protected internal virtual IndexWriterConfig GetConfig(Random random, IndexDeletionPolicy dp)
{
IndexWriterConfig conf = NewIndexWriterConfig(TEST_VERSION_CURRENT, new MockAnalyzer(random));
if (dp != null)
{
conf.SetIndexDeletionPolicy(dp);
}
return conf;
}
public RavenIndexWriter(Directory d, Analyzer a, IndexDeletionPolicy deletionPolicy, IndexWriter.MaxFieldLength mfl, int maximumNumberOfWritesBeforeRecreate, IndexWriter.IndexReaderWarmer indexReaderWarmer)
{
directory = d;
analyzer = a;
indexDeletionPolicy = deletionPolicy;
maxFieldLength = mfl;
_indexReaderWarmer = indexReaderWarmer;
this.maximumNumberOfWritesBeforeRecreate = maximumNumberOfWritesBeforeRecreate;
RecreateIfNecessary();
}
public RavenIndexWriter(Directory d, Analyzer a, IndexDeletionPolicy deletionPolicy, IndexWriter.MaxFieldLength mfl, int maximumNumberOfWritesBeforeRecreate, IndexWriter.IndexReaderWarmer indexReaderWarmer)
{
directory = d;
analyzer = a;
indexDeletionPolicy = deletionPolicy;
maxFieldLength = mfl;
_indexReaderWarmer = indexReaderWarmer;
this.maximumNumberOfWritesBeforeRecreate = maximumNumberOfWritesBeforeRecreate;
RecreateIfNecessary();
}
public LuceneIndexWriter(Directory d, Analyzer a, IndexDeletionPolicy deletionPolicy,
IndexWriter.MaxFieldLength mfl, IndexWriter.IndexReaderWarmer indexReaderWarmer, DocumentDatabase documentDatabase, IState state)
{
_directory = d;
_analyzer = a;
_indexDeletionPolicy = deletionPolicy;
_maxFieldLength = mfl;
_indexReaderWarmer = indexReaderWarmer;
_logger = LoggingSource.Instance.GetLogger <LuceneIndexWriter>(documentDatabase.Name);
RecreateIndexWriter(state);
}
/// <summary>Construct reading the named set of readers. </summary>
internal DirectoryReader(Directory directory, SegmentInfos sis, IndexDeletionPolicy deletionPolicy, bool readOnly, int termInfosIndexDivisor)
{
this.directory = directory;
this.readOnly = readOnly;
this.segmentInfos = sis;
this.deletionPolicy = deletionPolicy;
this.termInfosIndexDivisor = termInfosIndexDivisor;
if (!readOnly)
{
// We assume that this segments_N was previously
// properly sync'd:
SupportClass.CollectionsHelper.AddAllIfNotContains(synced, sis.Files(directory, true));
}
// To reduce the chance of hitting FileNotFound
// (and having to retry), we open segments in
// reverse because IndexWriter merges & deletes
// the newest segments first.
SegmentReader[] readers = new SegmentReader[sis.Count];
for (int i = sis.Count - 1; i >= 0; i--)
{
bool success = false;
try
{
readers[i] = SegmentReader.Get(readOnly, sis.Info(i), termInfosIndexDivisor);
success = true;
}
finally
{
if (!success)
{
// Close all readers we had opened:
for (i++; i < sis.Count; i++)
{
try
{
readers[i].Close();
}
catch (System.Exception ignore)
{
// keep going - we want to clean up as much as possible
}
}
}
}
}
Initialize(readers);
}
public RavenIndexWriter(Directory d, Analyzer a, IndexDeletionPolicy deletionPolicy, IndexWriter.MaxFieldLength mfl, int maximumNumberOfWritesBeforeRecreate, IndexWriter.IndexReaderWarmer indexReaderWarmer)
{
directory = d;
analyzer = a;
indexDeletionPolicy = deletionPolicy;
maxFieldLength = mfl;
_indexReaderWarmer = indexReaderWarmer;
this.maximumNumberOfWritesBeforeRecreate = maximumNumberOfWritesBeforeRecreate;
forceCommitDoc = new Document();
forceCommitDoc.Add(forceCommitField);
RecreateIfNecessary(force: true);
}
public RavenIndexWriter(Directory d, Analyzer a, IndexDeletionPolicy deletionPolicy, IndexWriter.MaxFieldLength mfl, int maximumNumberOfWritesBeforeRecreate, IndexWriter.IndexReaderWarmer indexReaderWarmer)
{
directory = d;
analyzer = a;
indexDeletionPolicy = deletionPolicy;
maxFieldLength = mfl;
_indexReaderWarmer = indexReaderWarmer;
this.maximumNumberOfWritesBeforeRecreate = maximumNumberOfWritesBeforeRecreate;
forceCommitDoc = new Document();
forceCommitDoc.Add(forceCommitField);
RecreateIfNecessary(force: true);
}
/// <summary>
/// Create index snapshot iterator for a writable index. </summary>
/// <param name="indexFolder"> index location folder </param>
/// <param name="indexWriter"> index writer </param>
/// <returns> index file name iterator </returns>
/// <exception cref="IOException"> </exception>
//JAVA TO C# CONVERTER WARNING: Method 'throws' clauses are not available in C#:
//ORIGINAL LINE: public static org.neo4j.graphdb.ResourceIterator<java.io.File> forIndex(java.io.File indexFolder, org.apache.lucene.index.IndexWriter indexWriter) throws java.io.IOException
public static ResourceIterator <File> ForIndex(File indexFolder, IndexWriter indexWriter)
{
IndexDeletionPolicy deletionPolicy = indexWriter.Config.IndexDeletionPolicy;
if (deletionPolicy is SnapshotDeletionPolicy)
{
SnapshotDeletionPolicy policy = ( SnapshotDeletionPolicy )deletionPolicy;
return(HasCommits(indexWriter) ? new WritableIndexSnapshotFileIterator(indexFolder, policy) : emptyResourceIterator());
}
else
{
//JAVA TO C# CONVERTER WARNING: The .NET Type.FullName property will not always yield results identical to the Java Class.getName method:
throw new UnsupportedIndexDeletionPolicy("Can't perform index snapshot with specified index deletion " + "policy: " + deletionPolicy.GetType().FullName + ". " + "Only " + typeof(SnapshotDeletionPolicy).FullName + " is " + "supported");
}
}
public IndexWriter(Directory d, bool autoCommit, Analyzer a, bool create, IndexDeletionPolicy deletionPolicy)
{
InitBlock();
Init(d, a, create, false, deletionPolicy, autoCommit, DEFAULT_MAX_FIELD_LENGTH, null, null);
}
/// <summary>Expert: returns an IndexReader reading the index in
/// the given Directory, using a specific commit and with
/// a custom {@link IndexDeletionPolicy}. You should pass
/// readOnly=true, since it gives much better concurrent
/// performance, unless you intend to do write operations
/// (delete documents or change norms) with the reader.
