Examples with ResizeFactor org.apache.datasketches.ResizeFactor used on opensource projects

Example 21 with ResizeFactor

use of org.apache.datasketches.ResizeFactor in project sketches-core by DataSketches.

the class SketchTest method checkBuilder.

@Test
public void checkBuilder() {
    int k = 2048;
    int lgK = Integer.numberOfTrailingZeros(k);
    long seed = 1021;
    float p = (float) 0.5;
    ResizeFactor rf = X4;
    Family fam = Family.ALPHA;
    UpdateSketch sk1 = UpdateSketch.builder().setSeed(seed).setP(p).setResizeFactor(rf).setFamily(fam).setNominalEntries(k).build();
    String nameS1 = sk1.getClass().getSimpleName();
    assertEquals(nameS1, "HeapAlphaSketch");
    assertEquals(sk1.getLgNomLongs(), lgK);
    assertEquals(sk1.getSeed(), seed);
    assertEquals(sk1.getP(), p);
    // check reset of defaults
    sk1 = UpdateSketch.builder().build();
    nameS1 = sk1.getClass().getSimpleName();
    assertEquals(nameS1, "HeapQuickSelectSketch");
    assertEquals(sk1.getLgNomLongs(), Integer.numberOfTrailingZeros(DEFAULT_NOMINAL_ENTRIES));
    assertEquals(sk1.getSeed(), DEFAULT_UPDATE_SEED);
    assertEquals(sk1.getP(), (float) 1.0);
    assertEquals(sk1.getResizeFactor(), ResizeFactor.X8);
}

Example 22 with ResizeFactor

use of org.apache.datasketches.ResizeFactor in project sketches-core by DataSketches.

the class SketchTest method checkBuilderResizeFactor.

@Test
public void checkBuilderResizeFactor() {
    ResizeFactor rf;
    rf = X1;
    assertEquals(rf.getValue(), 1);
    assertEquals(rf.lg(), 0);
    assertEquals(ResizeFactor.getRF(0), X1);
    rf = X2;
    assertEquals(rf.getValue(), 2);
    assertEquals(rf.lg(), 1);
    assertEquals(ResizeFactor.getRF(1), X2);
    rf = X4;
    assertEquals(rf.getValue(), 4);
    assertEquals(rf.lg(), 2);
    assertEquals(ResizeFactor.getRF(2), X4);
    rf = X8;
    assertEquals(rf.getValue(), 8);
    assertEquals(rf.lg(), 3);
    assertEquals(ResizeFactor.getRF(3), X8);
}

Example 23 with ResizeFactor

use of org.apache.datasketches.ResizeFactor in project sketches-core by DataSketches.

the class HeapAlphaSketch method resizeClean.

// curCount > hashTableThreshold
// Checks for rare lockup condition
// Used by hashUpdate()
private final void resizeClean() {
    // must resize, but are we at tgt size?
    final int lgTgtLongs = lgNomLongs_ + 1;
    if (lgTgtLongs > lgArrLongs_) {
        // not yet at tgt size
        final ResizeFactor rf = getResizeFactor();
        // must be > 0
        final int lgDeltaLongs = lgTgtLongs - lgArrLongs_;
        // rf_.lg() could be 0
        final int lgResizeFactor = max(min(rf.lg(), lgDeltaLongs), 1);
        forceResizeCleanCache(lgResizeFactor);
    } else {
        // at tgt size or larger, no dirty values, must take drastic measures, very rare.
        forceResizeCleanCache(1);
    }
}

Example 24 with ResizeFactor

use of org.apache.datasketches.ResizeFactor in project sketches-core by DataSketches.

the class PreambleUtil method preambleToString.

