@Test
public void test_inflate_impl() {
for (int l = 0; l < 2; ++l) {
NewChunk nc = new NewChunk(null, 0);
// -32.767, 0.34, 0, 32.767, NA for l==0
// NA, -32.767, 0.34, 0, 32.767, NA for l==1
long[] man = new long[] {-32767, 34, 0, 32767};
int[] exp = new int[] {-3, -2, 1, -3};
if (l == 1) nc.addNA(); // -32768
for (int i = 0; i < man.length; ++i) nc.addNum(man[i], exp[i]);
nc.addNA();
Chunk cc = nc.compress();
Assert.assertEquals(man.length + 1 + l, cc.len());
Assert.assertTrue(cc instanceof C2SChunk);
if (l == 1) {
Assert.assertTrue(cc.isNA0(0));
Assert.assertTrue(cc.isNA(0));
}
for (int i = 0; i < man.length; ++i) {
Assert.assertEquals((float) (man[i] * Math.pow(10, exp[i])), (float) cc.at0(l + i), 0);
Assert.assertEquals((float) (man[i] * Math.pow(10, exp[i])), (float) cc.at(l + i), 0);
}
Assert.assertTrue(cc.isNA0(man.length + l));
Assert.assertTrue(cc.isNA(man.length + l));
nc = cc.inflate_impl(new NewChunk(null, 0));
nc.values(0, nc.len());
Assert.assertEquals(man.length + 1 + l, nc.len());
Assert.assertEquals(man.length + 1 + l, nc.sparseLen());
if (l == 1) {
Assert.assertTrue(nc.isNA0(0));
Assert.assertTrue(nc.isNA(0));
}
for (int i = 0; i < man.length; ++i) {
Assert.assertEquals((float) (man[i] * Math.pow(10, exp[i])), (float) nc.at0(l + i), 0);
Assert.assertEquals((float) (man[i] * Math.pow(10, exp[i])), (float) nc.at(l + i), 0);
}
Assert.assertTrue(nc.isNA0(man.length + l));
Assert.assertTrue(nc.isNA(man.length + l));
Chunk cc2 = nc.compress();
Assert.assertEquals(man.length + 1 + l, cc.len());
if (l == 1) {
Assert.assertTrue(cc2.isNA0(0));
Assert.assertTrue(cc2.isNA(0));
}
for (int i = 0; i < man.length; ++i) {
Assert.assertEquals((float) (man[i] * Math.pow(10, exp[i])), (float) cc2.at0(l + i), 0);
Assert.assertEquals((float) (man[i] * Math.pow(10, exp[i])), (float) cc2.at(l + i), 0);
}
Assert.assertTrue(cc2.isNA0(man.length + l));
Assert.assertTrue(cc2.isNA(man.length + l));
Assert.assertTrue(cc2 instanceof C2SChunk);
Assert.assertTrue(Arrays.equals(cc._mem, cc2._mem));
}
}
public void addStr(Chunk c, int row) {
if (c.isNA(row)) addNA();
else {
addStr(c.atStr(new BufferedString(), row));
_isAllASCII &= ((CStrChunk) c)._isAllASCII;
}
}
@Test
public void test_setNA() {
// Create a vec with one chunk with 15 elements, and set its numbers
Vec vec = new Vec(Vec.newKey(), new long[] {0, 15}).makeZeros(1, null, null, null, null)[0];
int[] vals = new int[] {0, 3, 0, 6, 0, 0, 0, -32769, 0, 12, 234, 32765, 0, 0, 19};
Vec.Writer w = vec.open();
for (int i = 0; i < vals.length; ++i) w.set(i, vals[i]);
w.close();
Chunk cc = vec.chunkForChunkIdx(0);
assert cc instanceof C2SChunk;
Futures fs = new Futures();
fs.blockForPending();
for (int i = 0; i < vals.length; ++i) Assert.assertEquals(vals[i], cc.at80(i));
for (int i = 0; i < vals.length; ++i) Assert.assertEquals(vals[i], cc.at8(i));
int[] NAs = new int[] {1, 5, 2};
int[] notNAs = new int[] {0, 3, 4, 6, 7, 8, 9, 10, 11, 12, 13, 14};
for (int na : NAs) cc.setNA(na);
for (int na : NAs) Assert.assertTrue(cc.isNA0(na));
for (int na : NAs) Assert.assertTrue(cc.isNA(na));
for (int notna : notNAs) Assert.assertTrue(!cc.isNA0(notna));
for (int notna : notNAs) Assert.assertTrue(!cc.isNA(notna));
NewChunk nc = new NewChunk(null, 0);
cc.inflate_impl(nc);
nc.values(0, nc.len());
Assert.assertEquals(vals.length, nc.sparseLen());
Assert.assertEquals(vals.length, nc.len());
Iterator<NewChunk.Value> it = nc.values(0, vals.length);
for (int i = 0; i < vals.length; ++i) Assert.assertTrue(it.next().rowId0() == i);
Assert.assertTrue(!it.hasNext());
for (int na : NAs) Assert.assertTrue(cc.isNA0(na));
for (int na : NAs) Assert.assertTrue(cc.isNA(na));
for (int notna : notNAs) Assert.assertTrue(!cc.isNA0(notna));
for (int notna : notNAs) Assert.assertTrue(!cc.isNA(notna));
Chunk cc2 = nc.compress();
Assert.assertEquals(vals.length, cc.len());
Assert.assertTrue(cc2 instanceof C2SChunk);
for (int na : NAs) Assert.assertTrue(cc.isNA0(na));
for (int na : NAs) Assert.assertTrue(cc.isNA(na));