/// </summary>
/// <param name="commit">the specific {@link IndexCommit} to open;
/// see {@link IndexReader#listCommits} to list all commits
/// in a directory
/// </param>
/// <param name="deletionPolicy">a custom deletion policy (only used
/// if you use this reader to perform deletes or to set
/// norms); see {@link IndexWriter} for details.
/// </param>
/// <param name="readOnly">true if no changes (deletions, norms) will be made with this IndexReader
/// </param>
/// <param name="termInfosIndexDivisor">Subsambles which indexed
/// terms are loaded into RAM. This has the same effect as {@link
/// IndexWriter#setTermIndexInterval} except that setting
/// must be done at indexing time while this setting can be
/// set per reader. When set to N, then one in every
/// N*termIndexInterval terms in the index is loaded into
/// memory. By setting this to a value > 1 you can reduce
/// memory usage, at the expense of higher latency when
/// loading a TermInfo. The default value is 1. Set this
/// to -1 to skip loading the terms index entirely.
/// </param>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> IOException if there is a low-level IO error </throws>
public static IndexReader Open(IndexCommit commit, IndexDeletionPolicy deletionPolicy, bool readOnly, int termInfosIndexDivisor)
{
return Open(commit.GetDirectory(), deletionPolicy, commit, readOnly, termInfosIndexDivisor);
}
public TimeTrackingIndexWriter(Directory d, Analyzer a, IndexDeletionPolicy deletionPolicy, MaxFieldLength mfl, IState state) : base(d, a, deletionPolicy, mfl, state)
{
}
/// <summary> Initialize the deleter: find all previous commits in
/// the Directory, incref the files they reference, call
/// the policy to let it delete commits. This will remove
/// any files not referenced by any of the commits.
/// </summary>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> IOException if there is a low-level IO error </throws>
public IndexFileDeleter(Directory directory, IndexDeletionPolicy policy, SegmentInfos segmentInfos, System.IO.StreamWriter infoStream, DocumentsWriter docWriter, HashSet <string> synced)
{
this.docWriter = docWriter;
this.infoStream = infoStream;
this.synced = synced;
if (infoStream != null)
{
Message("init: current segments file is \"" + segmentInfos.GetCurrentSegmentFileName() + "\"; deletionPolicy=" + policy);
}
this.policy = policy;
this.directory = directory;
// First pass: walk the files and initialize our ref
// counts:
long currentGen = segmentInfos.Generation;
IndexFileNameFilter filter = IndexFileNameFilter.Filter;
System.String[] files = directory.ListAll();
CommitPoint currentCommitPoint = null;
for (int i = 0; i < files.Length; i++)
{
System.String fileName = files[i];
if (filter.Accept(null, fileName) && !fileName.Equals(IndexFileNames.SEGMENTS_GEN))
{
// Add this file to refCounts with initial count 0:
GetRefCount(fileName);
if (fileName.StartsWith(IndexFileNames.SEGMENTS))
{
// This is a commit (segments or segments_N), and
// it's valid (<= the max gen). Load it, then
// incref all files it refers to:
if (infoStream != null)
{
Message("init: load commit \"" + fileName + "\"");
}
SegmentInfos sis = new SegmentInfos();
try
{
sis.Read(directory, fileName);
}
catch (System.IO.FileNotFoundException)
{
// LUCENE-948: on NFS (and maybe others), if
// you have writers switching back and forth
// between machines, it's very likely that the
// dir listing will be stale and will claim a
// file segments_X exists when in fact it
// doesn't. So, we catch this and handle it
// as if the file does not exist
if (infoStream != null)
{
Message("init: hit FileNotFoundException when loading commit \"" + fileName + "\"; skipping this commit point");
}
sis = null;
}
catch (System.IO.IOException)
{
if (SegmentInfos.GenerationFromSegmentsFileName(fileName) <= currentGen)
{
throw;
}
else
{
// Most likely we are opening an index that
// has an aborted "future" commit, so suppress
// exc in this case
sis = null;
}
}
if (sis != null)
{
CommitPoint commitPoint = new CommitPoint(this, commitsToDelete, directory, sis);
if (sis.Generation == segmentInfos.Generation)
{
currentCommitPoint = commitPoint;
}
commits.Add(commitPoint);
IncRef(sis, true);
if (lastSegmentInfos == null || sis.Generation > lastSegmentInfos.Generation)
{
lastSegmentInfos = sis;
}
}
}
}
}
if (currentCommitPoint == null)
{
// We did not in fact see the segments_N file
// corresponding to the segmentInfos that was passed
// in. Yet, it must exist, because our caller holds
// the write lock. This can happen when the directory
// listing was stale (eg when index accessed via NFS
// client with stale directory listing cache). So we
// try now to explicitly open this commit point:
SegmentInfos sis = new SegmentInfos();
try
{
sis.Read(directory, segmentInfos.GetCurrentSegmentFileName());
}
catch (System.IO.IOException)
{
throw new CorruptIndexException("failed to locate current segments_N file");
}
if (infoStream != null)
{
Message("forced open of current segments file " + segmentInfos.GetCurrentSegmentFileName());
}
currentCommitPoint = new CommitPoint(this, commitsToDelete, directory, sis);
commits.Add(currentCommitPoint);
IncRef(sis, true);
}
// We keep commits list in sorted order (oldest to newest):
commits.Sort();
// Now delete anything with ref count at 0. These are
// presumably abandoned files eg due to crash of
// IndexWriter.