/**
 * Returns a human readable string summary of the preamble state of the given Memory.
 * Note: other than making sure that the given Memory size is large
 * enough for just the preamble, this does not do much value checking of the contents of the
 * preamble as this is primarily a tool for debugging the preamble visually.
 *
 * @param mem the given Memory.
 * @return the summary preamble string.
 */
static String preambleToString(final Memory mem) {
    final int preLongs = getAndCheckPreLongs(mem);
    final int rfId = extractLgResizeFactor(mem);
    final ResizeFactor rf = ResizeFactor.getRF(rfId);
    final int serVer = extractSerVer(mem);
    final int familyId = extractFamilyID(mem);
    final Family family = Family.idToFamily(familyId);
    final int lgNomLongs = extractLgNomLongs(mem);
    final int lgArrLongs = extractLgArrLongs(mem);
    // Flags
    final int flags = extractFlags(mem);
    final String flagsStr = (flags) + ", 0x" + (Integer.toHexString(flags)) + ", " + zeroPad(Integer.toBinaryString(flags), 8);
    final String nativeOrder = ByteOrder.nativeOrder().toString();
    final boolean bigEndian = (flags & BIG_ENDIAN_FLAG_MASK) > 0;
    final boolean readOnly = (flags & READ_ONLY_FLAG_MASK) > 0;
    final boolean empty = (flags & EMPTY_FLAG_MASK) > 0;
    final boolean compact = (flags & COMPACT_FLAG_MASK) > 0;
    final boolean ordered = (flags & ORDERED_FLAG_MASK) > 0;
    // !empty && (preLongs == 1);
    final boolean singleItem = (flags & SINGLEITEM_FLAG_MASK) > 0;
    final int seedHash = extractSeedHash(mem);
    // assumes preLongs == 1; empty or singleItem
    int curCount = singleItem ? 1 : 0;
    // preLongs 1 or 2
    float p = (float) 1.0;
    // preLongs 1 or 2
    long thetaLong = Long.MAX_VALUE;
    // preLongs 1, 2 or 3
    long thetaULong = thetaLong;
    if (preLongs == 2) {
        // exact (non-estimating) CompactSketch
        curCount = extractCurCount(mem);
        p = extractP(mem);
    } else if (preLongs == 3) {
        // Update Sketch
        curCount = extractCurCount(mem);
        p = extractP(mem);
        thetaLong = extractThetaLong(mem);
        thetaULong = thetaLong;
    } else if (preLongs == 4) {
        // Union
        curCount = extractCurCount(mem);
        p = extractP(mem);
        thetaLong = extractThetaLong(mem);
        thetaULong = extractUnionThetaLong(mem);
    }
    // else the same as an empty sketch or singleItem
    final double thetaDbl = thetaLong / Util.LONG_MAX_VALUE_AS_DOUBLE;
    final String thetaHex = zeroPad(Long.toHexString(thetaLong), 16);
    final double thetaUDbl = thetaULong / Util.LONG_MAX_VALUE_AS_DOUBLE;
    final String thetaUHex = zeroPad(Long.toHexString(thetaULong), 16);
    final StringBuilder sb = new StringBuilder();
    sb.append(LS);
    sb.append("### SKETCH PREAMBLE SUMMARY:").append(LS);
    sb.append("Native Byte Order             : ").append(nativeOrder).append(LS);
    sb.append("Byte  0: Preamble Longs       : ").append(preLongs).append(LS);
    sb.append("Byte  0: ResizeFactor         : ").append(rfId + ", " + rf.toString()).append(LS);
    sb.append("Byte  1: Serialization Version: ").append(serVer).append(LS);
    sb.append("Byte  2: Family               : ").append(familyId + ", " + family.toString()).append(LS);
    sb.append("Byte  3: LgNomLongs           : ").append(lgNomLongs).append(LS);
    sb.append("Byte  4: LgArrLongs           : ").append(lgArrLongs).append(LS);
    sb.append("Byte  5: Flags Field          : ").append(flagsStr).append(LS);
    sb.append("  Bit Flag Name               : State:").append(LS);
    sb.append("    0 BIG_ENDIAN_STORAGE      : ").append(bigEndian).append(LS);
    sb.append("    1 READ_ONLY               : ").append(readOnly).append(LS);
    sb.append("    2 EMPTY                   : ").append(empty).append(LS);
    sb.append("    3 COMPACT                 : ").append(compact).append(LS);
    sb.append("    4 ORDERED                 : ").append(ordered).append(LS);
    sb.append("    5 SINGLE_ITEM             : ").append(singleItem).append(LS);
    sb.append("Bytes 6-7  : Seed Hash Hex    : ").append(Integer.toHexString(seedHash)).append(LS);
    if (preLongs == 1) {
        sb.append(" --ABSENT FIELDS, ASSUMED:").append(LS);
        sb.append("Bytes 8-11 : CurrentCount     : ").append(curCount).append(LS);
        sb.append("Bytes 12-15: P                : ").append(p).append(LS);
        sb.append("Bytes 16-23: Theta (double)   : ").append(thetaDbl).append(LS);
        sb.append("             Theta (long)     : ").append(thetaLong).append(LS);
        sb.append("             Theta (long,hex) : ").append(thetaHex).append(LS);
    } else if (preLongs == 2) {
        sb.append("Bytes 8-11 : CurrentCount     : ").append(curCount).append(LS);
        sb.append("Bytes 12-15: P                : ").append(p).append(LS);
        sb.append(" --ABSENT, ASSUMED:").append(LS);
        sb.append("Bytes 16-23: Theta (double)   : ").append(thetaDbl).append(LS);
        sb.append("             Theta (long)     : ").append(thetaLong).append(LS);
        sb.append("             Theta (long,hex) : ").append(thetaHex).append(LS);
    } else if (preLongs == 3) {
        sb.append("Bytes 8-11 : CurrentCount     : ").append(curCount).append(LS);
        sb.append("Bytes 12-15: P                : ").append(p).append(LS);
        sb.append("Bytes 16-23: Theta (double)   : ").append(thetaDbl).append(LS);
        sb.append("             Theta (long)     : ").append(thetaLong).append(LS);
        sb.append("             Theta (long,hex) : ").append(thetaHex).append(LS);
    } else {
        // preLongs == 4
        sb.append("Bytes 8-11 : CurrentCount     : ").append(curCount).append(LS);
        sb.append("Bytes 12-15: P                : ").append(p).append(LS);
        sb.append("Bytes 16-23: Theta (double)   : ").append(thetaDbl).append(LS);
        sb.append("             Theta (long)     : ").append(thetaLong).append(LS);
        sb.append("             Theta (long,hex) : ").append(thetaHex).append(LS);
        sb.append("Bytes 25-31: ThetaU (double)  : ").append(thetaUDbl).append(LS);
        sb.append("             ThetaU (long)    : ").append(thetaULong).append(LS);
        sb.append("             ThetaU (long,hex): ").append(thetaUHex).append(LS);
    }
    sb.append("Preamble Bytes                : ").append(preLongs * 8).append(LS);
    sb.append("Data Bytes                    : ").append(curCount * 8).append(LS);
    sb.append("TOTAL Sketch Bytes            : ").append((preLongs + curCount) * 8).append(LS);
    sb.append("TOTAL Capacity Bytes          : ").append(mem.getCapacity()).append(LS);
    sb.append("### END SKETCH PREAMBLE SUMMARY").append(LS);
    return sb.toString();
}