public virtual void testUnmodifiableSet()
{
CharArraySet set = new CharArraySet(TEST_VERSION_CURRENT, 10, true);
set.addAll(TEST_STOP_WORDS);
set.add(Convert.ToInt32(1));
//JAVA TO C# CONVERTER WARNING: The original Java variable was marked 'final':
//ORIGINAL LINE: final int size = set.size();
int size = set.size();
set = CharArraySet.unmodifiableSet(set);
assertEquals("Set size changed due to unmodifiableSet call", size, set.size());
foreach (string stopword in TEST_STOP_WORDS)
{
assertTrue(set.contains(stopword));
}
assertTrue(set.contains(Convert.ToInt32(1)));
assertTrue(set.contains("1"));
assertTrue(set.contains(new char[] { '1' }));
try
{
CharArraySet.unmodifiableSet(null);
fail("can not make null unmodifiable");
}
catch (System.NullReferenceException)
{
// expected
}
}
/// <summary>
/// Creates a stopword set from the given stopword list. </summary>
/// <param name="matchVersion"> Lucene version to enable correct Unicode 4.0 behavior in the returned set if Version > 3.0 </param>
/// <param name="stopWords"> A List of Strings or char[] or any other toString()-able list representing the stopwords </param>
/// <param name="ignoreCase"> if true, all words are lower cased first </param>
/// <returns> A Set (<seealso cref="CharArraySet"/>) containing the words </returns>
public static CharArraySet makeStopSet <T1>(Version matchVersion, IList <T1> stopWords, bool ignoreCase)
{
CharArraySet stopSet = new CharArraySet(matchVersion, stopWords.Count, ignoreCase);
stopSet.addAll(stopWords);
return(stopSet);
}
public virtual void testClear()
{
CharArraySet set = new CharArraySet(TEST_VERSION_CURRENT, 10,true);
set.addAll(TEST_STOP_WORDS);
assertEquals("Not all words added", TEST_STOP_WORDS.Length, set.size());
set.clear();
assertEquals("not empty", 0, set.size());
for (int i = 0;i < TEST_STOP_WORDS.Length;i++)
{
assertFalse(set.contains(TEST_STOP_WORDS[i]));
}
set.addAll(TEST_STOP_WORDS);
assertEquals("Not all words added", TEST_STOP_WORDS.Length, set.size());
for (int i = 0;i < TEST_STOP_WORDS.Length;i++)
{
assertTrue(set.contains(TEST_STOP_WORDS[i]));
}
}
public virtual void testClear()
{
CharArraySet set = new CharArraySet(TEST_VERSION_CURRENT, 10, true);
set.addAll(TEST_STOP_WORDS);
assertEquals("Not all words added", TEST_STOP_WORDS.Length, set.size());
set.clear();
assertEquals("not empty", 0, set.size());
for (int i = 0; i < TEST_STOP_WORDS.Length; i++)
{
assertFalse(set.contains(TEST_STOP_WORDS[i]));
}
set.addAll(TEST_STOP_WORDS);
assertEquals("Not all words added", TEST_STOP_WORDS.Length, set.size());
for (int i = 0; i < TEST_STOP_WORDS.Length; i++)
{
assertTrue(set.contains(TEST_STOP_WORDS[i]));
}
}
/// <summary>
/// Test the static #copy() function with a CharArraySet as a source
/// </summary>
public virtual void testCopyCharArraySet()
{
CharArraySet setIngoreCase = new CharArraySet(TEST_VERSION_CURRENT, 10, true);
CharArraySet setCaseSensitive = new CharArraySet(TEST_VERSION_CURRENT, 10, false);
IList <string> stopwords = TEST_STOP_WORDS;
IList <string> stopwordsUpper = new List <string>();
foreach (string @string in stopwords)
{
stopwordsUpper.Add(@string.ToUpper(Locale.ROOT));
}
setIngoreCase.addAll(TEST_STOP_WORDS);
setIngoreCase.add(Convert.ToInt32(1));
setCaseSensitive.addAll(TEST_STOP_WORDS);
setCaseSensitive.add(Convert.ToInt32(1));
CharArraySet copy = CharArraySet.copy(TEST_VERSION_CURRENT, setIngoreCase);
CharArraySet copyCaseSens = CharArraySet.copy(TEST_VERSION_CURRENT, setCaseSensitive);
assertEquals(setIngoreCase.size(), copy.size());
assertEquals(setCaseSensitive.size(), copy.size());
assertTrue(copy.containsAll(stopwords));
assertTrue(copy.containsAll(stopwordsUpper));
assertTrue(copyCaseSens.containsAll(stopwords));
foreach (string @string in stopwordsUpper)
{
assertFalse(copyCaseSens.contains(@string));
}
// test adding terms to the copy
IList <string> newWords = new List <string>();
foreach (string @string in stopwords)
{
newWords.Add(@string + "_1");
}
copy.addAll(newWords);
assertTrue(copy.containsAll(stopwords));