foreach (KeyValuePair <string, RefCount> entry in refCounts)
{
string fileName = entry.Key;
RefCount rc = refCounts[fileName];
if (0 == rc.count)
{
if (infoStream != null)
{
Message("init: removing unreferenced file \"" + fileName + "\"");
}
DeleteFile(fileName);
}
}
// Finally, give policy a chance to remove things on
// startup:
policy.OnInit(commits);
// Always protect the incoming segmentInfos since
// sometime it may not be the most recent commit
Checkpoint(segmentInfos, false);
startingCommitDeleted = currentCommitPoint.IsDeleted;
DeleteCommits();
}
/// <summary>Expert: returns an IndexReader reading the index in
/// the given Directory, with a custom {@link
/// IndexDeletionPolicy}. You should pass readOnly=true,
/// since it gives much better concurrent performance,
/// unless you intend to do write operations (delete
/// documents or change norms) with the reader.
/// </summary>
/// <param name="directory">the index directory
/// </param>
/// <param name="deletionPolicy">a custom deletion policy (only used
/// if you use this reader to perform deletes or to set
/// norms); see {@link IndexWriter} for details.
/// </param>
/// <param name="readOnly">true if no changes (deletions, norms) will be made with this IndexReader
/// </param>
/// <param name="termInfosIndexDivisor">Subsamples which indexed
/// terms are loaded into RAM. This has the same effect as {@link
/// IndexWriter#setTermIndexInterval} except that setting
/// must be done at indexing time while this setting can be
/// set per reader. When set to N, then one in every
/// N*termIndexInterval terms in the index is loaded into
/// memory. By setting this to a value > 1 you can reduce
/// memory usage, at the expense of higher latency when
/// loading a TermInfo. The default value is 1. Set this
/// to -1 to skip loading the terms index entirely.
/// </param>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> IOException if there is a low-level IO error </throws>
public static IndexReader Open(Directory directory, IndexDeletionPolicy deletionPolicy, bool readOnly, int termInfosIndexDivisor)
{
return Open(directory, deletionPolicy, null, readOnly, termInfosIndexDivisor);
}
private void Init(Directory d, Analyzer a, IndexDeletionPolicy deletionPolicy, int maxFieldLength, DocumentsWriter.IndexingChain indexingChain, IndexCommit commit)
{
if (IndexReader.IndexExists(d))
{
Init(d, a, false, deletionPolicy, maxFieldLength, indexingChain, commit);
}
else
{
Init(d, a, true, deletionPolicy, maxFieldLength, indexingChain, commit);
}
}
/// <summary>
/// <seealso cref="PersistentSnapshotDeletionPolicy"/> wraps another
/// <seealso cref="IndexDeletionPolicy"/> to enable flexible
/// snapshotting, passing <seealso cref="OpenMode#CREATE_OR_APPEND"/>
/// by default.
/// </summary>
/// <param name="primary">
/// the <seealso cref="IndexDeletionPolicy"/> that is used on non-snapshotted
/// commits. Snapshotted commits, by definition, are not deleted until
/// explicitly released via <seealso cref="#release"/>. </param>
/// <param name="dir">
/// the <seealso cref="Directory"/> which will be used to persist the snapshots
/// information. </param>
public PersistentSnapshotDeletionPolicy(IndexDeletionPolicy primary, Directory dir)
: this(primary, dir, OpenMode.CREATE_OR_APPEND)
{
}
/// <summary>
/// Expert: allows an optional <seealso cref="IndexDeletionPolicy"/> implementation to be
/// specified. You can use this to control when prior commits are deleted from
/// the index. The default policy is <seealso cref="KeepOnlyLastCommitDeletionPolicy"/>
/// which removes all prior commits as soon as a new commit is done (this
/// matches behavior before 2.2). Creating your own policy can allow you to
/// explicitly keep previous "point in time" commits alive in the index for
/// some time, to allow readers to refresh to the new commit without having the
/// old commit deleted out from under them. this is necessary on filesystems
/// like NFS that do not support "delete on last close" semantics, which
/// Lucene's "point in time" search normally relies on.
/// <p>
/// <b>NOTE:</b> the deletion policy cannot be null.
///
/// <p>Only takes effect when IndexWriter is first created.
