C++ TopSegmentIteratorPtr类代码示例

void TopSegmentSimpleIteratorTest::createCallBack(Alignment *alignment) {
    Genome *ancGenome = alignment->addRootGenome("Anc0", 0);
    size_t numChildren = 9;
    for (size_t i = 0; i < numChildren; ++i) {
        alignment->addLeafGenome("Leaf" + std::to_string(i), "Anc0", 0.1);
    }
    vector<Sequence::Info> seqVec(1);
    seqVec[0] = Sequence::Info("Sequence", 1000000, 5000, 10000);
    ancGenome->setDimensions(seqVec);

    CuAssertTrue(_testCase, ancGenome->getNumChildren() == numChildren);

    _topSegments.clear();
    for (size_t i = 0; i < ancGenome->getNumTopSegments(); ++i) {
        TopSegmentStruct topSeg;
        topSeg.setRandom();
        topSeg._length = ancGenome->getSequenceLength() / ancGenome->getNumTopSegments();
        topSeg._startPosition = i * topSeg._length;
        _topSegments.push_back(topSeg);
    }

    TopSegmentIteratorPtr tsIt = ancGenome->getTopSegmentIterator(0);
    for (size_t i = 0; not tsIt->atEnd(); tsIt->toRight(), ++i) {
        CuAssertTrue(_testCase, (size_t)tsIt->getTopSegment()->getArrayIndex() == i);
        _topSegments[i].applyTo(tsIt);
    }
}

void GappedSegmentSimpleIteratorTest2::createCallBack(AlignmentPtr alignment)
{
  addIdenticalParentChild(alignment, 2, 100, 5);
  Genome* parent = alignment->openGenome(alignment->getRootName());
  Genome* child = parent->getChild(0);
  TopSegmentIteratorPtr ti = child->getTopSegmentIterator();
  BottomSegmentIteratorPtr bi = parent->getBottomSegmentIterator();
  hal_index_t i = 0;
  bool reversed = true;
  while (ti != child->getTopSegmentEndIterator())
  {
    if (i % 5 == 0)
    {
      reversed = !reversed;
      if (reversed && i < (hal_index_t)(parent->getNumBottomSegments() - 1))
      {
        makeInversion(ti, 5);
      }
    }

    ti->toRight();
    bi->toRight();
    ++i;
  }
}

void LodExtract::writeUnsampledSequence(const Sequence* outSequence,
                                        SegmentIteratorPtr outSegment)
{
  outSegment->setCoordinates(outSequence->getStartPosition(),
                             outSequence->getSequenceLength());
  if (outSegment->isTop())
  {
    assert(outSequence->getNumTopSegments() == 1);
    TopSegmentIteratorPtr top = outSegment.downCast<TopSegmentIteratorPtr>();
    top->setParentIndex(NULL_INDEX);
    top->setParentReversed(false);
    top->setNextParalogyIndex(NULL_INDEX);
    top->setBottomParseIndex(NULL_INDEX);
  }
  else
  {
    assert(outSequence->getNumBottomSegments() == 1);
    BottomSegmentIteratorPtr bottom = 
       outSegment.downCast<BottomSegmentIteratorPtr>();
    hal_size_t numChildren = bottom->getNumChildren();
    for (hal_size_t childNum = 0; childNum < numChildren; ++childNum)
    {
      bottom->setChildIndex(childNum, NULL_INDEX);
      bottom->setChildReversed(childNum, false);
    }
    bottom->setTopParseIndex(NULL_INDEX);
  }
}

void SequenceIteratorTest::checkCallBack(const Alignment *alignment) {
    const Genome *ancGenome = alignment->openGenome("AncGenome");

    hal_size_t numSequences = ancGenome->getNumSequences();
    CuAssertTrue(_testCase, numSequences = 1000);

    for (SequenceIteratorPtr seqIt = ancGenome->getSequenceIterator(); not seqIt->atEnd(); seqIt->toNext()) {
        const Sequence *seq = seqIt->getSequence();
        hal_size_t i = seq->getArrayIndex();

        TopSegmentIteratorPtr tsIt = seq->getTopSegmentIterator();
        hal_size_t numTopSegments = seq->getNumTopSegments();
        for (hal_size_t j = 0; j < numTopSegments; ++j) {
            TopSegmentIteratorPtr gtsIt = ancGenome->getTopSegmentIterator((i - 1) * 100 + j);
            const TopSegment *gsTopSegment = gtsIt->getTopSegment();
            const TopSegment *sqTopSegment = tsIt->getTopSegment();