assertTrue(copy.containsAll(stopwordsUpper));
assertTrue(copy.containsAll(newWords));
// new added terms are not in the source set
foreach (string @string in newWords)
{
assertFalse(setIngoreCase.contains(@string));
assertFalse(setCaseSensitive.contains(@string));
}
}
public virtual void testNonZeroOffset()
{
string[] words = new string[] { "Hello", "World", "this", "is", "a", "test" };
char[] findme = "xthisy".ToCharArray();
CharArraySet set = new CharArraySet(TEST_VERSION_CURRENT, 10, true);
set.addAll(words);
assertTrue(set.contains(findme, 1, 4));
assertTrue(set.contains(new string(findme, 1, 4)));
// test unmodifiable
set = CharArraySet.unmodifiableSet(set);
assertTrue(set.contains(findme, 1, 4));
assertTrue(set.contains(new string(findme, 1, 4)));
}
/// <summary>
/// Test the static #copy() function with a JDK <seealso cref="Set"/> as a source
/// </summary>
public virtual void testCopyJDKSet()
{
ISet <string> set = new HashSet <string>();
IList <string> stopwords = TEST_STOP_WORDS;
IList <string> stopwordsUpper = new List <string>();
foreach (string @string in stopwords)
{
stopwordsUpper.Add(@string.ToUpper(Locale.ROOT));
}
set.addAll(TEST_STOP_WORDS);
CharArraySet copy = CharArraySet.copy(TEST_VERSION_CURRENT, set);
assertEquals(set.Count, copy.size());
assertEquals(set.Count, copy.size());
assertTrue(copy.containsAll(stopwords));
foreach (string @string in stopwordsUpper)
{
assertFalse(copy.contains(@string));
}
IList <string> newWords = new List <string>();
foreach (string @string in stopwords)
{
newWords.Add(@string + "_1");
}
copy.addAll(newWords);
assertTrue(copy.containsAll(stopwords));
assertTrue(copy.containsAll(newWords));
// new added terms are not in the source set
foreach (string @string in newWords)
{
assertFalse(set.Contains(@string));
}
}
/// <summary>
/// Test the static #copy() function with a CharArraySet as a source
/// </summary>
public virtual void testCopyCharArraySet()
{
CharArraySet setIngoreCase = new CharArraySet(TEST_VERSION_CURRENT, 10, true);
CharArraySet setCaseSensitive = new CharArraySet(TEST_VERSION_CURRENT, 10, false);
IList<string> stopwords = TEST_STOP_WORDS;
IList<string> stopwordsUpper = new List<string>();
foreach (string @string in stopwords)
{
stopwordsUpper.Add(@string.ToUpper(Locale.ROOT));
}
setIngoreCase.addAll(TEST_STOP_WORDS);
setIngoreCase.add(Convert.ToInt32(1));
setCaseSensitive.addAll(TEST_STOP_WORDS);
setCaseSensitive.add(Convert.ToInt32(1));
CharArraySet copy = CharArraySet.copy(TEST_VERSION_CURRENT, setIngoreCase);
CharArraySet copyCaseSens = CharArraySet.copy(TEST_VERSION_CURRENT, setCaseSensitive);
assertEquals(setIngoreCase.size(), copy.size());
assertEquals(setCaseSensitive.size(), copy.size());
assertTrue(copy.containsAll(stopwords));
assertTrue(copy.containsAll(stopwordsUpper));
assertTrue(copyCaseSens.containsAll(stopwords));
foreach (string @string in stopwordsUpper)
{
assertFalse(copyCaseSens.contains(@string));
}
// test adding terms to the copy
IList<string> newWords = new List<string>();
foreach (string @string in stopwords)
{
newWords.Add(@string + "_1");
}
copy.addAll(newWords);
assertTrue(copy.containsAll(stopwords));
assertTrue(copy.containsAll(stopwordsUpper));
assertTrue(copy.containsAll(newWords));
// new added terms are not in the source set
foreach (string @string in newWords)
{
assertFalse(setIngoreCase.contains(@string));
assertFalse(setCaseSensitive.contains(@string));
}
}
public virtual void testUnmodifiableSet()
{
CharArraySet set = new CharArraySet(TEST_VERSION_CURRENT, 10,true);
set.addAll(TEST_STOP_WORDS);
set.add(Convert.ToInt32(1));
//JAVA TO C# CONVERTER WARNING: The original Java variable was marked 'final':
//ORIGINAL LINE: final int size = set.size();
int size = set.size();
set = CharArraySet.unmodifiableSet(set);
assertEquals("Set size changed due to unmodifiableSet call", size, set.size());
foreach (string stopword in TEST_STOP_WORDS)
{
assertTrue(set.contains(stopword));
}
assertTrue(set.contains(Convert.ToInt32(1)));
assertTrue(set.contains("1"));
assertTrue(set.contains(new char[]{'1'}));