/// </summary>
public IndexWriterConfig SetIndexDeletionPolicy(IndexDeletionPolicy deletionPolicy)
{
if (deletionPolicy == null)
{
throw new System.ArgumentException("indexDeletionPolicy must not be null");
}
this.delPolicy = deletionPolicy;
return this;
}
/// <summary> Expert: constructs an IndexWriter with a custom <see cref="IndexDeletionPolicy" />
/// and <see cref="DocumentsWriter.IndexingChain" />,
/// for the index in <c>d</c>.
/// Text will be analyzed with <c>a</c>. If
/// <c>create</c> is true, then a new, empty index
/// will be created in <c>d</c>, replacing the index
/// already there, if any.
///
/// </summary>
/// <param name="d">the index directory
/// </param>
/// <param name="a">the analyzer to use
/// </param>
/// <param name="create"><c>true</c> to create the index or overwrite
/// the existing one; <c>false</c> to append to the existing
/// index
/// </param>
/// <param name="deletionPolicy">see <a href="#deletionPolicy">above</a>
/// </param>
/// <param name="mfl">whether or not to limit field lengths, value is in number of terms/tokens. See <see cref="Lucene.Net.Index.IndexWriter.MaxFieldLength" />.
/// </param>
/// <param name="indexingChain">the <see cref="DocConsumer" /> chain to be used to
/// process documents
/// </param>
/// <param name="commit">which commit to open
/// </param>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> LockObtainFailedException if another writer </throws>
/// <summary> has this index open (<c>write.lock</c> could not
/// be obtained)
/// </summary>
/// <throws> IOException if the directory cannot be read/written to, or </throws>
/// <summary> if it does not exist and <c>create</c> is
/// <c>false</c> or if there is any other low-level
/// IO error
/// </summary>
internal IndexWriter(Directory d, Analyzer a, bool create, IndexDeletionPolicy deletionPolicy, MaxFieldLength mfl, DocumentsWriter.IndexingChain indexingChain, IndexCommit commit)
{
InitBlock();
Init(d, a, create, deletionPolicy, mfl.Limit, indexingChain, commit);
}
internal static IndexReader Open(Directory directory, IndexDeletionPolicy deletionPolicy, IndexCommit commit, bool readOnly, int termInfosIndexDivisor)
{
return (IndexReader) new AnonymousClassFindSegmentsFile(readOnly, deletionPolicy, termInfosIndexDivisor, directory).Run(commit);
}
internal ReadOnlyDirectoryReader(Directory directory, SegmentInfos sis, IndexDeletionPolicy deletionPolicy, int termInfosIndexDivisor) : base(directory, sis, deletionPolicy, true, termInfosIndexDivisor)
{
}
/// <summary> Expert: constructs an IndexWriter with a custom {@link
/// IndexDeletionPolicy} and {@link IndexingChain},
/// for the index in <code>d</code>.
/// Text will be analyzed with <code>a</code>. If
/// <code>create</code> is true, then a new, empty index
/// will be created in <code>d</code>, replacing the index
/// already there, if any.
///
/// <p/><b>NOTE</b>: autoCommit (see <a
/// href="#autoCommit">above</a>) is set to false with this
/// constructor.
///
/// </summary>
/// <param name="d">the index directory
/// </param>
/// <param name="a">the analyzer to use
/// </param>
/// <param name="create"><code>true</code> to create the index or overwrite
/// the existing one; <code>false</code> to append to the existing
/// index
/// </param>
/// <param name="deletionPolicy">see <a href="#deletionPolicy">above</a>
/// </param>
/// <param name="mfl">whether or not to limit field lengths, value is in number of terms/tokens. See {@link Lucene.Net.Index.IndexWriter.MaxFieldLength}.
/// </param>
/// <param name="indexingChain">the {@link DocConsumer} chain to be used to
/// process documents
/// </param>
/// <param name="commit">which commit to open
/// </param>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> LockObtainFailedException if another writer </throws>
/// <summary> has this index open (<code>write.lock</code> could not
/// be obtained)
/// </summary>
/// <throws> IOException if the directory cannot be read/written to, or </throws>
/// <summary> if it does not exist and <code>create</code> is
/// <code>false</code> or if there is any other low-level
/// IO error
/// </summary>
internal IndexWriter(Directory d, Analyzer a, bool create, IndexDeletionPolicy deletionPolicy, MaxFieldLength mfl, IndexingChain indexingChain, IndexCommit commit)
{
InitBlock();
Init(d, a, create, false, deletionPolicy, false, mfl.GetLimit(), indexingChain, commit);
}
/// <summary> Expert: constructs an IndexWriter with a custom {@link
/// IndexDeletionPolicy}, for the index in <code>d</code>,
/// first creating it if it does not already exist. Text
/// will be analyzed with <code>a</code>.
///
/// <p/><b>NOTE</b>: autoCommit (see <a
/// href="#autoCommit">above</a>) is set to false with this
/// constructor.