            CuAssertTrue(_testCase, gsTopSegment->getArrayIndex() == sqTopSegment->getArrayIndex());
            tsIt->toRight();
        }

        BottomSegmentIteratorPtr bsIt = seq->getBottomSegmentIterator();
        hal_size_t numBottomSegments = seq->getNumBottomSegments();
        for (hal_size_t j = 0; j < numBottomSegments; ++j) {
            BottomSegmentIteratorPtr gbsIt = ancGenome->getBottomSegmentIterator((i - 1) * 100 + j);
            const BottomSegment *gsBottomSegment = gbsIt->getBottomSegment();
            const BottomSegment *sqBottomSegment = bsIt->getBottomSegment();

            CuAssertTrue(_testCase, gsBottomSegment->getArrayIndex() == sqBottomSegment->getArrayIndex());
            bsIt->toRight();
        }
    }
}

// note: takes smart pointer as it maybe added to the results
static hal_size_t mapDown(MappedSegmentPtr mappedSeg, hal_size_t childIndex, list<MappedSegmentPtr> &results,
                          hal_size_t minLength) {
    const Genome *child = mappedSeg->getGenome()->getChild(childIndex);
    assert(child != NULL);
    hal_size_t added = 0;
    if (mappedSeg->isTop() == false) {
        TopSegmentIteratorPtr topSegIt = child->getTopSegmentIterator();
        SegmentIteratorPtr targetSegIt = mappedSeg->getTargetIteratorPtr();
        BottomSegmentIteratorPtr botSegIt = std::dynamic_pointer_cast<BottomSegmentIterator>(targetSegIt);

        if (botSegIt->bseg()->hasChild(childIndex) == true && botSegIt->getLength() >= minLength) {
            topSegIt->toChild(botSegIt, childIndex);
            mappedSeg->setTarget(std::dynamic_pointer_cast<SegmentIterator>(topSegIt));
            results.push_back(MappedSegmentPtr(mappedSeg));
            ++added;
        }
    } else {
        hal_index_t rightCutoff = mappedSeg->getEndPosition();
        TopSegmentIteratorPtr topSegIt = mappedSeg->targetAsTop();
        hal_index_t startOffset = (hal_index_t)topSegIt->getStartOffset();
        hal_index_t endOffset = (hal_index_t)topSegIt->getEndOffset();
        BottomSegmentIteratorPtr botSegIt = mappedSeg->getGenome()->getBottomSegmentIterator();
        botSegIt->toParseDown(topSegIt);
        do {
            BottomSegmentIteratorPtr newBotSegIt = botSegIt->clone();

            // we map the new target back to see how the offsets have
            // changed.  these changes are then applied to the source segment
            // as deltas
            TopSegmentIteratorPtr backTopSegIt = topSegIt->clone();
            backTopSegIt->toParseUp(newBotSegIt);
            hal_index_t startBack = (hal_index_t)backTopSegIt->getStartOffset();
            hal_index_t endBack = (hal_index_t)backTopSegIt->getEndOffset();
            assert(startBack >= startOffset);
            assert(endBack >= endOffset);
            SegmentIteratorPtr newSourceSegIt = mappedSeg->sourceClone();
            hal_index_t startDelta = startBack - startOffset;
            hal_index_t endDelta = endBack - endOffset;
            assert((hal_index_t)newSourceSegIt->getLength() > startDelta + endDelta);
            newSourceSegIt->slice(newSourceSegIt->getStartOffset() + startDelta, newSourceSegIt->getEndOffset() + endDelta);

            MappedSegmentPtr newMappedSeg(new MappedSegment(newSourceSegIt, newBotSegIt));

            assert(newMappedSeg->isTop() == false);
            assert(newMappedSeg->getSource()->getGenome() == mappedSeg->getSource()->getGenome());

            added += mapDown(newMappedSeg, childIndex, results, minLength);

            // stupid that we have to make this check but odn't want to
            // make fundamental api change now
            if (botSegIt->getEndPosition() != rightCutoff) {
                botSegIt->toRight(rightCutoff);
            } else {
                break;
            }
        } while (true);
    }
    return added;
}

void Genome::fixParseInfo()
{
  if (getParent() == NULL || getNumChildren() == 0)
  {
    return;
  }
  