try
{
CharArraySet.unmodifiableSet(null);
fail("can not make null unmodifiable");
}
catch (System.NullReferenceException)
{
// expected
}
}
public virtual void testNonZeroOffset()
{
string[] words = new string[] {"Hello","World","this","is","a","test"};
char[] findme = "xthisy".ToCharArray();
CharArraySet set = new CharArraySet(TEST_VERSION_CURRENT, 10, true);
set.addAll(words);
assertTrue(set.contains(findme, 1, 4));
assertTrue(set.contains(new string(findme,1,4)));
// test unmodifiable
set = CharArraySet.unmodifiableSet(set);
assertTrue(set.contains(findme, 1, 4));
assertTrue(set.contains(new string(findme,1,4)));
}
public virtual void testModifyOnUnmodifiable()
{
CharArraySet set = new CharArraySet(TEST_VERSION_CURRENT, 10, true);
set.addAll(TEST_STOP_WORDS);
//JAVA TO C# CONVERTER WARNING: The original Java variable was marked 'final':
//ORIGINAL LINE: final int size = set.size();
int size = set.size();
set = CharArraySet.unmodifiableSet(set);
assertEquals("Set size changed due to unmodifiableSet call", size, set.size());
string NOT_IN_SET = "SirGallahad";
assertFalse("Test String already exists in set", set.contains(NOT_IN_SET));
try
{
set.add(NOT_IN_SET.ToCharArray());
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.add(NOT_IN_SET);
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.add(new StringBuilder(NOT_IN_SET));
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.clear();
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Changed unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.add((object) NOT_IN_SET);
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
// This test was changed in 3.1, as a contains() call on the given Collection using the "correct" iterator's
// current key (now a char[]) on a Set<String> would not hit any element of the CAS and therefor never call
// remove() on the iterator
try
{
set.removeAll(new CharArraySet(TEST_VERSION_CURRENT, TEST_STOP_WORDS, true));
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.retainAll(new CharArraySet(TEST_VERSION_CURRENT, NOT_IN_SET, true));
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.addAll(NOT_IN_SET);
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
}
for (int i = 0; i < TEST_STOP_WORDS.Length; i++)
{
assertTrue(set.contains(TEST_STOP_WORDS[i]));
}
}
public virtual void TestModifyOnUnmodifiable()
{
CharArraySet set = new CharArraySet(TEST_VERSION_CURRENT, 10, true);
set.AddAll(TEST_STOP_WORDS);
int size = set.size();
set = CharArraySet.UnmodifiableSet(set);
assertEquals("Set size changed due to unmodifiableSet call", size, set.size());
string NOT_IN_SET = "SirGallahad";
assertFalse("Test String already exists in set", set.Contains(NOT_IN_SET));
try
{
set.Add(NOT_IN_SET.ToCharArray());
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.add(NOT_IN_SET);
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.Add(new StringBuilder(NOT_IN_SET));
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.clear();
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Changed unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.add(NOT_IN_SET);
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
// NOTE: This results in a StackOverflow exception. Since this is not a public member of CharArraySet,
// but an extension method for the test fixture (which apparently has a bug), this test is non-critical
//// This test was changed in 3.1, as a contains() call on the given Collection using the "correct" iterator's
//// current key (now a char[]) on a Set<String> would not hit any element of the CAS and therefor never call
//// remove() on the iterator
//try
//{
// set.removeAll(new CharArraySet(TEST_VERSION_CURRENT, TEST_STOP_WORDS, true));
// fail("Modified unmodifiable set");
//}
//catch (System.NotSupportedException)
//{
// // expected
// assertEquals("Size of unmodifiable set has changed", size, set.size());
//}
#region Added for better .NET support
// This test was added for .NET to check the Remove method, since the extension method
// above fails to execute.
