//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: static void assertLegalOffsets(String in) throws Exception
internal static void assertLegalOffsets(string @in)
{
int length = @in.Length;
HTMLStripCharFilter reader = new HTMLStripCharFilter(new System.IO.StreamReader(new StringReader(@in)));
int ch = 0;
int off = 0;
while ((ch = reader.read()) != -1)
{
int correction = reader.correctOffset(off);
assertTrue("invalid offset correction: " + off + "->" + correction + " for doc of length: " + length, correction <= length);
off++;
}
}
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: public void testMSWord14GeneratedHTML() throws Exception
public virtual void testMSWord14GeneratedHTML()
{
System.IO.Stream stream = this.GetType().getResourceAsStream("MS-Word 14 generated.htm");
HTMLStripCharFilter reader = new HTMLStripCharFilter(new System.IO.StreamReader(stream, Encoding.UTF8));
string gold = "This is a test";
StringBuilder builder = new StringBuilder();
int ch = 0;
while ((ch = reader.read()) != -1)
{
builder.Append((char)ch);
}
// Compare trim()'d output to gold
assertEquals("'" + builder.ToString().Trim() + "' is not equal to '" + gold + "'", gold, builder.ToString().Trim());
}
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: public void testRandomText() throws Exception
public virtual void testRandomText()
{
StringBuilder text = new StringBuilder();
int minNumWords = 10;
int maxNumWords = 10000;
int minWordLength = 3;
int maxWordLength = 20;
int numWords = TestUtil.Next(random(), minNumWords, maxNumWords);
switch (TestUtil.Next(random(), 0, 4))
{
case 0:
{
for (int wordNum = 0; wordNum < numWords; ++wordNum)
{
text.Append(TestUtil.randomUnicodeString(random(), maxWordLength));
text.Append(' ');
}
break;
}
case 1:
{
for (int wordNum = 0; wordNum < numWords; ++wordNum)
{
text.Append(TestUtil.randomRealisticUnicodeString(random(), minWordLength, maxWordLength));
text.Append(' ');
}
break;
}
default:
{ // ASCII 50% of the time
for (int wordNum = 0; wordNum < numWords; ++wordNum)
{
text.Append(TestUtil.randomSimpleString(random()));
text.Append(' ');
}
}
break;
}
Reader reader = new HTMLStripCharFilter(new StringReader(text.ToString()));
while (reader.read() != -1)
{
;
}
}
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: public void doTestOffsets(String in) throws Exception
public virtual void doTestOffsets(string @in)
{
HTMLStripCharFilter reader = new HTMLStripCharFilter(new System.IO.StreamReader(new StringReader(@in)));
int ch = 0;
int off = 0; // offset in the reader
int strOff = -1; // offset in the original string
while ((ch = reader.read()) != -1)
{
int correctedOff = reader.correctOffset(off);
if (ch == 'X')
{
strOff = @in.IndexOf('X',strOff + 1);
assertEquals(strOff, correctedOff);
}
off++;
}
}
//Some sanity checks, but not a full-fledged check
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: public void testHTML() throws Exception
public virtual void testHTML()
{
System.IO.Stream stream = this.GetType().getResourceAsStream("htmlStripReaderTest.html");
HTMLStripCharFilter reader = new HTMLStripCharFilter(new System.IO.StreamReader(stream, Encoding.UTF8));
StringBuilder builder = new StringBuilder();
int ch = -1;
while ((ch = reader.read()) != -1)
{
builder.Append((char)ch);
}
string str = builder.ToString();
assertTrue("Entity not properly escaped", str.IndexOf("<", StringComparison.Ordinal) == -1); //there is one > in the text
assertTrue("Forrest should have been stripped out", str.IndexOf("forrest", StringComparison.Ordinal) == -1 && str.IndexOf("Forrest", StringComparison.Ordinal) == -1);
assertTrue("File should start with 'Welcome to Solr' after trimming", str.Trim().StartsWith("Welcome to Solr", StringComparison.Ordinal));
assertTrue("File should start with 'Foundation.' after trimming", str.Trim().EndsWith("Foundation.", StringComparison.Ordinal));
}
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: public void doTestOffsets(String in) throws Exception
public virtual void doTestOffsets(string @in)
{
HTMLStripCharFilter reader = new HTMLStripCharFilter(new System.IO.StreamReader(new StringReader(@in)));
int ch = 0;
int off = 0; // offset in the reader
int strOff = -1; // offset in the original string
while ((ch = reader.read()) != -1)
{
int correctedOff = reader.correctOffset(off);
if (ch == 'X')
{
strOff = @in.IndexOf('X', strOff + 1);
assertEquals(strOff, correctedOff);
}
off++;
}
}
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: public void testReserved() throws Exception
public virtual void testReserved()
{
string test = "aaa bbb <reserved ccc=\"ddddd\"> eeee </reserved> ffff <reserved ggg=\"hhhh\"/> <other/>";
ISet <string> set = new HashSet <string>();
set.Add("reserved");
Reader reader = new HTMLStripCharFilter(new StringReader(test), set);
StringBuilder builder = new StringBuilder();
int ch = 0;
while ((ch = reader.read()) != -1)
{
builder.Append((char)ch);
}
string result = builder.ToString();
// System.out.println("Result: " + result);