///
/// </summary>
/// <param name="d">the index directory
/// </param>
/// <param name="a">the analyzer to use
/// </param>
/// <param name="deletionPolicy">see <a href="#deletionPolicy">above</a>
/// </param>
/// <param name="mfl">whether or not to limit field lengths
/// </param>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> LockObtainFailedException if another writer </throws>
/// <summary> has this index open (<code>write.lock</code> could not
/// be obtained)
/// </summary>
/// <throws> IOException if the directory cannot be </throws>
/// <summary> read/written to or if there is any other low-level
/// IO error
/// </summary>
public IndexWriter(Directory d, Analyzer a, IndexDeletionPolicy deletionPolicy, MaxFieldLength mfl)
{
InitBlock();
Init(d, a, false, deletionPolicy, false, mfl.GetLimit(), null, null);
}
private void InitBlock(bool readOnly, IndexDeletionPolicy deletionPolicy, int termInfosIndexDivisor)
{
this.readOnly = readOnly;
this.deletionPolicy = deletionPolicy;
this.termInfosIndexDivisor = termInfosIndexDivisor;
}
public static IndexWriterConfig CreateWriterConfig(Config config, PerfRunData runData, OpenMode mode, IndexCommit commit)
{
// :Post-Release-Update-Version.LUCENE_XY:
LuceneVersion version = (LuceneVersion)Enum.Parse(typeof(LuceneVersion), config.Get("writer.version", LuceneVersion.LUCENE_48.ToString()));
IndexWriterConfig iwConf = new IndexWriterConfig(version, runData.Analyzer);
iwConf.OpenMode = mode;
IndexDeletionPolicy indexDeletionPolicy = GetIndexDeletionPolicy(config);
iwConf.IndexDeletionPolicy = indexDeletionPolicy;
if (commit != null)
{
iwConf.IndexCommit = commit;
}
string mergeScheduler = config.Get("merge.scheduler",
"Lucene.Net.Index.ConcurrentMergeScheduler, Lucene.Net");
Type mergeSchedulerType = Type.GetType(mergeScheduler);
if (mergeSchedulerType is null)
{
throw RuntimeException.Create("Unrecognized merge scheduler type '" + mergeScheduler + "'"); // LUCENENET: We don't get an exception in this case, so throwing one for compatibility
}
else if (mergeSchedulerType.Equals(typeof(NoMergeScheduler)))
{
iwConf.MergeScheduler = NoMergeScheduler.INSTANCE;
}
else
{
try
{
iwConf.MergeScheduler = (IMergeScheduler)Activator.CreateInstance(mergeSchedulerType);
}
catch (Exception e) when(e.IsException())
{
throw RuntimeException.Create("unable to instantiate class '" + mergeScheduler + "' as merge scheduler", e);
}
if (mergeScheduler.Equals("Lucene.Net.Index.ConcurrentMergeScheduler", StringComparison.Ordinal))
{
ConcurrentMergeScheduler cms = (ConcurrentMergeScheduler)iwConf.MergeScheduler;
int maxThreadCount = config.Get("concurrent.merge.scheduler.max.thread.count", ConcurrentMergeScheduler.DEFAULT_MAX_THREAD_COUNT);
int maxMergeCount = config.Get("concurrent.merge.scheduler.max.merge.count", ConcurrentMergeScheduler.DEFAULT_MAX_MERGE_COUNT);
cms.SetMaxMergesAndThreads(maxMergeCount, maxThreadCount);
}
}
string defaultCodec = config.Get("default.codec", null);
if (defaultCodec != null)
{
try
{
Type clazz = Type.GetType(defaultCodec);
iwConf.Codec = (Codec)Activator.CreateInstance(clazz);
}
catch (Exception e) when(e.IsException())
{
throw RuntimeException.Create("Couldn't instantiate Codec: " + defaultCodec, e);
}
}
string mergePolicy = config.Get("merge.policy",
"Lucene.Net.Index.LogByteSizeMergePolicy, Lucene.Net");
bool isCompound = config.Get("compound", true);
Type mergePolicyType = Type.GetType(mergePolicy);
if (mergePolicyType is null)
{
throw RuntimeException.Create("Unrecognized merge policy type '" + mergePolicy + "'"); // LUCENENET: We don't get an exception in this case, so throwing one for compatibility
}
else if (mergePolicyType.Equals(typeof(NoMergePolicy)))
{
iwConf.MergePolicy = isCompound ? NoMergePolicy.COMPOUND_FILES : NoMergePolicy.NO_COMPOUND_FILES;
}
else
{
try
{
iwConf.MergePolicy = (MergePolicy)Activator.CreateInstance(mergePolicyType);
}
catch (Exception e) when(e.IsException())
{
throw RuntimeException.Create("unable to instantiate class '" + mergePolicy + "' as merge policy", e);
}
iwConf.MergePolicy.NoCFSRatio = isCompound ? 1.0 : 0.0;
if (iwConf.MergePolicy is LogMergePolicy logMergePolicy)
{
logMergePolicy.MergeFactor = config.Get("merge.factor", OpenIndexTask.DEFAULT_MERGE_PFACTOR);
}
}
double ramBuffer = config.Get("ram.flush.mb", OpenIndexTask.DEFAULT_RAM_FLUSH_MB);
int maxBuffered = config.Get("max.buffered", OpenIndexTask.DEFAULT_MAX_BUFFERED);
if (maxBuffered == IndexWriterConfig.DISABLE_AUTO_FLUSH)
{
iwConf.RAMBufferSizeMB = ramBuffer;
iwConf.MaxBufferedDocs = maxBuffered;
}
else
{
iwConf.MaxBufferedDocs = maxBuffered;
iwConf.RAMBufferSizeMB = ramBuffer;
}
return(iwConf);
}
private void Init(Directory d, Analyzer a, bool closeDir, IndexDeletionPolicy deletionPolicy, bool autoCommit, int maxFieldLength, IndexingChain indexingChain, IndexCommit commit)
{
if (IndexReader.IndexExists(d))
{
Init(d, a, false, closeDir, deletionPolicy, autoCommit, maxFieldLength, indexingChain, commit);
}
else
{
Init(d, a, true, closeDir, deletionPolicy, autoCommit, maxFieldLength, indexingChain, commit);
}
}
/// <summary> Initialize the deleter: find all previous commits in
/// the Directory, incref the files they reference, call
/// the policy to let it delete commits. This will remove
/// any files not referenced by any of the commits.
