Examples with MRTask water.MRTask used on opensource projects

Example 51 with MRTask

use of water.MRTask in project h2o-3 by h2oai.

the class AstEntropy method entropyStringCol.

private Vec entropyStringCol(Vec vec) {
    return new MRTask() {

        @Override
        public void map(Chunk chk, NewChunk newChk) {
            if (//all NAs
            chk instanceof C0DChunk)
                newChk.addNAs(chk.len());
            else if (//fast-path operations
            ((CStrChunk) chk)._isAllASCII)
                ((CStrChunk) chk).asciiEntropy(newChk);
            else {
                //UTF requires Java string methods
                BufferedString tmpStr = new BufferedString();
                for (int i = 0; i < chk._len; i++) {
                    if (chk.isNA(i))
                        newChk.addNA();
                    else {
                        String str = chk.atStr(tmpStr, i).toString();
                        newChk.addNum(calcEntropy(str));
                    }
                }
            }
        }
    }.doAll(new byte[] { Vec.T_NUM }, vec).outputFrame().anyVec();
}

Example 52 with MRTask

use of water.MRTask in project h2o-3 by h2oai.

the class AstReLevel method apply.

@Override
public ValFrame apply(Env env, Env.StackHelp stk, AstRoot[] asts) {
    Frame fr = stk.track(asts[1].exec(env)).getFrame();
    if (fr.numCols() != 1)
        throw new IllegalArgumentException("`setLevel` works on a single column at a time.");
    String[] doms = fr.anyVec().domain().clone();
    if (doms == null)
        throw new IllegalArgumentException("Cannot set the level on a non-factor column!");
    String lvl = asts[2].exec(env).getStr();
    final int idx = Arrays.asList(doms).indexOf(lvl);
    if (idx == -1)
        throw new IllegalArgumentException("Did not find level `" + lvl + "` in the column.");
    if (idx == 0)
        return new ValFrame(new Frame(fr.names(), new Vec[] { fr.anyVec().makeCopy() }));
    String[] srcDom = fr.anyVec().domain();
    final String[] dom = new String[srcDom.length];
    dom[0] = srcDom[idx];
    int j = 1;
    for (int i = 0; i < srcDom.length; ++i) if (i != idx)
        dom[j++] = srcDom[i];
    return new ValFrame(new MRTask() {

        @Override
        public void map(Chunk c, NewChunk nc) {
            int[] vals = new int[c._len];
            c.getIntegers(vals, 0, c._len, -1);
            for (int i = 0; i < vals.length; ++i) if (vals[i] == -1)
                nc.addNA();
            else if (vals[i] == idx)
                nc.addNum(0);
            else
                nc.addNum(vals[i] + (vals[i] < idx ? 1 : 0));
        }
    }.doAll(1, Vec.T_CAT, fr).outputFrame(fr.names(), new String[][] { dom }));
}

Example 53 with MRTask

use of water.MRTask in project h2o-3 by h2oai.

the class AstSetLevel method apply.

@Override
public ValFrame apply(Env env, Env.StackHelp stk, AstRoot[] asts) {
    Frame fr = stk.track(asts[1].exec(env)).getFrame();
    if (fr.numCols() != 1)
        throw new IllegalArgumentException("`setLevel` works on a single column at a time.");
    String[] doms = fr.anyVec().domain().clone();
    if (doms == null)
        throw new IllegalArgumentException("Cannot set the level on a non-factor column!");
    String lvl = asts[2].exec(env).getStr();
    final int idx = Arrays.asList(doms).indexOf(lvl);
    if (idx == -1)
        throw new IllegalArgumentException("Did not find level `" + lvl + "` in the column.");
    // COW semantics
    Frame fr2 = new MRTask() {

        @Override
        public void map(Chunk c, NewChunk nc) {
            for (int i = 0; i < c._len; ++i) nc.addNum(idx);
        }
    }.doAll(new byte[] { Vec.T_NUM }, fr.anyVec()).outputFrame(null, fr.names(), fr.domains());
    return new ValFrame(fr2);
}

Example 54 with MRTask

use of water.MRTask in project h2o-3 by h2oai.

the class AstBinOp method frame_op_scalar.

/**
   * Auto-widen the scalar to every element of the frame
   */
private ValFrame frame_op_scalar(Frame fr, final String str) {
    Frame res = new MRTask() {

        @Override
        public void map(Chunk[] chks, NewChunk[] cress) {
            BufferedString vstr = new BufferedString();
            for (int c = 0; c < chks.length; c++) {
                Chunk chk = chks[c];
                NewChunk cres = cress[c];
                Vec vec = chk.vec();
                // String Vectors: apply str_op as BufferedStrings to all elements
                if (vec.isString()) {
                    final BufferedString conStr = new BufferedString(str);
                    for (int i = 0; i < chk._len; i++) cres.addNum(str_op(chk.atStr(vstr, i), conStr));
                } else if (vec.isCategorical()) {
                    // categorical Vectors: convert string to domain value; apply op (not
                    // str_op).  Not sure what the "right" behavior here is, can
                    // easily argue that should instead apply str_op to the categorical
                    // string domain value - except that this whole operation only
                    // makes sense for EQ/NE, and is much faster when just comparing
                    // doubles vs comparing strings.  Note that if the string is not
                    // part of the categorical domain, the find op returns -1 which is never
                    // equal to any categorical dense integer (which are always 0+).
                    final double d = (double) ArrayUtils.find(vec.domain(), str);
                    for (int i = 0; i < chk._len; i++) cres.addNum(op(chk.atd(i), d));
                } else {
                    // mixing string and numeric
                    // false or true only
                    final double d = op(1, 2);
                    for (int i = 0; i < chk._len; i++) cres.addNum(d);
                }
            }
        }
    }.doAll(fr.numCols(), Vec.T_NUM, fr).outputFrame(fr._names, null);
    return new ValFrame(res);
}

Example 55 with MRTask

use of water.MRTask in project h2o-3 by h2oai.

the class AstGroupedPermute method buildOutput.

private static Frame buildOutput(final double[][][] a, String[] names, String[][] domains) {
    Frame dVec = new Frame(Vec.makeSeq(0, a.length));
    long s = System.currentTimeMillis();
    Frame res = new MRTask() {

        @Override
        public void map(Chunk[] cs, NewChunk[] ncs) {
            for (int i = 0; i < cs[0]._len; ++i) for (double[] anAa : a[(int) cs[0].at8(i)]) for (int k = 0; k < anAa.length; ++k) ncs[k].addNum(anAa[k]);
        }
    }.doAll(5, Vec.T_NUM, dVec).outputFrame(null, names, domains);
    Log.info("Elapsed time: " + (System.currentTimeMillis() - s) / 1000. + "s");
    dVec.delete();
    return res;
}