public virtual void TestRandomUnicodeStrings()
{
char[] buffer = new char[20];
char[] expected = new char[20];
BytesRef utf8 = new BytesRef(20);
CharsRef utf16 = new CharsRef(20);
int num = AtLeast(100000);
for (int iter = 0; iter < num; iter++)
{
bool hasIllegal = FillUnicode(buffer, expected, 0, 20);
UnicodeUtil.UTF16toUTF8(buffer, 0, 20, utf8);
if (!hasIllegal)
{
#pragma warning disable 612, 618
var b = (new string(buffer, 0, 20)).GetBytes(IOUtils.CHARSET_UTF_8);
#pragma warning restore 612, 618
Assert.AreEqual(b.Length, utf8.Length);
for (int i = 0; i < b.Length; i++)
{
Assert.AreEqual(b[i], utf8.Bytes[i]);
}
}
UnicodeUtil.UTF8toUTF16(utf8.Bytes, 0, utf8.Length, utf16);
Assert.AreEqual(utf16.Length, 20);
for (int i = 0; i < 20; i++)
{
Assert.AreEqual(expected[i], utf16.Chars[i]);
}
}
}
/// <summary>Writes a string.</summary>
/// <seealso cref="IndexInput.ReadString()">
/// </seealso>
public virtual void WriteString(System.String s)
{
UnicodeUtil.UTF8Result utf8Result = new UnicodeUtil.UTF8Result();
UnicodeUtil.UTF16toUTF8(s, 0, s.Length, utf8Result);
WriteVInt(utf8Result.length);
WriteBytes(utf8Result.result, 0, utf8Result.length);
}
/// <summary>
/// Adds an input string and it's stemmer override output to this builder.
/// </summary>
/// <param name="input"> the input char sequence </param>
/// <param name="output"> the stemmer override output char sequence </param>
/// <returns> <code>false</code> iff the input has already been added to this builder otherwise <code>true</code>. </returns>
public virtual bool add(ICharSequence input, ICharSequence output)
{
int length = input.length();
if (ignoreCase)
{
// convert on the fly to lowercase
charsSpare.grow(length);
char[] buffer = charsSpare.chars;
for (int i = 0; i < length;)
{
i += char.toChars(char.ToLower(char.codePointAt(input, i)), buffer, i);
}
UnicodeUtil.UTF16toUTF8(buffer, 0, length, spare);
}
else
{
UnicodeUtil.UTF16toUTF8(input, 0, length, spare);
}
if (hash.add(spare) >= 0)
{
outputValues.Add(output);
return(true);
}
return(false);
}
/// <summary>
/// Compresses the <see cref="string"/> value using the specified
/// <paramref name="compressionLevel"/>.
/// </summary>
public static byte[] CompressString(string value, CompressionLevel compressionLevel)
{
var result = new BytesRef();
UnicodeUtil.UTF16toUTF8(value.ToCharArray(), 0, value.Length, result);
return(Compress(result.Bytes, 0, result.Length, compressionLevel));
}
public void Read(IndexInput input, FieldInfos fieldInfos)
{
this.term = null; // invalidate cache
int start = input.ReadVInt();
int length = input.ReadVInt();
int totalLength = start + length;
if (preUTF8Strings)
{
text.SetLength(totalLength);
input.ReadChars(text.result, start, length);
}
else
{
if (dirty)
{
// Fully convert all bytes since bytes is dirty
UnicodeUtil.UTF16toUTF8(text.result, 0, text.length, bytes);
bytes.SetLength(totalLength);
input.ReadBytes(bytes.result, start, length);
UnicodeUtil.UTF8toUTF16(bytes.result, 0, totalLength, text);
dirty = false;
}
else
{
// Incrementally convert only the UTF8 bytes that are new:
bytes.SetLength(totalLength);
input.ReadBytes(bytes.result, start, length);
UnicodeUtil.UTF8toUTF16(bytes.result, start, length, text);
}
}
this.field = fieldInfos.FieldName(input.ReadVInt());
}
/// <summary>
/// Adds an input string and it's stemmer override output to this builder.