/// </summary>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> IOException if there is a low-level IO error </throws>
public IndexFileDeleter(Directory directory, IndexDeletionPolicy policy, SegmentInfos segmentInfos, System.IO.StreamWriter infoStream, DocumentsWriter docWriter, System.Collections.Generic.Dictionary<string, string> synced)
{
this.docWriter = docWriter;
this.infoStream = infoStream;
this.synced = synced;
if (infoStream != null)
{
Message("init: current segments file is \"" + segmentInfos.GetCurrentSegmentFileName() + "\"; deletionPolicy=" + policy);
}
this.policy = policy;
this.directory = directory;
// First pass: walk the files and initialize our ref
// counts:
long currentGen = segmentInfos.GetGeneration();
IndexFileNameFilter filter = IndexFileNameFilter.GetFilter();
System.String[] files = directory.ListAll();
CommitPoint currentCommitPoint = null;
for (int i = 0; i < files.Length; i++)
{
System.String fileName = files[i];
if (filter.Accept(null, fileName) && !fileName.Equals(IndexFileNames.SEGMENTS_GEN))
{
// Add this file to refCounts with initial count 0:
GetRefCount(fileName);
if (fileName.StartsWith(IndexFileNames.SEGMENTS))
{
// This is a commit (segments or segments_N), and
// it's valid (<= the max gen). Load it, then
// incref all files it refers to:
if (infoStream != null)
{
Message("init: load commit \"" + fileName + "\"");
}
SegmentInfos sis = new SegmentInfos();
try
{
sis.Read(directory, fileName);
}
catch (System.IO.FileNotFoundException e)
{
// LUCENE-948: on NFS (and maybe others), if
// you have writers switching back and forth
// between machines, it's very likely that the
// dir listing will be stale and will claim a
// file segments_X exists when in fact it
// doesn't. So, we catch this and handle it
// as if the file does not exist
if (infoStream != null)
{
Message("init: hit FileNotFoundException when loading commit \"" + fileName + "\"; skipping this commit point");
}
sis = null;
}
catch (System.IO.IOException e)
{
if (SegmentInfos.GenerationFromSegmentsFileName(fileName) <= currentGen)
{
throw e;
}
else
{
// Most likely we are opening an index that
// has an aborted "future" commit, so suppress
// exc in this case
sis = null;
}
}
if (sis != null)
{
CommitPoint commitPoint = new CommitPoint(this,commitsToDelete, directory, sis);
if (sis.GetGeneration() == segmentInfos.GetGeneration())
{
currentCommitPoint = commitPoint;
}
commits.Add(commitPoint);
IncRef(sis, true);
if (lastSegmentInfos == null || sis.GetGeneration() > lastSegmentInfos.GetGeneration())
{
lastSegmentInfos = sis;
}
}
}
}
}
if (currentCommitPoint == null)
{
// We did not in fact see the segments_N file
// corresponding to the segmentInfos that was passed
// in. Yet, it must exist, because our caller holds
// the write lock. This can happen when the directory
// listing was stale (eg when index accessed via NFS
// client with stale directory listing cache). So we
// try now to explicitly open this commit point:
SegmentInfos sis = new SegmentInfos();
try
{
sis.Read(directory, segmentInfos.GetCurrentSegmentFileName());
}
catch (System.IO.IOException e)
{
throw new CorruptIndexException("failed to locate current segments_N file");
}
if (infoStream != null)
Message("forced open of current segments file " + segmentInfos.GetCurrentSegmentFileName());
currentCommitPoint = new CommitPoint(this, commitsToDelete, directory, sis);
commits.Add(currentCommitPoint);
IncRef(sis, true);
}
// We keep commits list in sorted order (oldest to newest):
commits.Sort();
// Now delete anything with ref count at 0. These are
// presumably abandoned files eg due to crash of
// IndexWriter.
System.Collections.Generic.IEnumerator<System.Collections.Generic.KeyValuePair<System.String, RefCount>> it = refCounts.GetEnumerator();
while (it.MoveNext())
{
System.String fileName = (System.String) it.Current.Key;
RefCount rc = (RefCount) refCounts[fileName];
if (0 == rc.count)
{
if (infoStream != null)
{
Message("init: removing unreferenced file \"" + fileName + "\"");
}
DeleteFile(fileName);
}
}
// Finally, give policy a chance to remove things on
// startup:
policy.OnInit(commits);
// Always protect the incoming segmentInfos since
// sometime it may not be the most recent commit
Checkpoint(segmentInfos, false);
startingCommitDeleted = currentCommitPoint.IsDeleted();
DeleteCommits();
}
public SnapshotDeletionPolicy(IndexDeletionPolicy primary)
{
this.primary = primary;
}
/// <summary> Expert: constructs an IndexWriter with a custom <see cref="IndexDeletionPolicy" />
///, for the index in <c>d</c>.
/// Text will be analyzed with <c>a</c>. If
/// <c>create</c> is true, then a new, empty index
/// will be created in <c>d</c>, replacing the index
/// already there, if any.
