From 6829418183902a104edf7f7572da0ef42c56ee9e Mon Sep 17 00:00:00 2001 From: zhouhui Date: Thu, 30 May 2024 15:32:12 +0800 Subject: [PATCH] Init. --- .../lucene90/blocktree/SegmentTermsEnum.java | 27 +++++++++++++++++ .../blocktree/SegmentTermsEnumFrame.java | 30 +++++++++++++++++++ .../index/BasePostingsFormatTestCase.java | 5 ++++ 3 files changed, 62 insertions(+) diff --git a/lucene/core/src/java/org/apache/lucene/codecs/lucene90/blocktree/SegmentTermsEnum.java b/lucene/core/src/java/org/apache/lucene/codecs/lucene90/blocktree/SegmentTermsEnum.java index e3389931be7..6dec0d7bb10 100644 --- a/lucene/core/src/java/org/apache/lucene/codecs/lucene90/blocktree/SegmentTermsEnum.java +++ b/lucene/core/src/java/org/apache/lucene/codecs/lucene90/blocktree/SegmentTermsEnum.java @@ -584,6 +584,22 @@ final class SegmentTermsEnum extends BaseTermsEnum { throw new IllegalStateException("terms index was not loaded"); } + if (fr.size() > 0) { + if(target.compareTo(fr.getMin()) < 0){ + // TODO: set current frame to min entry's block. + termExists = true; + term.copyBytes(fr.getMin()); + System.out.println("target: " + target.utf8ToString() + ", term: " + term().utf8ToString() + ", frame: " + currentFrame); + return SeekStatus.NOT_FOUND; + }else if(target.compareTo(fr.getMax()) > 0){ + // TODO: set current frame to max entry's block. + termExists = false; + term.clear(); + System.out.println("target: " + target.utf8ToString() + ", term: " + term().utf8ToString() + ", frame: " + currentFrame); + return SeekStatus.END; + } + } + term.grow(1 + target.length); assert clearEOF(); @@ -695,6 +711,8 @@ final class SegmentTermsEnum extends BaseTermsEnum { // if (DEBUG) { // System.out.println(" target is same as current; return FOUND"); // } + System.out.println("target: " + target.utf8ToString() + ", term: " + term().utf8ToString() + ", frame: " + currentFrame); + return SeekStatus.FOUND; } else { // if (DEBUG) { @@ -768,11 +786,13 @@ final class SegmentTermsEnum extends BaseTermsEnum { // System.out.println(" return NOT_FOUND term=" + // ToStringUtils.bytesRefToString(term)); // } + System.out.println("target: " + target.utf8ToString() + ", term: " + term().utf8ToString() + ", frame: " + currentFrame); return SeekStatus.NOT_FOUND; } else { // if (DEBUG) { // System.out.println(" return END"); // } + System.out.println("target: " + target.utf8ToString() + ", term: " + term().utf8ToString() + ", frame: " + currentFrame); return SeekStatus.END; } } else { @@ -780,6 +800,7 @@ final class SegmentTermsEnum extends BaseTermsEnum { // System.out.println(" return " + result + " term=" + // ToStringUtils.bytesRefToString(term)); // } + System.out.println("target: " + target.utf8ToString() + ", term: " + term().utf8ToString() + ", frame: " + currentFrame); return result; } } else { @@ -823,14 +844,20 @@ final class SegmentTermsEnum extends BaseTermsEnum { // if (DEBUG) { // System.out.println(" return NOT_FOUND term=" + term.get().utf8ToString() + " " + term); // } + System.out.println("target: " + target.utf8ToString() + ", term: " + term().utf8ToString() + ", frame: " + currentFrame); + return SeekStatus.NOT_FOUND; } else { // if (DEBUG) { // System.out.println(" return END"); // } + System.out.println("target: " + target.utf8ToString() + ", term: " + term().utf8ToString() + ", frame: " + currentFrame); + return SeekStatus.END; } } else { + System.out.println("target: " + target.utf8ToString() + ", term: " + term().utf8ToString() + ", frame: " + currentFrame); + return result; } } diff --git a/lucene/core/src/java/org/apache/lucene/codecs/lucene90/blocktree/SegmentTermsEnumFrame.java b/lucene/core/src/java/org/apache/lucene/codecs/lucene90/blocktree/SegmentTermsEnumFrame.java index c1552c8ce60..49750fd49e0 100644 --- a/lucene/core/src/java/org/apache/lucene/codecs/lucene90/blocktree/SegmentTermsEnumFrame.java +++ b/lucene/core/src/java/org/apache/lucene/codecs/lucene90/blocktree/SegmentTermsEnumFrame.java @@ -854,4 +854,34 @@ final class SegmentTermsEnumFrame { ste.term.grow(termLength); System.arraycopy(suffixBytes, startBytePos, ste.term.bytes(), prefix, suffix); } + + @Override + public String toString() { + return "fp: " + + fp + + ", fpOrig: " + + fpOrig + + ", fpEnd: " + + fpEnd + + ", lastSubFP: " + + lastSubFP + + ", entCount: " + + entCount + + ", nextEnt: " + + nextEnt + + ", isLeafBlock: " + + isLeafBlock + + ", isFloor: " + + isFloor + + ", isLastInFloor: " + + isLastInFloor + + ", nextFloorLabel: " + + nextFloorLabel + + ", suffixesPos: " + + suffixesReader.getPosition() + + ", suffixLengthsPos: " + + suffixLengthsReader.getPosition() + + ", floorDataPos: " + + floorDataReader.getPosition(); + } } diff --git a/lucene/test-framework/src/java/org/apache/lucene/tests/index/BasePostingsFormatTestCase.java b/lucene/test-framework/src/java/org/apache/lucene/tests/index/BasePostingsFormatTestCase.java index 8f8233ee680..b12bd7b00e7 100644 --- a/lucene/test-framework/src/java/org/apache/lucene/tests/index/BasePostingsFormatTestCase.java +++ b/lucene/test-framework/src/java/org/apache/lucene/tests/index/BasePostingsFormatTestCase.java @@ -455,6 +455,11 @@ public abstract class BasePostingsFormatTestCase extends BaseIndexFileFormatTest } } assertEquals(SeekStatus.END, termsEnum.seekCeil(new BytesRef(100400 + ""))); + + assertEquals(SeekStatus.NOT_FOUND, termsEnum.seekCeil(new BytesRef(100000 + ""))); + assertEquals(new BytesRef(100001 + ""), termsEnum.term()); + assertEquals(SeekStatus.END, termsEnum.seekCeil(new BytesRef(9 + ""))); + reader.close(); iw.close(); dir.close();