Java Code Examples for org.apache.lucene.codecs.PostingsFormat#forName()

The following examples show how to use org.apache.lucene.codecs.PostingsFormat#forName() . You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may check out the related API usage on the sidebar.
Example 1
Source File: MtasCodec.java    From mtas with Apache License 2.0 6 votes vote down vote up
@Override
public PostingsFormat postingsFormat() {
  initDelegate();
  if (delegate.postingsFormat() instanceof PerFieldPostingsFormat) {
    Codec defaultCodec = Codec.getDefault();
    PostingsFormat defaultPostingsFormat = defaultCodec.postingsFormat();
    if (defaultPostingsFormat instanceof PerFieldPostingsFormat) {
      defaultPostingsFormat = ((PerFieldPostingsFormat) defaultPostingsFormat)
          .getPostingsFormatForField(null);
      if ((defaultPostingsFormat == null)
          || (defaultPostingsFormat instanceof PerFieldPostingsFormat)) {
        // fallback option
        return new MtasCodecPostingsFormat(
            PostingsFormat.forName("Lucene70"));
      } else {
        return new MtasCodecPostingsFormat(defaultPostingsFormat);
      }
    } else {
      return new MtasCodecPostingsFormat(defaultPostingsFormat);
    }
  } else {
    return new MtasCodecPostingsFormat(delegate.postingsFormat());
  }
}
 
Example 2
Source File: Completion090PostingsFormat.java    From Elasticsearch with Apache License 2.0 5 votes vote down vote up
public CompletionFieldsProducer(SegmentReadState state) throws IOException {
    String suggestFSTFile = IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, EXTENSION);
    IndexInput input = state.directory.openInput(suggestFSTFile, state.context);
    version = CodecUtil.checkHeader(input, CODEC_NAME, SUGGEST_CODEC_VERSION, SUGGEST_VERSION_CURRENT);
    FieldsProducer delegateProducer = null;
    boolean success = false;
    try {
        PostingsFormat delegatePostingsFormat = PostingsFormat.forName(input.readString());
        String providerName = input.readString();
        CompletionLookupProvider completionLookupProvider = providers.get(providerName);
        if (completionLookupProvider == null) {
            throw new IllegalStateException("no provider with name [" + providerName + "] registered");
        }
        // TODO: we could clone the ReadState and make it always forward IOContext.MERGE to prevent unecessary heap usage?
        delegateProducer = delegatePostingsFormat.fieldsProducer(state);
        /*
         * If we are merging we don't load the FSTs at all such that we
         * don't consume so much memory during merge
         */
        if (state.context.context != Context.MERGE) {
            // TODO: maybe we can do this in a fully lazy fashion based on some configuration
            // eventually we should have some kind of curciut breaker that prevents us from going OOM here
            // with some configuration
            this.lookupFactory = completionLookupProvider.load(input);
        } else {
            this.lookupFactory = null;
        }
        this.delegateProducer = delegateProducer;
        success = true;
    } finally {
        if (!success) {
            IOUtils.closeWhileHandlingException(delegateProducer, input);
        } else {
            IOUtils.close(input);
        }
    }
}
 
Example 3
Source File: BloomFilteringPostingsFormat.java    From lucene-solr with Apache License 2.0 5 votes vote down vote up
public BloomFilteredFieldsProducer(SegmentReadState state)
    throws IOException {
  
  String bloomFileName = IndexFileNames.segmentFileName(
      state.segmentInfo.name, state.segmentSuffix, BLOOM_EXTENSION);
  ChecksumIndexInput bloomIn = null;
  boolean success = false;
  try {
    bloomIn = state.directory.openChecksumInput(bloomFileName, state.context);
    CodecUtil.checkIndexHeader(bloomIn, BLOOM_CODEC_NAME, VERSION_START, VERSION_CURRENT, state.segmentInfo.getId(), state.segmentSuffix);
    // // Load the hash function used in the BloomFilter
    // hashFunction = HashFunction.forName(bloomIn.readString());
    // Load the delegate postings format
    PostingsFormat delegatePostingsFormat = PostingsFormat.forName(bloomIn
        .readString());
    
    this.delegateFieldsProducer = delegatePostingsFormat
        .fieldsProducer(state);
    int numBlooms = bloomIn.readInt();
    for (int i = 0; i < numBlooms; i++) {
      int fieldNum = bloomIn.readInt();
      FuzzySet bloom = FuzzySet.deserialize(bloomIn);
      FieldInfo fieldInfo = state.fieldInfos.fieldInfo(fieldNum);
      bloomsByFieldName.put(fieldInfo.name, bloom);
    }
    CodecUtil.checkFooter(bloomIn);
    IOUtils.close(bloomIn);
    success = true;
  } finally {
    if (!success) {
      IOUtils.closeWhileHandlingException(bloomIn, delegateFieldsProducer);
    }
  }
}
 