for (int notna : notNAs) Assert.assertTrue(!cc.isNA0(notna));
for (int notna : notNAs) Assert.assertTrue(!cc.isNA(notna));
Assert.assertTrue(Arrays.equals(cc._mem, cc2._mem));
vec.remove();
}
@Override
public void map(Chunk cs) {
int idx = _chunkOffset + cs.cidx();
Key ckey = Vec.chunkKey(_v._key, idx);
if (_cmap != null) {
assert !cs.hasFloat()
: "Input chunk (" + cs.getClass() + ") has float, but is expected to be categorical";
NewChunk nc = new NewChunk(_v, idx);
// loop over rows and update ints for new domain mapping according to vecs[c].domain()
for (int r = 0; r < cs._len; ++r) {
if (cs.isNA(r)) nc.addNA();
else nc.addNum(_cmap[(int) cs.at8(r)], 0);
}
nc.close(_fs);
} else {
DKV.put(ckey, cs.deepCopy(), _fs, true);
}
}
public void addUUID(Chunk c, int row) {
if (c.isNA(row)) addUUID(C16Chunk._LO_NA, C16Chunk._HI_NA);
else addUUID(c.at16l(row), c.at16h(row));
}
public void addStr(Chunk c, int row) {
if (c.isNA(row)) addNA();
else addStr(c.atStr(new ValueString(), row));
}
/**
* Extract (sparse) rows from given chunks. Note: 0 remains 0 - _normSub of DataInfo isn't used
* (mean shift during standarization is not reverted) - UNLESS offset is specified (for GLM only)
* Essentially turns the dataset 90 degrees.
*
* @param chunks - chunk of dataset
* @return array of sparse rows
*/
public final Row[] extractSparseRows(Chunk[] chunks) {
Row[] rows = new Row[chunks[0]._len];
long startOff = chunks[0].start();
for (int i = 0; i < rows.length; ++i) {
rows[i] =
new Row(
true,
Math.min(_nums, 16),
_cats,
_responses,
i,
startOff); // if sparse, _nums is the correct number of nonzero values! i.e., do not
// use numNums()
rows[i].rid = chunks[0].start() + i;
if (_offset) {
rows[i].offset = chunks[offsetChunkId()].atd(i);
if (Double.isNaN(rows[i].offset)) rows[i].bad = true;
}
if (_weights) {
rows[i].weight = chunks[weightChunkId()].atd(i);
if (Double.isNaN(rows[i].weight)) rows[i].bad = true;
}
if (_skipMissing) {
int N = _cats + _nums;
for (int c = 0; c < N; ++c) if (chunks[c].isNA(i)) rows[i].bad = true;
}
}
// categoricals
for (int i = 0; i < _cats; ++i) {
for (int r = 0; r < chunks[0]._len; ++r) {
Row row = rows[r];
if (row.bad) continue;
int cid = getCategoricalId(i, chunks[i].isNA(r) ? _catModes[i] : (int) chunks[i].at8(r));
if (cid >= 0) row.binIds[row.nBins++] = cid;
}
}
// generic numbers + interactions
int interactionOffset = 0;
for (int cid = 0; cid < _nums; ++cid) {
Chunk c = chunks[_cats + cid];
int oldRow = -1;
if (c
instanceof
InteractionWrappedVec
.InteractionWrappedChunk) { // for each row, only 1 value in an interaction is 'hot'
// all other values are off (i.e., are 0)
for (int r = 0;
r < c._len;
++r) { // the vec is "vertically" dense and "horizontally" sparse (i.e., every row has
// one, and only one, value)
Row row = rows[r];
if (row.bad) continue;
if (c.isNA(r)) row.bad = _skipMissing;
int cidVirtualOffset =
getInteractionOffset(
chunks, _cats + cid, r); // the "virtual" offset into the hot-expanded interaction
row.addNum(
_numOffsets[cid] + cidVirtualOffset,
c.atd(r)); // FIXME: if this produces a "true" NA then should sub with mean? with?
}
interactionOffset += nextNumericIdx(cid);
} else {
for (int r = c.nextNZ(-1); r < c._len; r = c.nextNZ(r)) {
if (c.atd(r) == 0) continue;
assert r > oldRow;
oldRow = r;
Row row = rows[r];
if (row.bad) continue;
if (c.isNA(r)) row.bad = _skipMissing;
double d = c.atd(r);
if (Double.isNaN(d)) d = _numMeans[cid];
if (_normMul != null) d *= _normMul[interactionOffset];
row.addNum(_numOffsets[cid], d);
}
interactionOffset++;
}
}
// response(s)
for (int i = 1; i <= _responses; ++i) {
int rid = responseChunkId(i - 1);
Chunk rChunk = chunks[rid];
for (int r = 0; r < chunks[0]._len; ++r) {
Row row = rows[r];
if (row.bad) continue;
row.response[i - 1] = rChunk.atd(r);
if (_normRespMul != null) {
row.response[i - 1] = (row.response[i - 1] - _normRespSub[i - 1]) * _normRespMul[i - 1];
}
if (Double.isNaN(row.response[row.response.length - i])) row.bad = true;
}
}
return rows;
}