/// </summary>
/// <param name="input"> the input char sequence </param>
/// <param name="output"> the stemmer override output char sequence </param>
/// <returns> <code>false</code> iff the input has already been added to this builder otherwise <code>true</code>. </returns>
public virtual bool add(CharSequence input, CharSequence output)
{
//JAVA TO C# CONVERTER WARNING: The original Java variable was marked 'final':
//ORIGINAL LINE: final int length = input.length();
int length = input.length();
if (ignoreCase)
{
// convert on the fly to lowercase
charsSpare.grow(length);
//JAVA TO C# CONVERTER WARNING: The original Java variable was marked 'final':
//ORIGINAL LINE: final char[] buffer = charsSpare.chars;
char[] buffer = charsSpare.chars;
for (int i = 0; i < length;)
{
i += char.toChars(char.ToLower(char.codePointAt(input, i)), buffer, i);
}
UnicodeUtil.UTF16toUTF8(buffer, 0, length, spare);
}
else
{
UnicodeUtil.UTF16toUTF8(input, 0, length, spare);
}
if (hash.add(spare) >= 0)
{
outputValues.Add(output);
return(true);
}
return(false);
}
/// <summary>
/// Adds an input string and it's stemmer override output to this builder.
/// </summary>
/// <param name="input"> the input char sequence </param>
/// <param name="output"> the stemmer override output char sequence </param>
/// <returns> <c>false</c> if the input has already been added to this builder otherwise <c>true</c>. </returns>
public virtual bool Add(string input, string output)
{
int length = input.Length;
if (ignoreCase)
{
// convert on the fly to lowercase
charsSpare.Grow(length);
char[] buffer = charsSpare.Chars;
for (int i = 0; i < length;)
{
i += Character.ToChars(Character.ToLower(Character.CodePointAt(input, i), CultureInfo.InvariantCulture), buffer, i);
}
UnicodeUtil.UTF16toUTF8(buffer, 0, length, spare);
}
else
{
UnicodeUtil.UTF16toUTF8(input.ToCharArray(), 0, length, spare);
}
if (hash.Add(spare) >= 0)
{
outputValues.Add(output);
return(true);
}
return(false);
}
/// <summary>
/// Writes a string.
/// <p>
/// Writes strings as UTF-8 encoded bytes. First the length, in bytes, is
/// written as a <seealso cref="#writeVInt VInt"/>, followed by the bytes.
/// </summary>
/// <seealso cref= DataInput#readString() </seealso>
public virtual void WriteString(string s)
{
BytesRef utf8Result = new BytesRef(10);
UnicodeUtil.UTF16toUTF8(s.ToCharArray(), 0, s.Length, utf8Result);
WriteVInt(utf8Result.Length);
WriteBytes(utf8Result.Bytes, 0, utf8Result.Length);
}
/// <summary>
/// Writes a string.
/// <para/>
/// Writes strings as UTF-8 encoded bytes. First the length, in bytes, is
/// written as a <see cref="WriteVInt32"/>, followed by the bytes.
/// </summary>
/// <seealso cref="DataInput.ReadString()"/>
public virtual void WriteString(string s)
{
var utf8Result = new BytesRef(10);
UnicodeUtil.UTF16toUTF8(s, 0, s.Length, utf8Result);
WriteVInt32(utf8Result.Length);
WriteBytes(utf8Result.Bytes, 0, utf8Result.Length);
}
// NOTE: while it's tempting to make this public, since
// caller's parser likely knows the
// numInput/numOutputWords, sneaky exceptions, much later
// on, will result if these values are wrong; so we always
// recompute ourselves to be safe:
internal virtual void Add(CharsRef input, int numInputWords, CharsRef output, int numOutputWords, bool includeOrig)
{
// first convert to UTF-8
if (numInputWords <= 0)
{
throw new ArgumentOutOfRangeException(nameof(numInputWords), "numInputWords must be > 0 (got " + numInputWords + ")"); // LUCENENET specific - changed from IllegalArgumentException to ArgumentOutOfRangeException (.NET convention)
