/// <summary>Retrieve the length (in bytes) of the tvd and tvf
/// entries for the next numDocs starting with
/// startDocID. This is used for bulk copying when
/// merging segments, if the field numbers are
/// congruent. Once this returns, the tvf & tvd streams
/// are seeked to the startDocID.
/// </summary>
internal void RawDocs(int[] tvdLengths, int[] tvfLengths, int startDocID, int numDocs)
{
if (tvx == null)
{
for (int i = 0; i < tvdLengths.Length; i++)
{
tvdLengths[i] = 0;
}
for (int i = 0; i < tvfLengths.Length; i++)
{
tvfLengths[i] = 0;
}
return;
}
// SegmentMerger calls canReadRawDocs() first and should
// not call us if that returns false.
if (format < FORMAT_VERSION2)
{
throw new System.SystemException("cannot read raw docs with older term vector formats");
}
SeekTvx(startDocID);
long tvdPosition = tvx.ReadLong();
tvd.Seek(tvdPosition);
long tvfPosition = tvx.ReadLong();
tvf.Seek(tvfPosition);
long lastTvdPosition = tvdPosition;
long lastTvfPosition = tvfPosition;
int count = 0;
while (count < numDocs)
{
int docID = docStoreOffset + startDocID + count + 1;
System.Diagnostics.Debug.Assert(docID <= numTotalDocs);
if (docID < numTotalDocs)
{
tvdPosition = tvx.ReadLong();
tvfPosition = tvx.ReadLong();
}
else
{
tvdPosition = tvd.Length();
tvfPosition = tvf.Length();
System.Diagnostics.Debug.Assert(count == numDocs - 1);
}
tvdLengths[count] = (int)(tvdPosition - lastTvdPosition);
tvfLengths[count] = (int)(tvfPosition - lastTvfPosition);
count++;
lastTvdPosition = tvdPosition;
lastTvfPosition = tvfPosition;
}
}
public CompoundFileReader(Directory dir, System.String name, int readBufferSize)
{
directory = dir;
fileName = name;
this.readBufferSize = readBufferSize;
bool success = false;
try
{
stream = dir.OpenInput(name, readBufferSize);
// read the directory and init files
int count = stream.ReadVInt();
FileEntry entry = null;
for (int i = 0; i < count; i++)
{
long offset = stream.ReadLong();
System.String id = stream.ReadString();
if (entry != null)
{
// set length of the previous entry
entry.length = offset - entry.offset;
}
entry = new FileEntry();
entry.offset = offset;
entries[id] = entry;
}
// set the length of the final entry
if (entry != null)
{
entry.length = stream.Length() - entry.offset;
}
success = true;
}
finally
{
if (!success && (stream != null))
{
try
{
stream.Close();
}
catch (System.IO.IOException e)
{
}
}
}
}
private void Read(IndexInput input, System.String fileName)
{
int firstInt = input.ReadVInt();
if (firstInt < 0)
{
// This is a real format
format = firstInt;
}
else
{
format = FORMAT_PRE;
}
if (format != FORMAT_PRE & format != FORMAT_START)
{
throw new CorruptIndexException("unrecognized format " + format + " in file \"" + fileName + "\"");
}
int size;
if (format == FORMAT_PRE)
{
size = firstInt;
}
else
{
size = input.ReadVInt(); //read in the size
}
for (int i = 0; i < size; i++)
{
System.String name = StringHelper.Intern(input.ReadString());
byte bits = input.ReadByte();
bool isIndexed = (bits & IS_INDEXED) != 0;
bool storeTermVector = (bits & STORE_TERMVECTOR) != 0;
bool storePositionsWithTermVector = (bits & STORE_POSITIONS_WITH_TERMVECTOR) != 0;
bool storeOffsetWithTermVector = (bits & STORE_OFFSET_WITH_TERMVECTOR) != 0;
bool omitNorms = (bits & OMIT_NORMS) != 0;
bool storePayloads = (bits & STORE_PAYLOADS) != 0;
bool omitTermFreqAndPositions = (bits & OMIT_TERM_FREQ_AND_POSITIONS) != 0;
AddInternal(name, isIndexed, storeTermVector, storePositionsWithTermVector, storeOffsetWithTermVector, omitNorms, storePayloads, omitTermFreqAndPositions);
}
if (input.GetFilePointer() != input.Length())
{
throw new CorruptIndexException("did not read all bytes from file \"" + fileName + "\": read " + input.GetFilePointer() + " vs size " + input.Length());
}
}
/// <summary>Copy the contents of the file with specified extension into the
/// provided output stream. Use the provided buffer for moving data
/// to reduce memory allocation.