///
/// </summary>
/// <param name="d">the index directory
/// </param>
/// <param name="a">the analyzer to use
/// </param>
/// <param name="create"><c>true</c> to create the index or overwrite
/// the existing one; <c>false</c> to append to the existing
/// index
/// </param>
/// <param name="deletionPolicy">see <a href="#deletionPolicy">above</a>
/// </param>
/// <param name="mfl"><see cref="Lucene.Net.Index.IndexWriter.MaxFieldLength" />, whether or not to limit field lengths. Value is in number of terms/tokens
/// </param>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> LockObtainFailedException if another writer </throws>
/// <summary> has this index open (<c>write.lock</c> could not
/// be obtained)
/// </summary>
/// <throws> IOException if the directory cannot be read/written to, or </throws>
/// <summary> if it does not exist and <c>create</c> is
/// <c>false</c> or if there is any other low-level
/// IO error
/// </summary>
public IndexWriter(Directory d, Analyzer a, bool create, IndexDeletionPolicy deletionPolicy, MaxFieldLength mfl)
{
InitBlock();
Init(d, a, create, deletionPolicy, mfl.Limit, null, null);
}
public SnapshotDeletionPolicy(IndexDeletionPolicy primary)
{
this.primary = primary;
}
/// <summary> Expert: constructs an IndexWriter on specific commit
/// point, with a custom <see cref="IndexDeletionPolicy" />, for
/// the index in <c>d</c>. Text will be analyzed
/// with <c>a</c>.
///
/// <p/> This is only meaningful if you've used a <see cref="IndexDeletionPolicy" />
/// in that past that keeps more than
/// just the last commit.
///
/// <p/>This operation is similar to <see cref="Rollback()" />,
/// except that method can only rollback what's been done
/// with the current instance of IndexWriter since its last
/// commit, whereas this method can rollback to an
/// arbitrary commit point from the past, assuming the
/// <see cref="IndexDeletionPolicy" /> has preserved past
/// commits.
///
/// </summary>
/// <param name="d">the index directory
/// </param>
/// <param name="a">the analyzer to use
/// </param>
/// <param name="deletionPolicy">see <a href="#deletionPolicy">above</a>
/// </param>
/// <param name="mfl">whether or not to limit field lengths, value is in number of terms/tokens. See <see cref="Lucene.Net.Index.IndexWriter.MaxFieldLength" />.
/// </param>
/// <param name="commit">which commit to open
/// </param>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> LockObtainFailedException if another writer </throws>
/// <summary> has this index open (<c>write.lock</c> could not
/// be obtained)
/// </summary>
/// <throws> IOException if the directory cannot be read/written to, or </throws>
/// <summary> if it does not exist and <c>create</c> is
/// <c>false</c> or if there is any other low-level
/// IO error
/// </summary>
public IndexWriter(Directory d, Analyzer a, IndexDeletionPolicy deletionPolicy, MaxFieldLength mfl, IndexCommit commit)
{
InitBlock();
Init(d, a, false, deletionPolicy, mfl.Limit, null, commit);
}
public virtual void SetDeletionPolicy(IndexDeletionPolicy deletionPolicy)
{
this.deletionPolicy = deletionPolicy;
}
private void Init(Directory d, Analyzer a, bool create, IndexDeletionPolicy deletionPolicy, int maxFieldLength, DocumentsWriter.IndexingChain indexingChain, IndexCommit commit)
{
directory = d;
analyzer = a;
SetMessageID(defaultInfoStream);
this.maxFieldLength = maxFieldLength;
if (indexingChain == null)
indexingChain = DocumentsWriter.DefaultIndexingChain;
if (create)
{
// Clear the write lock in case it's leftover:
directory.ClearLock(WRITE_LOCK_NAME);
}
Lock writeLock = directory.MakeLock(WRITE_LOCK_NAME);
if (!writeLock.Obtain(writeLockTimeout))
// obtain write lock
{
throw new LockObtainFailedException("Index locked for write: " + writeLock);
}
this.writeLock = writeLock; // save it
bool success = false;
try
{
if (create)
{
// Try to read first. This is to allow create
// against an index that's currently open for
// searching. In this case we write the next
// segments_N file with no segments:
bool doCommit;
try
{
segmentInfos.Read(directory);
segmentInfos.Clear();
doCommit = false;
}
catch (System.IO.IOException)
{
// Likely this means it's a fresh directory
doCommit = true;
}
if (doCommit)
{
// Only commit if there is no segments file
// in this dir already.
segmentInfos.Commit(directory);
synced.UnionWith(segmentInfos.Files(directory, true));
}
else
{
// Record that we have a change (zero out all
// segments) pending:
changeCount++;
}
}
else
{
segmentInfos.Read(directory);
if (commit != null)
{
// Swap out all segments, but, keep metadata in
// SegmentInfos, like version & generation, to
// preserve write-once. This is important if
// readers are open against the future commit
// points.
if (commit.Directory != directory)
throw new System.ArgumentException("IndexCommit's directory doesn't match my directory");
SegmentInfos oldInfos = new SegmentInfos();
oldInfos.Read(directory, commit.SegmentsFileName);
segmentInfos.Replace(oldInfos);
changeCount++;
if (infoStream != null)
Message("init: loaded commit \"" + commit.SegmentsFileName + "\"");
}
// We assume that this segments_N was previously
// properly sync'd:
synced.UnionWith(segmentInfos.Files(directory, true));
}
SetRollbackSegmentInfos(segmentInfos);
docWriter = new DocumentsWriter(directory, this, indexingChain);
docWriter.SetInfoStream(infoStream);
docWriter.SetMaxFieldLength(maxFieldLength);
// Default deleter (for backwards compatibility) is
// KeepOnlyLastCommitDeleter:
deleter = new IndexFileDeleter(directory, deletionPolicy == null?new KeepOnlyLastCommitDeletionPolicy():deletionPolicy, segmentInfos, infoStream, docWriter, synced);
if (deleter.startingCommitDeleted)
// Deletion policy deleted the "head" commit point.