}
if (input.Length <= 0)
{
throw new ArgumentOutOfRangeException(nameof(input.Length), "input.Length must be > 0 (got " + input.Length + ")"); // LUCENENET specific - changed from IllegalArgumentException to ArgumentOutOfRangeException (.NET convention)
}
if (numOutputWords <= 0)
{
throw new ArgumentOutOfRangeException(nameof(numOutputWords), "numOutputWords must be > 0 (got " + numOutputWords + ")"); // LUCENENET specific - changed from IllegalArgumentException to ArgumentOutOfRangeException (.NET convention)
}
if (output.Length <= 0)
{
throw new ArgumentOutOfRangeException(nameof(output.Length), "output.Length must be > 0 (got " + output.Length + ")"); // LUCENENET specific - changed from IllegalArgumentException to ArgumentOutOfRangeException (.NET convention)
}
if (Debugging.AssertsEnabled)
{
Debugging.Assert(!HasHoles(input), "input has holes: {0}", input);
Debugging.Assert(!HasHoles(output), "output has holes: {0}", output);
}
//System.out.println("fmap.add input=" + input + " numInputWords=" + numInputWords + " output=" + output + " numOutputWords=" + numOutputWords);
UnicodeUtil.UTF16toUTF8(output.Chars, output.Offset, output.Length, utf8Scratch);
// lookup in hash
int ord = words.Add(utf8Scratch);
if (ord < 0)
{
// already exists in our hash
ord = (-ord) - 1;
//System.out.println(" output=" + output + " old ord=" + ord);
}
else
{
//System.out.println(" output=" + output + " new ord=" + ord);
}
if (!workingSet.TryGetValue(input, out MapEntry e) || e is null)
{
e = new MapEntry();
workingSet[CharsRef.DeepCopyOf(input)] = e; // make a copy, since we will keep around in our map
}
e.ords.Add(ord);
e.includeOrig |= includeOrig;
maxHorizontalContext = Math.Max(maxHorizontalContext, numInputWords);
maxHorizontalContext = Math.Max(maxHorizontalContext, numOutputWords);
}
// NOTE: while it's tempting to make this public, since
// caller's parser likely knows the
// numInput/numOutputWords, sneaky exceptions, much later
// on, will result if these values are wrong; so we always
// recompute ourselves to be safe:
internal virtual void Add(CharsRef input, int numInputWords, CharsRef output, int numOutputWords, bool includeOrig)
{
// first convert to UTF-8
if (numInputWords <= 0)
{
throw new System.ArgumentException("numInputWords must be > 0 (got " + numInputWords + ")");
}
if (input.Length <= 0)
{
throw new System.ArgumentException("input.length must be > 0 (got " + input.Length + ")");
}
if (numOutputWords <= 0)
{
throw new System.ArgumentException("numOutputWords must be > 0 (got " + numOutputWords + ")");
}
if (output.Length <= 0)
{
throw new System.ArgumentException("output.length must be > 0 (got " + output.Length + ")");
}
Debug.Assert(!HasHoles(input), "input has holes: " + input);
Debug.Assert(!HasHoles(output), "output has holes: " + output);
//System.out.println("fmap.add input=" + input + " numInputWords=" + numInputWords + " output=" + output + " numOutputWords=" + numOutputWords);
UnicodeUtil.UTF16toUTF8(output.Chars, output.Offset, output.Length, utf8Scratch);
// lookup in hash
int ord = words.Add(utf8Scratch);
if (ord < 0)
{
// already exists in our hash
ord = (-ord) - 1;
//System.out.println(" output=" + output + " old ord=" + ord);
}
else
{
//System.out.println(" output=" + output + " new ord=" + ord);
}
MapEntry e = workingSet.ContainsKey(input) ? workingSet[input] : null;
if (e == null)
{
e = new MapEntry();
workingSet[CharsRef.DeepCopyOf(input)] = e; // make a copy, since we will keep around in our map
}
e.ords.Add(ord);