/// </summary>
private void CopyFile(FileEntry source, IndexOutput os, byte[] buffer)
{
IndexInput is_Renamed = null;
try
{
long startPtr = os.GetFilePointer();
is_Renamed = directory.OpenInput(source.file);
long length = is_Renamed.Length();
long remainder = length;
int chunk = buffer.Length;
while (remainder > 0)
{
int len = (int)System.Math.Min(chunk, remainder);
is_Renamed.ReadBytes(buffer, 0, len, false);
os.WriteBytes(buffer, len);
remainder -= len;
if (checkAbort != null)
{
// Roughly every 2 MB we will check if
// it's time to abort
checkAbort.Work(80);
}
}
// Verify that remainder is 0
if (remainder != 0)
{
throw new System.IO.IOException("Non-zero remainder length after copying: " + remainder + " (id: " + source.file + ", length: " + length + ", buffer size: " + chunk + ")");
}
// Verify that the output length diff is equal to original file
long endPtr = os.GetFilePointer();
long diff = endPtr - startPtr;
if (diff != length)
{
throw new System.IO.IOException("Difference in the output file offsets " + diff + " does not match the original file length " + length);
}
}
finally
{
if (is_Renamed != null)
{
is_Renamed.Close();
}
}
}
private void Read(IndexInput input, System.String fileName)
{
int firstInt = input.ReadVInt();
if (firstInt < 0)
{
// This is a real format
format = firstInt;
}
else
{
format = FORMAT_PRE;
}
if (format != FORMAT_PRE & format != FORMAT_START)
{
throw new CorruptIndexException("unrecognized format " + format + " in file \"" + fileName + "\"");
}
int size;
if (format == FORMAT_PRE)
{
size = firstInt;
}
else
{
size = input.ReadVInt(); //read in the size
}
for (int i = 0; i < size; i++)
{
System.String name = StringHelper.Intern(input.ReadString());
byte bits = input.ReadByte();
bool isIndexed = (bits & IS_INDEXED) != 0;
bool storeTermVector = (bits & STORE_TERMVECTOR) != 0;
bool storePositionsWithTermVector = (bits & STORE_POSITIONS_WITH_TERMVECTOR) != 0;
bool storeOffsetWithTermVector = (bits & STORE_OFFSET_WITH_TERMVECTOR) != 0;
bool omitNorms = (bits & OMIT_NORMS) != 0;
bool storePayloads = (bits & STORE_PAYLOADS) != 0;
bool omitTermFreqAndPositions = (bits & OMIT_TERM_FREQ_AND_POSITIONS) != 0;
AddInternal(name, isIndexed, storeTermVector, storePositionsWithTermVector, storeOffsetWithTermVector, omitNorms, storePayloads, omitTermFreqAndPositions);
}
if (input.GetFilePointer() != input.Length())
{
throw new CorruptIndexException("did not read all bytes from file \"" + fileName + "\": read " + input.GetFilePointer() + " vs size " + input.Length());
}
}
internal TermVectorsReader(Directory d, System.String segment, FieldInfos fieldInfos, int readBufferSize, int docStoreOffset, int size)
{
bool success = false;
try
{
if (d.FileExists(segment + "." + IndexFileNames.VECTORS_INDEX_EXTENSION))
{
tvx = d.OpenInput(segment + "." + IndexFileNames.VECTORS_INDEX_EXTENSION, readBufferSize);
format = CheckValidFormat(tvx);
tvd = d.OpenInput(segment + "." + IndexFileNames.VECTORS_DOCUMENTS_EXTENSION, readBufferSize);
int tvdFormat = CheckValidFormat(tvd);
tvf = d.OpenInput(segment + "." + IndexFileNames.VECTORS_FIELDS_EXTENSION, readBufferSize);
int tvfFormat = CheckValidFormat(tvf);
System.Diagnostics.Debug.Assert(format == tvdFormat);
System.Diagnostics.Debug.Assert(format == tvfFormat);
if (format >= FORMAT_VERSION2)
{
System.Diagnostics.Debug.Assert((tvx.Length() - FORMAT_SIZE) % 16 == 0);
numTotalDocs = (int)(tvx.Length() >> 4);
}
else
{
System.Diagnostics.Debug.Assert((tvx.Length() - FORMAT_SIZE) % 8 == 0);
numTotalDocs = (int)(tvx.Length() >> 3);
}
if (-1 == docStoreOffset)
{
this.docStoreOffset = 0;
this.size = numTotalDocs;
System.Diagnostics.Debug.Assert(size == 0 || numTotalDocs == size);
}
else
{
this.docStoreOffset = docStoreOffset;
this.size = size;
// Verify the file is long enough to hold all of our
// docs
System.Diagnostics.Debug.Assert(numTotalDocs >= size + docStoreOffset, "numTotalDocs=" + numTotalDocs + " size=" + size + " docStoreOffset=" + docStoreOffset);
}
}
else
{
// If all documents flushed in a segment had hit
// non-aborting exceptions, it's possible that
// FieldInfos.hasVectors returns true yet the term
// vector files don't exist.