Example 4
Source File: CompletionFieldsProducer.java    From lucene-solr with Apache License 2.0 5 votes vote down vote up
CompletionFieldsProducer(String codecName, SegmentReadState state, FSTLoadMode fstLoadMode) throws IOException {
  String indexFile = IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, INDEX_EXTENSION);
  delegateFieldsProducer = null;
  boolean success = false;

  try (ChecksumIndexInput index = state.directory.openChecksumInput(indexFile, state.context)) {
    // open up dict file containing all fsts
    String dictFile = IndexFileNames.segmentFileName(state.segmentInfo.name, state.segmentSuffix, DICT_EXTENSION);
    dictIn = state.directory.openInput(dictFile, state.context);
    CodecUtil.checkIndexHeader(dictIn, codecName, COMPLETION_CODEC_VERSION, COMPLETION_VERSION_CURRENT, state.segmentInfo.getId(), state.segmentSuffix);
    // just validate the footer for the dictIn
    CodecUtil.retrieveChecksum(dictIn);

    // open up index file (fieldNumber, offset)
    CodecUtil.checkIndexHeader(index, codecName, COMPLETION_CODEC_VERSION, COMPLETION_VERSION_CURRENT, state.segmentInfo.getId(), state.segmentSuffix);
    // load delegate PF
    PostingsFormat delegatePostingsFormat = PostingsFormat.forName(index.readString());
    delegateFieldsProducer = delegatePostingsFormat.fieldsProducer(state);

    // read suggest field numbers and their offsets in the terms file from index
    int numFields = index.readVInt();
    readers = new HashMap<>(numFields);
    for (int i = 0; i < numFields; i++) {
      int fieldNumber = index.readVInt();
      long offset = index.readVLong();
      long minWeight = index.readVLong();
      long maxWeight = index.readVLong();
      byte type = index.readByte();
      FieldInfo fieldInfo = state.fieldInfos.fieldInfo(fieldNumber);
      // we don't load the FST yet
      readers.put(fieldInfo.name, new CompletionsTermsReader(dictIn, offset, minWeight, maxWeight, type, fstLoadMode));
    }
    CodecUtil.checkFooter(index);
    success = true;
  } finally {
    if (success == false) {
      IOUtils.closeWhileHandlingException(delegateFieldsProducer, dictIn);
    }
  }
}
 
Example 5
Source File: PerFieldPostingsFormat.java    From lucene-solr with Apache License 2.0 5 votes vote down vote up
public FieldsReader(final SegmentReadState readState) throws IOException {

      // Read _X.per and init each format:
      boolean success = false;
      try {
        // Read field name -> format name
        for (FieldInfo fi : readState.fieldInfos) {
          if (fi.getIndexOptions() != IndexOptions.NONE) {
            final String fieldName = fi.name;
            final String formatName = fi.getAttribute(PER_FIELD_FORMAT_KEY);
            if (formatName != null) {
              // null formatName means the field is in fieldInfos, but has no postings!
              final String suffix = fi.getAttribute(PER_FIELD_SUFFIX_KEY);
              if (suffix == null) {
                throw new IllegalStateException("missing attribute: " + PER_FIELD_SUFFIX_KEY + " for field: " + fieldName);
              }
              PostingsFormat format = PostingsFormat.forName(formatName);
              String segmentSuffix = getSuffix(formatName, suffix);
              if (!formats.containsKey(segmentSuffix)) {
                formats.put(segmentSuffix, format.fieldsProducer(new SegmentReadState(readState, segmentSuffix)));
              }
              fields.put(fieldName, formats.get(segmentSuffix));
            }
          }
        }
        success = true;
      } finally {
        if (!success) {
          IOUtils.closeWhileHandlingException(formats.values());
        }
      }

      this.segment = readState.segmentInfo.name;
    }
 
Example 6
Source File: Completion84PostingsFormat.java    From lucene-solr with Apache License 2.0 4 votes vote down vote up
@Override
protected PostingsFormat delegatePostingsFormat() {
  return PostingsFormat.forName("Lucene84");
}
 
Example 7
Source File: Completion50PostingsFormat.java    From lucene-solr with Apache License 2.0 4 votes vote down vote up
@Override
protected PostingsFormat delegatePostingsFormat() {
  return PostingsFormat.forName("Lucene50");
}