e.includeOrig |= includeOrig;
maxHorizontalContext = Math.Max(maxHorizontalContext, numInputWords);
maxHorizontalContext = Math.Max(maxHorizontalContext, numOutputWords);
}
public virtual void Reset(string token)
{
Debug.Assert(Level != 0);
this.token = token;
shapeRel = SpatialRelation.NULL_VALUE;
//converting string t0 byte[]
//bytes = Encoding.UTF8.GetBytes(token);
BytesRef utf8Result = new BytesRef(token.Length);
UnicodeUtil.UTF16toUTF8(token.ToCharArray(), 0, token.Length, utf8Result);
bytes = utf8Result.bytes.ToByteArray();
b_off = 0;
b_len = bytes.Length;
B_fixLeaf();
}
/// <summary>Called when we are done adding docs to this term </summary>
internal override void Finish()
{
long skipPointer = skipListWriter.WriteSkip(out_Renamed);
// TODO: this is abstraction violation -- we should not
// peek up into parents terms encoding format
termInfo.Set(df, parent.freqStart, parent.proxStart, (int)(skipPointer - parent.freqStart));
// TODO: we could do this incrementally
UnicodeUtil.UTF16toUTF8(parent.currentTerm, parent.currentTermStart, utf8);
if (df > 0)
{
parent.termsOut.Add(fieldInfo.number, utf8.result, utf8.length, termInfo);
}
lastDocID = 0;
df = 0;
}
public virtual void TestAllUnicodeChars()
{
BytesRef utf8 = new BytesRef(10);
CharsRef utf16 = new CharsRef(10);
char[] chars = new char[2];
for (int ch = 0; ch < 0x0010FFFF; ch++)
{
if (ch == 0xd800)
// Skip invalid code points
{
ch = 0xe000;
}
int len = 0;
if (ch <= 0xffff)
{
chars[len++] = (char)ch;
}
else
{
chars[len++] = (char)(((ch - 0x0010000) >> 10) + UnicodeUtil.UNI_SUR_HIGH_START);
chars[len++] = (char)(((ch - 0x0010000) & 0x3FFL) + UnicodeUtil.UNI_SUR_LOW_START);
}
UnicodeUtil.UTF16toUTF8(chars, 0, len, utf8);
string s1 = new string(chars, 0, len);
string s2 = Encoding.UTF8.GetString(utf8.Bytes, utf8.Offset, utf8.Length);
Assert.AreEqual(s1, s2, "codepoint " + ch);
UnicodeUtil.UTF8toUTF16(utf8.Bytes, 0, utf8.Length, utf16);
Assert.AreEqual(s1, new string(utf16.Chars, 0, utf16.Length), "codepoint " + ch);
var b = s1.GetBytes(Encoding.UTF8);
Assert.AreEqual(utf8.Length, b.Length);
for (int j = 0; j < utf8.Length; j++)
{
Assert.AreEqual(utf8.Bytes[j], b[j]);
}
}
}
/// <summary>Note: doesn't contain a trailing leaf byte.</summary>
/// <remarks>Note: doesn't contain a trailing leaf byte.</remarks>
public virtual byte[] GetTokenBytes()
{
if (bytes != null)
{
if (b_off != 0 || b_len != bytes.Length)
{
throw new InvalidOperationException("Not supported if byte[] needs to be recreated.");
}
}
else
{
//converting string t0 byte[]
//bytes = Encoding.UTF8.GetBytes(token);
BytesRef utf8Result = new BytesRef(token.Length);
UnicodeUtil.UTF16toUTF8(token.ToCharArray(), 0, token.Length, utf8Result);
bytes = utf8Result.bytes.ToByteArray();
b_off = 0;
b_len = bytes.Length;
}
return(bytes);
}
/// <summary>Compresses the String value using the specified
/// compressionLevel (constants are defined in
/// java.util.zip.Deflater).
/// </summary>
public static byte[] CompressString(System.String value_Renamed, int compressionLevel)
{
UnicodeUtil.UTF8Result result = new UnicodeUtil.UTF8Result();
UnicodeUtil.UTF16toUTF8(value_Renamed, 0, value_Renamed.Length, result);
return(Compress(result.result, 0, result.length, compressionLevel));
}
public static void Write(DataOutput output, string s, BytesRef scratch)
{
UnicodeUtil.UTF16toUTF8(s.ToCharArray(), 0, s.Length, scratch);
Write(output, scratch);
}
/// <summary> Add a complete document specified by all its term vectors. If document has no
/// term vectors, add value for tvx.