  // copied from CactusHalConverter::updateRootParseInfo() in
  // cactus2hal/src/cactusHalConverter.cpp 
  BottomSegmentIteratorPtr bottomIterator = 
    getBottomSegmentIterator();
  TopSegmentIteratorPtr topIterator = getTopSegmentIterator();
  BottomSegmentIteratorConstPtr bend = getBottomSegmentEndIterator();
  TopSegmentIteratorConstPtr tend = getTopSegmentEndIterator();
  int top = 0, bot = 0;
  while (bottomIterator != bend && topIterator != tend)
  {
    bool bright = false;
    bool tright = false;
    BottomSegment* bseg = bottomIterator->getBottomSegment();
    TopSegment* tseg = topIterator->getTopSegment();
    hal_index_t bstart = bseg->getStartPosition();
    hal_index_t bendidx = bstart + (hal_index_t)bseg->getLength();
    hal_index_t tstart = tseg->getStartPosition();
    hal_index_t tendidx = tstart + (hal_index_t)tseg->getLength();

    if (bstart >= tstart && bstart < tendidx)
    {
      bseg->setTopParseIndex(tseg->getArrayIndex());
    }
    if (bendidx <= tendidx || bstart == bendidx)
    {
      bright = true;
    }
        
    if (tstart >= bstart && tstart < bendidx)
    {
      tseg->setBottomParseIndex(bseg->getArrayIndex());
    }
    if (tendidx <= bendidx || tstart == tendidx)
    {
      tright = true;
    }

    assert(bright || tright);
    if (bright == true)
    {
      bot += 1;
      bottomIterator->toRight();
    }
    if (tright == true)
    {
      top += 1;
      topIterator->toRight();
    }
  }
}

void MappedSegmentMapDupeTest::createCallBack(AlignmentPtr alignment)
{
  vector<Sequence::Info> seqVec(1);
  
  BottomSegmentIteratorPtr bi;
  BottomSegmentStruct bs;
  TopSegmentIteratorPtr ti;
  TopSegmentStruct ts;
  
  // setup simple case were there is an edge from a parent to 
  // child and it is reversed
  Genome* parent = alignment->addRootGenome("parent");
  Genome* child1 = alignment->addLeafGenome("child1", "parent", 1);
  Genome* child2 = alignment->addLeafGenome("child2", "parent", 1);
  seqVec[0] = Sequence::Info("Sequence", 3, 0, 1);
  parent->setDimensions(seqVec);
  seqVec[0] = Sequence::Info("Sequence", 9, 3, 0);
  child1->setDimensions(seqVec);
  seqVec[0] = Sequence::Info("Sequence", 9, 3, 0);
  child2->setDimensions(seqVec);

  parent->setString("CCC");
  child1->setString("CCCTACGTG");
  child2->setString("CCCTACGTG");

  bi = parent->getBottomSegmentIterator();
  bs.set(0, 3);
  bs._children.push_back(pair<hal_size_t, bool>(0, true));
  bs._children.push_back(pair<hal_size_t, bool>(0, false));
  bs.applyTo(bi);
     
  ti = child1->getTopSegmentIterator();
  ts.set(0, 3, 0, true, NULL_INDEX, 1);
  ts.applyTo(ti);
  ti->toRight();
  ts.set(3, 3, 0, true, NULL_INDEX, 2);
  ts.applyTo(ti);
  ti->toRight();
  ts.set(6, 3, 0, true, NULL_INDEX, 0);
  ts.applyTo(ti);

  ti = child2->getTopSegmentIterator();
  ts.set(0, 3, 0, false);
  ts.applyTo(ti);
  ti->toRight();
  ts.set(3, 3, NULL_INDEX, true);
  ts.applyTo(ti);
  ti->toRight();
  ts.set(6, 3, NULL_INDEX, false);
  ts.applyTo(ti);
}

// note: takes smart pointer as it maybe added to the results
static hal_size_t mapUp(MappedSegmentPtr mappedSeg, list<MappedSegmentPtr> &results, bool doDupes, hal_size_t minLength) {
    const Genome *parent = mappedSeg->getGenome()->getParent();
    assert(parent != NULL);
    hal_size_t added = 0;
    if (mappedSeg->isTop() == true) {
        BottomSegmentIteratorPtr botSegIt = parent->getBottomSegmentIterator();
        TopSegmentIteratorPtr topSegIt = mappedSeg->targetAsTop();
        if (topSegIt->tseg()->hasParent() == true && topSegIt->getLength() >= minLength &&
            (doDupes == true || topSegIt->tseg()->isCanonicalParalog() == true)) {
            botSegIt->toParent(topSegIt);
            mappedSeg->setTarget(std::dynamic_pointer_cast<SegmentIterator>(botSegIt));
            results.push_back(mappedSeg);
            ++added;
        }
    } else {
        hal_index_t rightCutoff = mappedSeg->getEndPosition();
        BottomSegmentIteratorPtr botSegIt = mappedSeg->targetAsBottom();
        hal_index_t startOffset = (hal_index_t)botSegIt->getStartOffset();
        hal_index_t endOffset = (hal_index_t)botSegIt->getEndOffset();
        TopSegmentIteratorPtr topSegIt = mappedSeg->getGenome()->getTopSegmentIterator();
        topSegIt->toParseUp(botSegIt);
        do {
            TopSegmentIteratorPtr newTopSegIt = topSegIt->clone();