// We have to mark ourself as changed so that if we
// are closed w/o any further changes we write a new
// segments_N file.
changeCount++;
PushMaxBufferedDocs();
if (infoStream != null)
{
Message("init: create=" + create);
MessageState();
}
success = true;
}
finally
{
if (!success)
{
if (infoStream != null)
{
Message("init: hit exception on init; releasing write lock");
}
try
{
writeLock.Release();
}
catch (Exception)
{
// don't mask the original exception
}
writeLock = null;
}
}
}
internal static DirectoryIndexReader Open(Directory directory, bool closeDirectory, IndexDeletionPolicy deletionPolicy)
{
return Open(directory, closeDirectory, deletionPolicy, null, false);
}
/// <summary>
/// Builder method for <see cref="IndexWriterConfig.IndexDeletionPolicy"/>.
/// </summary>
/// <param name="config">this <see cref="IndexWriterConfig"/> instance</param>
/// <param name="deletionPolicy"></param>
/// <returns>this <see cref="IndexWriterConfig"/> instance</returns>
public static IndexWriterConfig SetIndexDeletionPolicy(this IndexWriterConfig config, IndexDeletionPolicy deletionPolicy)
{
config.IndexDeletionPolicy = deletionPolicy;
return(config);
}
internal static DirectoryIndexReader Open(Directory directory, bool closeDirectory, IndexDeletionPolicy deletionPolicy, IndexCommit commit, bool readOnly)
{
SegmentInfos.FindSegmentsFile finder = new AnonymousClassFindSegmentsFile(closeDirectory, deletionPolicy, directory, readOnly);
if (commit == null)
return (DirectoryIndexReader) finder.Run();
else
{
if (directory != commit.GetDirectory())
throw new System.IO.IOException("the specified commit does not match the specified Directory");
// this can and will directly throw IOException if the specified commit point has been deleted
return (DirectoryIndexReader)finder.DoBody(commit.GetSegmentsFileName());
}
}
/// <summary>Expert: returns an IndexReader reading the index in
/// the given Directory, with a custom {@link
/// IndexDeletionPolicy}. You should pass readOnly=true,
/// since it gives much better concurrent performance,
/// unless you intend to do write operations (delete
/// documents or change norms) with the reader.
/// </summary>
/// <param name="directory">the index directory
/// </param>
/// <param name="deletionPolicy">a custom deletion policy (only used
/// if you use this reader to perform deletes or to set
/// norms); see {@link IndexWriter} for details.
/// </param>
/// <param name="readOnly">true if no changes (deletions, norms) will be made with this IndexReader
/// </param>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> IOException if there is a low-level IO error </throws>
public static IndexReader Open(Directory directory, IndexDeletionPolicy deletionPolicy, bool readOnly)
{
return Open(directory, deletionPolicy, null, readOnly, DEFAULT_TERMS_INDEX_DIVISOR);
}
internal ReadOnlyDirectoryReader(Directory directory, SegmentInfos sis, IndexDeletionPolicy deletionPolicy, int termInfosIndexDivisor)
: base(directory, sis, deletionPolicy, true, termInfosIndexDivisor)
{
}
/// <summary>Expert: returns an IndexReader reading the index in
/// the given Directory, using a specific commit and with
/// a custom {@link IndexDeletionPolicy}. You should pass
/// readOnly=true, since it gives much better concurrent
/// performance, unless you intend to do write operations
/// (delete documents or change norms) with the reader.
/// </summary>
/// <param name="commit">the specific {@link IndexCommit} to open;
/// see {@link IndexReader#listCommits} to list all commits
/// in a directory
/// </param>
/// <param name="deletionPolicy">a custom deletion policy (only used
/// if you use this reader to perform deletes or to set
/// norms); see {@link IndexWriter} for details.
/// </param>
/// <param name="readOnly">true if no changes (deletions, norms) will be made with this IndexReader
/// </param>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> IOException if there is a low-level IO error </throws>
public static IndexReader Open(IndexCommit commit, IndexDeletionPolicy deletionPolicy, bool readOnly)
{
return Open(commit.GetDirectory(), deletionPolicy, commit, readOnly, DEFAULT_TERMS_INDEX_DIVISOR);
}
/// <summary>Expert: returns an IndexReader reading the index in the given
/// Directory, with a custom {@link IndexDeletionPolicy}.
/// </summary>
/// <param name="directory">the index directory
/// </param>
/// <param name="deletionPolicy">a custom deletion policy (only used
/// if you use this reader to perform deletes or to set
/// norms); see {@link IndexWriter} for details.
/// </param>
/// <throws> CorruptIndexException if the index is corrupt </throws>
/// <throws> IOException if there is a low-level IO error </throws>
public static IndexReader Open(Directory directory, IndexDeletionPolicy deletionPolicy)
{
return Open(directory, false, deletionPolicy);
}
private static IndexReader Open(Directory directory, IndexDeletionPolicy deletionPolicy, IndexCommit commit, bool readOnly, int termInfosIndexDivisor)
{
return DirectoryReader.Open(directory, deletionPolicy, commit, readOnly, termInfosIndexDivisor);
}
private static IndexReader Open(Directory directory, bool closeDirectory, IndexDeletionPolicy deletionPolicy)
{
return DirectoryIndexReader.Open(directory, closeDirectory, deletionPolicy);
}