///
/// </summary>
/// <param name="vectors">
/// </param>
/// <throws> IOException </throws>
public void AddAllDocVectors(ITermFreqVector[] vectors)
{
tvx.WriteLong(tvd.FilePointer);
tvx.WriteLong(tvf.FilePointer);
if (vectors != null)
{
int numFields = vectors.Length;
tvd.WriteVInt(numFields);
var fieldPointers = new long[numFields];
for (int i = 0; i < numFields; i++)
{
fieldPointers[i] = tvf.FilePointer;
int fieldNumber = fieldInfos.FieldNumber(vectors[i].Field);
// 1st pass: write field numbers to tvd
tvd.WriteVInt(fieldNumber);
int numTerms = vectors[i].Size;
tvf.WriteVInt(numTerms);
TermPositionVector tpVector;
byte bits;
bool storePositions;
bool storeOffsets;
if (vectors[i] is TermPositionVector)
{
// May have positions & offsets
tpVector = (TermPositionVector)vectors[i];
storePositions = tpVector.Size > 0 && tpVector.GetTermPositions(0) != null;
storeOffsets = tpVector.Size > 0 && tpVector.GetOffsets(0) != null;
bits = (byte)((storePositions?TermVectorsReader.STORE_POSITIONS_WITH_TERMVECTOR: (byte)0) + (storeOffsets?TermVectorsReader.STORE_OFFSET_WITH_TERMVECTOR: (byte)0));
}
else
{
tpVector = null;
bits = 0;
storePositions = false;
storeOffsets = false;
}
tvf.WriteVInt(bits);
System.String[] terms = vectors[i].GetTerms();
int[] freqs = vectors[i].GetTermFrequencies();
int utf8Upto = 0;
utf8Results[1].length = 0;
for (int j = 0; j < numTerms; j++)
{
UnicodeUtil.UTF16toUTF8(terms[j], 0, terms[j].Length, utf8Results[utf8Upto]);
int start = StringHelper.BytesDifference(utf8Results[1 - utf8Upto].result, utf8Results[1 - utf8Upto].length, utf8Results[utf8Upto].result, utf8Results[utf8Upto].length);
int length = utf8Results[utf8Upto].length - start;
tvf.WriteVInt(start); // write shared prefix length
tvf.WriteVInt(length); // write delta length
tvf.WriteBytes(utf8Results[utf8Upto].result, start, length); // write delta bytes
utf8Upto = 1 - utf8Upto;
int termFreq = freqs[j];
tvf.WriteVInt(termFreq);
if (storePositions)
{
int[] positions = tpVector.GetTermPositions(j);
if (positions == null)
{
throw new System.SystemException("Trying to write positions that are null!");
}
System.Diagnostics.Debug.Assert(positions.Length == termFreq);
// use delta encoding for positions
int lastPosition = 0;
foreach (int position in positions)
{
tvf.WriteVInt(position - lastPosition);
lastPosition = position;
}
}
if (storeOffsets)
{
TermVectorOffsetInfo[] offsets = tpVector.GetOffsets(j);
if (offsets == null)
{
throw new System.SystemException("Trying to write offsets that are null!");
}
System.Diagnostics.Debug.Assert(offsets.Length == termFreq);
// use delta encoding for offsets
int lastEndOffset = 0;
foreach (TermVectorOffsetInfo t in offsets)
{
int startOffset = t.StartOffset;
int endOffset = t.EndOffset;
tvf.WriteVInt(startOffset - lastEndOffset);
tvf.WriteVInt(endOffset - startOffset);
lastEndOffset = endOffset;
}
}
}
}
// 2nd pass: write field pointers to tvd
if (numFields > 1)
{
long lastFieldPointer = fieldPointers[0];
for (int i = 1; i < numFields; i++)
{
long fieldPointer = fieldPointers[i];
tvd.WriteVLong(fieldPointer - lastFieldPointer);
lastFieldPointer = fieldPointer;
}
}
}
else
{
tvd.WriteVInt(0);
}
}
public virtual void FillBytesRef()
{
UnicodeUtil.UTF16toUTF8(TermBuffer, 0, TermLength, Bytes);
}
/// <summary>Called once per field per document if term vectors
/// are enabled, to write the vectors to
/// RAMOutputStream, which is then quickly flushed to
/// the real term vectors files in the Directory.