try
{
#pragma warning disable 612, 618
set.Remove(TEST_STOP_WORDS[0]);
#pragma warning restore 612, 618
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
#endregion
try
{
set.retainAll(new CharArraySet(TEST_VERSION_CURRENT, new [] { NOT_IN_SET }, true));
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.addAll(new[] { NOT_IN_SET});
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
}
// LUCENENET Specific - added to test .NETified UnionWith method
try
{
set.UnionWith(new[] { NOT_IN_SET });
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
}
for (int i = 0; i < TEST_STOP_WORDS.Length; i++)
{
assertTrue(set.contains(TEST_STOP_WORDS[i]));
}
}
public virtual void testModifyOnUnmodifiable()
{
CharArraySet set = new CharArraySet(TEST_VERSION_CURRENT, 10, true);
set.addAll(TEST_STOP_WORDS);
//JAVA TO C# CONVERTER WARNING: The original Java variable was marked 'final':
//ORIGINAL LINE: final int size = set.size();
int size = set.size();
set = CharArraySet.unmodifiableSet(set);
assertEquals("Set size changed due to unmodifiableSet call", size, set.size());
string NOT_IN_SET = "SirGallahad";
assertFalse("Test String already exists in set", set.contains(NOT_IN_SET));
try
{
set.add(NOT_IN_SET.ToCharArray());
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.add(NOT_IN_SET);
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.add(new StringBuilder(NOT_IN_SET));
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.clear();
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Changed unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.add((object)NOT_IN_SET);
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
// This test was changed in 3.1, as a contains() call on the given Collection using the "correct" iterator's
// current key (now a char[]) on a Set<String> would not hit any element of the CAS and therefor never call
// remove() on the iterator
try
{
set.removeAll(new CharArraySet(TEST_VERSION_CURRENT, TEST_STOP_WORDS, true));
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.retainAll(new CharArraySet(TEST_VERSION_CURRENT, NOT_IN_SET, true));
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertEquals("Size of unmodifiable set has changed", size, set.size());
}
try
{
set.addAll(NOT_IN_SET);
fail("Modified unmodifiable set");
}
catch (System.NotSupportedException)
{
// expected
assertFalse("Test String has been added to unmodifiable set", set.contains(NOT_IN_SET));
}
for (int i = 0; i < TEST_STOP_WORDS.Length; i++)
{
assertTrue(set.contains(TEST_STOP_WORDS[i]));
}
}
/// <summary>
/// Returns as <seealso cref="CharArraySet"/> from wordFiles, which
/// can be a comma-separated list of filenames
/// </summary>
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: protected final CharArraySet getWordSet(ResourceLoader loader, String wordFiles, boolean ignoreCase) throws java.io.IOException
protected internal CharArraySet GetWordSet(ResourceLoader 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(luceneMatchVersion, files.Count * 10, ignoreCase);
foreach (string file in files)
{
IList<string> wlist = getLines(loader, file.Trim());
words.addAll(StopFilter.makeStopSet(luceneMatchVersion, wlist, ignoreCase));
}
}
return words;
}