cgatools: SuperlocusIterator.hpp Source File

00001 // Copyright 2010 Complete Genomics, Inc.
00002 //
00003 // Licensed under the Apache License, Version 2.0 (the "License"); you
00004 // may not use this file except in compliance with the License. You
00005 // may obtain a copy of the License at
00006 //
00007 // http://www.apache.org/licenses/LICENSE-2.0
00008 //
00009 // Unless required by applicable law or agreed to in writing, software
00010 // distributed under the License is distributed on an "AS IS" BASIS,
00011 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
00012 // implied. See the License for the specific language governing
00013 // permissions and limitations under the License.
00014 
00015 #ifndef CGATOOLS_VARIANTS_SUPERLOCUSITERATOR_HPP_
00016 #define CGATOOLS_VARIANTS_SUPERLOCUSITERATOR_HPP_ 1
00017 
00019 
00020 #include "cgatools/core.hpp"
00021 #include "cgatools/variants/Superlocus.hpp"
00022 #include "cgatools/variants/VariantFileIterator.hpp"
00023 
00024 #include <deque>
00025 
00026 namespace cgatools { namespace variants {
00027 
00028     class SuperlocusQueueIterator;
00029 
00043     class SuperlocusIterator : private boost::noncopyable
00044     {
00045     public:
00053         SuperlocusIterator(uint32_t extend3Mers = 4, uint32_t extendBases = 0);
00054 
00056         void setVariantFile(VariantFileIterator& iter);
00057 
00059         void setVariantFiles(VariantFileIterator& iterA, VariantFileIterator& iterB);
00060 
00062         void setVariantFiles(const std::vector<VariantFileIterator*> iters);
00063 
00066         void skipToVariant(const reference::Range& range, const std::string& alleleSeq);
00067 
00069         void seekFirst();
00070 
00072         bool eof() const
00073         {
00074             return eof_;
00075         }
00076 
00078         const Superlocus& operator*() const
00079         {
00080             return sl_;
00081         }
00082 
00084         const Superlocus* operator->() const
00085         {
00086             return &sl_;
00087         }
00088 
00090         SuperlocusIterator& operator++()
00091         {
00092             next();
00093             return *this;
00094         }
00095 
00096     private:
00097         void next();
00098         bool findVariant(
00099             const reference::Location& loc, reference::Range& range, reference::Range& searchRange);
00100         void retireLoci();
00101         void extendVariant();
00102         void extendVariant(reference::Range& slRange, reference::Range& searchRange);
00103 
00104         reference::Range extendSearchRange(
00105             const reference::Range& range, const Locus& locus) const;
00106         uint32_t extendLeftBySuffixMatching(const Locus& locus) const;
00107         uint32_t extendRightByPrefixMatching(const Locus& locus) const;
00108         uint32_t extendLeftBySuffixMatching(
00109             const reference::Location& loc, const std::string& sequence) const;
00110         uint32_t extendRightByPrefixMatching(
00111             const reference::Location& loc, const std::string& sequence) const;
00112         size_t findContig(const reference::Location& loc) const;
00113 
00114         Superlocus sl_;
00115         std::vector<VariantFileIterator*> iters_;
00116         std::vector< std::deque<Locus> >& queues_;
00117         std::vector< boost::shared_ptr<SuperlocusQueueIterator> > qIters_;
00118         bool started_;
00119         bool eof_;
00120         uint32_t extend3Mers_;
00121         uint32_t extendBases_;
00122         uint32_t padBases_;
00123         reference::Location queueCutoff_;
00124         reference::Location slRetired_;
00125         const reference::CrrFile* crr_;
00126         std::vector<reference::Range> contigs_;
00127     };
00128 
00129 } } // cgatools::variants
00130 
00131 #endif // CGATOOLS_VARIANTS_SUPERLOCUSITERATOR_HPP_