            // we map the new target back to see how the offsets have
            // changed.  these changes are then applied to the source segment
            // as deltas
            BottomSegmentIteratorPtr backBotSegIt = botSegIt->clone();
            backBotSegIt->toParseDown(newTopSegIt);
            hal_index_t startBack = (hal_index_t)backBotSegIt->getStartOffset();
            hal_index_t endBack = (hal_index_t)backBotSegIt->getEndOffset();
            assert(startBack >= startOffset);
            assert(endBack >= endOffset);
            SegmentIteratorPtr newSourceSegIt = mappedSeg->sourceClone();
            hal_index_t startDelta = startBack - startOffset;
            hal_index_t endDelta = endBack - endOffset;
            assert((hal_index_t)newSourceSegIt->getLength() > startDelta + endDelta);
            newSourceSegIt->slice(newSourceSegIt->getStartOffset() + startDelta, newSourceSegIt->getEndOffset() + endDelta);

            MappedSegmentPtr newMappedSeg(new MappedSegment(newSourceSegIt, newTopSegIt));

            assert(newMappedSeg->isTop() == true);
            assert(newMappedSeg->getSource()->getGenome() == mappedSeg->getSource()->getGenome());

            added += mapUp(newMappedSeg, results, doDupes, minLength);
            // stupid that we have to make this check but odn't want to
            // make fundamental api change now
            if (topSegIt->getEndPosition() != rightCutoff) {
                topSegIt->toRight(rightCutoff);
            } else {
                break;
            }
        } while (true);
    }
    return added;
}

void TopSegmentIteratorToSiteTest::checkGenome(const Genome *genome) {
    TopSegmentIteratorPtr ti = genome->getTopSegmentIterator();
    for (hal_index_t pos = 0; pos < (hal_index_t)genome->getSequenceLength(); ++pos) {
        ti->toSite(pos);
        CuAssertTrue(_testCase, ti->getStartPosition() == pos);
        CuAssertTrue(_testCase, ti->getLength() == 1);
        ti->toSite(pos, false);
        CuAssertTrue(_testCase, pos >= ti->getStartPosition() && pos < ti->getStartPosition() + (hal_index_t)ti->getLength());
        CuAssertTrue(_testCase, ti->getLength() == ti->getTopSegment()->getLength());
    }
}

// Set top segments to be equal width and so that segment 1, 2, 3,
// etc. corresponds to parent segment 1, 2, 3, etc.
void setTopSegments(Genome *genome, hal_size_t width) {
    TopSegmentIteratorPtr topIt = genome->getTopSegmentIterator();
    hal_size_t n = genome->getNumTopSegments();
    hal_index_t startPos = 0;
    for (; topIt->getArrayIndex() < n; topIt->toRight(), startPos += width) {
        topIt->setCoordinates(startPos, width);
        topIt->tseg()->setParentIndex(topIt->getArrayIndex());
        topIt->tseg()->setParentReversed(false);
        topIt->tseg()->setBottomParseIndex(NULL_INDEX);
        topIt->tseg()->setNextParalogyIndex(NULL_INDEX);
    }
}

void TopSegmentStruct::applyTo(TopSegmentIteratorPtr it) const {
    TopSegment *seg = it->getTopSegment();
    seg->setCoordinates(_startPosition, _length);
    seg->setNextParalogyIndex(_nextParalogyIndex);
    seg->setParentIndex(_parentIndex);
    seg->setParentReversed(_parentReversed);
    seg->setBottomParseIndex(_bottomParseIndex);
}

void TopSegmentStruct::compareTo(TopSegmentIteratorPtr it, CuTest *testCase) const {
    const TopSegment *seg = it->getTopSegment();
    CuAssertTrue(testCase, _length == seg->getLength());
    CuAssertTrue(testCase, _startPosition == seg->getStartPosition());
    CuAssertTrue(testCase, _nextParalogyIndex == seg->getNextParalogyIndex());
    CuAssertTrue(testCase, _parentIndex == seg->getParentIndex());
    CuAssertTrue(testCase, _bottomParseIndex == seg->getBottomParseIndex());
}

void Genome::copyTopSegments(Genome *dest) const
{
  const Genome *inParent = getParent();
  const Genome *outParent = dest->getParent();