format = 0;
}
this.fieldInfos = fieldInfos;
success = true;
}
finally
{
// With lock-less commits, it's entirely possible (and
// fine) to hit a FileNotFound exception above. In
// this case, we want to explicitly close any subset
// of things that were opened so that we don't have to
// wait for a GC to do so.
if (!success)
{
Close();
}
}
}
internal FieldsReader(Directory d, System.String segment, FieldInfos fn, int readBufferSize, int docStoreOffset, int size)
{
bool success = false;
isOriginal = true;
try
{
fieldInfos = fn;
cloneableFieldsStream = d.OpenInput(segment + "." + IndexFileNames.FIELDS_EXTENSION, readBufferSize);
cloneableIndexStream = d.OpenInput(segment + "." + IndexFileNames.FIELDS_INDEX_EXTENSION, readBufferSize);
// First version of fdx did not include a format
// header, but, the first int will always be 0 in that
// case
int firstInt = cloneableIndexStream.ReadInt();
if (firstInt == 0)
format = 0;
else
format = firstInt;
if (format > FieldsWriter.FORMAT_CURRENT
/* extra support for Lucene 3.0 indexes: */ && format != FieldsWriter.FORMAT_LUCENE_3_0_NO_COMPRESSED_FIELDS
)
throw new CorruptIndexException("Incompatible format version: " + format + " expected " + FieldsWriter.FORMAT_CURRENT + " or lower");
if (format > FieldsWriter.FORMAT)
formatSize = 4;
else
formatSize = 0;
if (format < FieldsWriter.FORMAT_VERSION_UTF8_LENGTH_IN_BYTES)
cloneableFieldsStream.SetModifiedUTF8StringsMode();
fieldsStream = (IndexInput) cloneableFieldsStream.Clone();
long indexSize = cloneableIndexStream.Length() - formatSize;
if (docStoreOffset != - 1)
{
// We read only a slice out of this shared fields file
this.docStoreOffset = docStoreOffset;
this.size = size;
// Verify the file is long enough to hold all of our
// docs
System.Diagnostics.Debug.Assert(((int)(indexSize / 8)) >= size + this.docStoreOffset, "indexSize=" + indexSize + " size=" + size + " docStoreOffset=" + docStoreOffset);
}
else
{
this.docStoreOffset = 0;
this.size = (int) (indexSize >> 3);
}
indexStream = (IndexInput) cloneableIndexStream.Clone();
numTotalDocs = (int) (indexSize >> 3);
success = true;
}
finally
{
// With lock-less commits, it's entirely possible (and
// fine) to hit a FileNotFound exception above. In
// this case, we want to explicitly close any subset
// of things that were opened so that we don't have to
// wait for a GC to do so.
if (!success)
{
Close();
}
}
}
internal FieldsReader(Directory d, System.String segment, FieldInfos fn, int readBufferSize, int docStoreOffset, int size)
{
bool success = false;
isOriginal = true;
try
{
fieldInfos = fn;
cloneableFieldsStream = d.OpenInput(segment + "." + IndexFileNames.FIELDS_EXTENSION, readBufferSize);
cloneableIndexStream = d.OpenInput(segment + "." + IndexFileNames.FIELDS_INDEX_EXTENSION, readBufferSize);
// First version of fdx did not include a format
// header, but, the first int will always be 0 in that
// case
int firstInt = cloneableIndexStream.ReadInt();
if (firstInt == 0)
{
format = 0;
}
else
{
format = firstInt;
}
if (format > FieldsWriter.FORMAT_CURRENT &&
/* extra support for Lucene 3.0 indexes: */ format != FieldsWriter.FORMAT_LUCENE_3_0_NO_COMPRESSED_FIELDS
)
{
throw new CorruptIndexException("Incompatible format version: " + format + " expected " + FieldsWriter.FORMAT_CURRENT + " or lower");
}
if (format > FieldsWriter.FORMAT)
{
formatSize = 4;
}
else
{
formatSize = 0;
}
if (format < FieldsWriter.FORMAT_VERSION_UTF8_LENGTH_IN_BYTES)
{
cloneableFieldsStream.SetModifiedUTF8StringsMode();
}
fieldsStream = (IndexInput)cloneableFieldsStream.Clone();
long indexSize = cloneableIndexStream.Length() - formatSize;
if (docStoreOffset != -1)
{
// We read only a slice out of this shared fields file
this.docStoreOffset = docStoreOffset;
this.size = size;
// Verify the file is long enough to hold all of our
// docs
System.Diagnostics.Debug.Assert(((int)(indexSize / 8)) >= size + this.docStoreOffset, "indexSize=" + indexSize + " size=" + size + " docStoreOffset=" + docStoreOffset);
}
else
{
this.docStoreOffset = 0;
this.size = (int)(indexSize >> 3);
}
indexStream = (IndexInput)cloneableIndexStream.Clone();
numTotalDocs = (int)(indexSize >> 3);
success = true;
}
finally
{
// With lock-less commits, it's entirely possible (and
// fine) to hit a FileNotFound exception above. In
// this case, we want to explicitly close any subset
// of things that were opened so that we don't have to
// wait for a GC to do so.