/// </summary>
internal override void Finish()
{
System.Diagnostics.Debug.Assert(docState.TestPoint("TermVectorsTermsWriterPerField.finish start"));
int numPostings = termsHashPerField.numPostings;
System.Diagnostics.Debug.Assert(numPostings >= 0);
if (!doVectors || numPostings == 0)
{
return;
}
if (numPostings > maxNumPostings)
{
maxNumPostings = numPostings;
}
IndexOutput tvf = perThread.doc.perDocTvf;
// This is called once, after inverting all occurences
// of a given field in the doc. At this point we flush
// our hash into the DocWriter.
System.Diagnostics.Debug.Assert(fieldInfo.storeTermVector);
System.Diagnostics.Debug.Assert(perThread.VectorFieldsInOrder(fieldInfo));
perThread.doc.AddField(termsHashPerField.fieldInfo.number);
RawPostingList[] postings = termsHashPerField.SortPostings();
tvf.WriteVInt(numPostings);
byte bits = (byte)(0x0);
if (doVectorPositions)
{
bits |= TermVectorsReader.STORE_POSITIONS_WITH_TERMVECTOR;
}
if (doVectorOffsets)
{
bits |= TermVectorsReader.STORE_OFFSET_WITH_TERMVECTOR;
}
tvf.WriteByte(bits);
int encoderUpto = 0;
int lastTermBytesCount = 0;
ByteSliceReader reader = perThread.vectorSliceReader;
char[][] charBuffers = perThread.termsHashPerThread.charPool.buffers;
for (int j = 0; j < numPostings; j++)
{
TermVectorsTermsWriter.PostingList posting = (TermVectorsTermsWriter.PostingList)postings[j];
int freq = posting.freq;
char[] text2 = charBuffers[posting.textStart >> DocumentsWriter.CHAR_BLOCK_SHIFT];
int start2 = posting.textStart & DocumentsWriter.CHAR_BLOCK_MASK;
// We swap between two encoders to save copying
// last Term's byte array
UnicodeUtil.UTF8Result utf8Result = perThread.utf8Results[encoderUpto];
// TODO: we could do this incrementally
UnicodeUtil.UTF16toUTF8(text2, start2, utf8Result);
int termBytesCount = utf8Result.length;
// TODO: UTF16toUTF8 could tell us this prefix
// Compute common prefix between last term and
// this term
int prefix = 0;
if (j > 0)
{
byte[] lastTermBytes = perThread.utf8Results[1 - encoderUpto].result;
byte[] termBytes = perThread.utf8Results[encoderUpto].result;
while (prefix < lastTermBytesCount && prefix < termBytesCount)
{
if (lastTermBytes[prefix] != termBytes[prefix])
{
break;
}
prefix++;
}
}
encoderUpto = 1 - encoderUpto;
lastTermBytesCount = termBytesCount;
int suffix = termBytesCount - prefix;
tvf.WriteVInt(prefix);
tvf.WriteVInt(suffix);
tvf.WriteBytes(utf8Result.result, prefix, suffix);
tvf.WriteVInt(freq);
if (doVectorPositions)
{
termsHashPerField.InitReader(reader, posting, 0);
reader.WriteTo(tvf);
}
if (doVectorOffsets)
{
termsHashPerField.InitReader(reader, posting, 1);
reader.WriteTo(tvf);
}
}
termsHashPerField.Reset();
// NOTE: we clear, per-field, at the thread level,
// because term vectors fully write themselves on each
// field; this saves RAM (eg if large doc has two large
// fields w/ term vectors on) because we recycle/reuse
// all RAM after each field:
perThread.termsHashPerThread.Reset(false);
}
private bool Matches(ByteRunAutomaton a, int code)
{
char[] chars = Character.ToChars(code);
UnicodeUtil.UTF16toUTF8(chars, 0, chars.Length, b);
return(a.Run(b.Bytes, 0, b.Length));
}
internal void Add(Term term, TermInfo ti)
{
UnicodeUtil.UTF16toUTF8(term.Text, 0, term.Text.Length, utf8Result);
Add(fieldInfos.FieldNumber(term.Field), utf8Result.result, utf8Result.length, ti);
}