assertTrue("Escaped tag not preserved: " + result.IndexOf("reserved", StringComparison.Ordinal), result.IndexOf("reserved", StringComparison.Ordinal) == 9);
assertTrue("Escaped tag not preserved: " + result.IndexOf("reserved", 15, StringComparison.Ordinal), result.IndexOf("reserved", 15, StringComparison.Ordinal) == 38);
assertTrue("Escaped tag not preserved: " + result.IndexOf("reserved", 41, StringComparison.Ordinal), result.IndexOf("reserved", 41, StringComparison.Ordinal) == 54);
assertTrue("Other tag should be removed", result.IndexOf("other", StringComparison.Ordinal) == -1);
}
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: static void assertLegalOffsets(String in) throws Exception
internal static void assertLegalOffsets(string @in)
{
int length = @in.Length;
HTMLStripCharFilter reader = new HTMLStripCharFilter(new System.IO.StreamReader(new StringReader(@in)));
int ch = 0;
int off = 0;
while ((ch = reader.read()) != -1)
{
int correction = reader.correctOffset(off);
assertTrue("invalid offset correction: " + off + "->" + correction + " for doc of length: " + length, correction <= length);
off++;
}
}
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: public void testReserved() throws Exception
public virtual void testReserved()
{
string test = "aaa bbb <reserved ccc=\"ddddd\"> eeee </reserved> ffff <reserved ggg=\"hhhh\"/> <other/>";
ISet<string> set = new HashSet<string>();
set.Add("reserved");
Reader reader = new HTMLStripCharFilter(new StringReader(test), set);
StringBuilder builder = new StringBuilder();
int ch = 0;
while ((ch = reader.read()) != -1)
{
builder.Append((char)ch);
}
string result = builder.ToString();
// System.out.println("Result: " + result);
assertTrue("Escaped tag not preserved: " + result.IndexOf("reserved", StringComparison.Ordinal), result.IndexOf("reserved", StringComparison.Ordinal) == 9);
assertTrue("Escaped tag not preserved: " + result.IndexOf("reserved", 15, StringComparison.Ordinal), result.IndexOf("reserved", 15, StringComparison.Ordinal) == 38);
assertTrue("Escaped tag not preserved: " + result.IndexOf("reserved", 41, StringComparison.Ordinal), result.IndexOf("reserved", 41, StringComparison.Ordinal) == 54);
assertTrue("Other tag should be removed", result.IndexOf("other", StringComparison.Ordinal) == -1);
}
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: public void testRandomText() throws Exception
public virtual void testRandomText()
{
StringBuilder text = new StringBuilder();
int minNumWords = 10;
int maxNumWords = 10000;
int minWordLength = 3;
int maxWordLength = 20;
int numWords = TestUtil.Next(random(), minNumWords, maxNumWords);
switch (TestUtil.Next(random(), 0, 4))
{
case 0:
{
for (int wordNum = 0 ; wordNum < numWords ; ++wordNum)
{
text.Append(TestUtil.randomUnicodeString(random(), maxWordLength));
text.Append(' ');
}
break;
}
case 1:
{
for (int wordNum = 0 ; wordNum < numWords ; ++wordNum)
{
text.Append(TestUtil.randomRealisticUnicodeString(random(), minWordLength, maxWordLength));
text.Append(' ');
}
break;
}
default:
{ // ASCII 50% of the time
for (int wordNum = 0 ; wordNum < numWords ; ++wordNum)
{
text.Append(TestUtil.randomSimpleString(random()));
text.Append(' ');
}
}
break;
}
Reader reader = new HTMLStripCharFilter(new StringReader(text.ToString()));
while (reader.read() != -1);
}
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: public void testMSWord14GeneratedHTML() throws Exception
public virtual void testMSWord14GeneratedHTML()
{
System.IO.Stream stream = this.GetType().getResourceAsStream("MS-Word 14 generated.htm");
HTMLStripCharFilter reader = new HTMLStripCharFilter(new System.IO.StreamReader(stream, Encoding.UTF8));
string gold = "This is a test";
StringBuilder builder = new StringBuilder();
int ch = 0;
while ((ch = reader.read()) != -1)
{
builder.Append((char)ch);
}
// Compare trim()'d output to gold
assertEquals("'" + builder.ToString().Trim() + "' is not equal to '" + gold + "'", gold, builder.ToString().Trim());
}
//Some sanity checks, but not a full-fledged check
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in .NET:
//ORIGINAL LINE: public void testHTML() throws Exception
public virtual void testHTML()
{
System.IO.Stream stream = this.GetType().getResourceAsStream("htmlStripReaderTest.html");
HTMLStripCharFilter reader = new HTMLStripCharFilter(new System.IO.StreamReader(stream, Encoding.UTF8));
StringBuilder builder = new StringBuilder();
int ch = -1;
while ((ch = reader.read()) != -1)
{
builder.Append((char)ch);
}
string str = builder.ToString();
assertTrue("Entity not properly escaped", str.IndexOf("<", StringComparison.Ordinal) == -1); //there is one > in the text
assertTrue("Forrest should have been stripped out", str.IndexOf("forrest", StringComparison.Ordinal) == -1 && str.IndexOf("Forrest", StringComparison.Ordinal) == -1);
assertTrue("File should start with 'Welcome to Solr' after trimming", str.Trim().StartsWith("Welcome to Solr", StringComparison.Ordinal));
assertTrue("File should start with 'Foundation.' after trimming", str.Trim().EndsWith("Foundation.", StringComparison.Ordinal));
}