if (!success)
{
Close();
}
}
}
internal TermVectorsReader(Directory d, System.String segment, FieldInfos fieldInfos, int readBufferSize, int docStoreOffset, int size)
{
bool success = false;
try
{
if (d.FileExists(segment + "." + IndexFileNames.VECTORS_INDEX_EXTENSION))
{
tvx = d.OpenInput(segment + "." + IndexFileNames.VECTORS_INDEX_EXTENSION, readBufferSize);
format = CheckValidFormat(tvx);
tvd = d.OpenInput(segment + "." + IndexFileNames.VECTORS_DOCUMENTS_EXTENSION, readBufferSize);
int tvdFormat = CheckValidFormat(tvd);
tvf = d.OpenInput(segment + "." + IndexFileNames.VECTORS_FIELDS_EXTENSION, readBufferSize);
int tvfFormat = CheckValidFormat(tvf);
System.Diagnostics.Debug.Assert(format == tvdFormat);
System.Diagnostics.Debug.Assert(format == tvfFormat);
if (format >= FORMAT_VERSION2)
{
System.Diagnostics.Debug.Assert((tvx.Length() - FORMAT_SIZE) % 16 == 0);
numTotalDocs = (int)(tvx.Length() >> 4);
}
else
{
System.Diagnostics.Debug.Assert((tvx.Length() - FORMAT_SIZE) % 8 == 0);
numTotalDocs = (int)(tvx.Length() >> 3);
}
if (-1 == docStoreOffset)
{
this.docStoreOffset = 0;
this.size = numTotalDocs;
System.Diagnostics.Debug.Assert(size == 0 || numTotalDocs == size);
}
else
{
this.docStoreOffset = docStoreOffset;
this.size = size;
// Verify the file is long enough to hold all of our
// docs
System.Diagnostics.Debug.Assert(numTotalDocs >= size + docStoreOffset, "numTotalDocs=" + numTotalDocs + " size=" + size + " docStoreOffset=" + docStoreOffset);
}
}
else
{
// If all documents flushed in a segment had hit
// non-aborting exceptions, it's possible that
// FieldInfos.hasVectors returns true yet the term
// vector files don't exist.
format = 0;
}
this.fieldInfos = fieldInfos;
success = true;
}
finally
{
// With lock-less commits, it's entirely possible (and
// fine) to hit a FileNotFound exception above. In
// this case, we want to explicitly close any subset
// of things that were opened so that we don't have to
// wait for a GC to do so.
if (!success)
{
Close();
}
}
}
public CompoundFileReader(Directory dir, System.String name, int readBufferSize)
{
directory = dir;
fileName = name;
this.readBufferSize = readBufferSize;
bool success = false;
try
{
stream = dir.OpenInput(name, readBufferSize);
// read the directory and init files
int count = stream.ReadVInt();
FileEntry entry = null;
for (int i = 0; i < count; i++)
{
long offset = stream.ReadLong();
System.String id = stream.ReadString();
if (entry != null)
{
// set length of the previous entry
entry.length = offset - entry.offset;
}
entry = new FileEntry();
entry.offset = offset;
entries[id] = entry;
}
// set the length of the final entry
if (entry != null)
{
entry.length = stream.Length() - entry.offset;
}
success = true;
}
finally
{
if (!success && (stream != null))
{
try
{
stream.Close();
}
catch (System.IO.IOException e)
{
}
}
}
}
