreach (string file in files)
{
using (Stream stream = loader.OpenResource(file.Trim()))
using (TextReader reader = new StreamReader(stream, Encoding.UTF8))
{
WordlistLoader.GetSnowballWordSet(reader, words);
}
}
}
return(words);
}