/// <summary>
/// Creates a <see cref="CharArraySet"/> from a file.
/// </summary>
/// <param name="stopwords">
/// the stopwords reader to load
/// </param>
/// <param name="matchVersion">
/// the Lucene version for cross version compatibility </param>
/// <returns> a <see cref="CharArraySet"/> containing the distinct stopwords from the given
/// reader </returns>
/// <exception cref="IOException">
/// if loading the stopwords throws an <see cref="IOException"/> </exception>
protected static CharArraySet LoadStopwordSet(TextReader stopwords, LuceneVersion matchVersion)
{
try
{
return(WordlistLoader.GetWordSet(stopwords, matchVersion));
}
finally
{
IOUtils.Dispose(stopwords);
}
}
/// <summary>
/// Creates a <see cref="CharArraySet"/> from a file.
/// </summary>
/// <param name="stopwords">
/// the stopwords file to load
/// </param>
/// <param name="matchVersion">
/// the Lucene version for cross version compatibility </param>
/// <returns> a <see cref="CharArraySet"/> containing the distinct stopwords from the given
/// file </returns>
/// <exception cref="IOException">
/// if loading the stopwords throws an <see cref="IOException"/> </exception>
protected static CharArraySet LoadStopwordSet(FileInfo stopwords, LuceneVersion matchVersion)
{
TextReader reader = null;
try
{
reader = IOUtils.GetDecodingReader(stopwords, Encoding.UTF8);
return(WordlistLoader.GetWordSet(reader, matchVersion));
}
finally
{
IOUtils.Dispose(reader);
}
}
/// <summary>
/// Creates a <see cref="CharArraySet"/> from an embedded resource associated with a class. (See
/// <see cref="Assembly.GetManifestResourceStream(string)"/>).
/// </summary>
/// <param name="ignoreCase">
/// <c>true</c> if the set should ignore the case of the
/// stopwords, otherwise <c>false</c> </param>
/// <param name="aClass">
/// a class that is associated with the given stopwordResource </param>
/// <param name="resource">
/// name of the resource file associated with the given class </param>
/// <param name="comment">
/// comment string to ignore in the stopword file </param>
/// <returns> a <see cref="CharArraySet"/> containing the distinct stopwords from the given
/// file </returns>
/// <exception cref="IOException">
/// if loading the stopwords throws an <see cref="IOException"/> </exception>
protected static CharArraySet LoadStopwordSet(bool ignoreCase, Type aClass, string resource, string comment)
{
TextReader reader = null;
try
{
var resourceStream = aClass.FindAndGetManifestResourceStream(resource);
reader = IOUtils.GetDecodingReader(resourceStream, Encoding.UTF8);
return(WordlistLoader.GetWordSet(reader, comment, new CharArraySet(
#pragma warning disable 612, 618
LuceneVersion.LUCENE_CURRENT, 16, ignoreCase)));
#pragma warning restore 612, 618
}
finally
{
IOUtils.Dispose(reader);
}
}
/// <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);
}
/// <summary>
/// Returns the resource's lines (with content treated as UTF-8)
/// </summary>
protected IList <string> GetLines(IResourceLoader loader, string resource)
{
return(WordlistLoader.GetLines(loader.OpenResource(resource), Encoding.UTF8));
}