  TopSegmentIteratorConstPtr inTop = getTopSegmentIterator();
  TopSegmentIteratorPtr outTop = dest->getTopSegmentIterator();
  hal_size_t n = dest->getNumTopSegments();
  assert(n == 0 || n == getNumTopSegments());

  if (n == 0) {
    // Nothing to do if there are no top segments.
    return;
  }

  BottomSegmentIteratorConstPtr inParentBottomSegIt = inParent->getBottomSegmentIterator();
  BottomSegmentIteratorConstPtr outParentBottomSegIt = outParent->getBottomSegmentIterator();

  for (; (hal_size_t)inTop->getArrayIndex() < n; inTop->toRight(),
         outTop->toRight())
  {
    hal_index_t genomePos = inTop->getStartPosition();
    assert(genomePos != NULL_INDEX);
    string inSeqName = getSequenceBySite(genomePos)->getName();
    string outSeqName = dest->getSequenceBySite(genomePos)->getName();
    // if (inSeqName != outSeqName) {
    //   stringstream ss;
    //   ss << "When copying top segments from " << getName() << " to " << dest->getName() << ": sequence " << inSeqName << " != " << outSeqName << " at site " << genomePos;
    //   throw hal_exception(ss.str());
    // }

    outTop->setCoordinates(inTop->getStartPosition(), inTop->getLength());
    outTop->setParentIndex(inTop->getParentIndex());
    outTop->setParentReversed(inTop->getParentReversed());
    outTop->setBottomParseIndex(inTop->getBottomParseIndex());
    outTop->setNextParalogyIndex(inTop->getNextParalogyIndex());

    // Check that the sequences from the bottom segments we point to are the same. If not, correct the indices so that they are.
    if (inTop->getParentIndex() != NULL_INDEX) {
      inParentBottomSegIt->toParent(inTop);

      const Sequence *inParentSequence = inParentBottomSegIt->getSequence();

      const Sequence *outParentSequence = outParent->getSequence(inParentSequence->getName());

      hal_index_t inParentSegmentOffset = inTop->getParentIndex() - inParentSequence->getBottomSegmentArrayIndex();
      hal_index_t outParentSegmentIndex = inParentSegmentOffset + outParentSequence->getBottomSegmentArrayIndex();

      outTop->setParentIndex(outParentSegmentIndex);
    }
  }
}

void TopSegmentIteratorToSiteTest::createCallBack(AlignmentPtr alignment)
{
  vector<Sequence::Info> seqVec(1);
  
  TopSegmentIteratorPtr ti;
  TopSegmentStruct ts;
  
  // case 1: single segment
  Genome* case1 = alignment->addRootGenome("case1");
  seqVec[0] = Sequence::Info("Sequence", 10, 2, 0);
  case1->setDimensions(seqVec);
  ti = case1->getTopSegmentIterator();
  ts.set(0, 9);
  ts.applyTo(ti);
  ti->toRight();
  ts.set(9, 1);
  ts.applyTo(ti);
  case1 = NULL;

  // case 2: bunch of random segments
  const hal_size_t numSegs = 1133;
  hal_size_t total = 0;
  vector<hal_size_t> segLens(numSegs);
  for (size_t i = 0 ; i < numSegs; ++i)
  {
    hal_size_t len = rand() % 77 + 1;
    segLens[i] = len;
    total += len;
    assert(len > 0);
  }
  Genome* case2 = alignment->addRootGenome("case2");
  seqVec[0] = Sequence::Info("Sequence", total, numSegs, 0);
  case2->setDimensions(seqVec);
  hal_index_t prev = 0;
  for (size_t i = 0 ; i < numSegs; ++i)
  {
    ti = case2->getTopSegmentIterator((hal_index_t)i);
    ts.set(prev, segLens[i]);
    prev += segLens[i];
    ts.applyTo(ti);
  }
}

void GappedBottomSegmentIterator::toRightNextUngapped(TopSegmentIteratorPtr topSeqIt) const {
    while (topSeqIt->tseg()->hasParent() == false && topSeqIt->getLength() <= _gapThreshold) {
        if ((!topSeqIt->getReversed() && topSeqIt->getTopSegment()->isLast()) ||
            (topSeqIt->getReversed() && topSeqIt->getTopSegment()->isFirst())) {
            break;
        }
        topSeqIt->toRight();
    }
}