Use vectors with small buffer optimization to represent intermediate …

…lists of gene indexes

Makefile

@@ -14,7 +14,7 @@ shark: main.o
 	@echo '* Compiling $<'
 	$(CXX) $(CXXFLAGS) -o $@ -c $<
 
-main.o: common.hpp argument_parser.hpp bloomfilter.h BloomfilterFiller.hpp KmerBuilder.hpp FastaSplitter.hpp FastqSplitter.hpp ReadAnalyzer.hpp ReadOutput.hpp kmer_utils.hpp
+main.o: common.hpp argument_parser.hpp bloomfilter.h BloomfilterFiller.hpp KmerBuilder.hpp FastaSplitter.hpp FastqSplitter.hpp ReadAnalyzer.hpp ReadOutput.hpp kmer_utils.hpp small_vector.hpp
 
 clean:
 	rm -rf *.o

bloomfilter.h

@@ -28,13 +28,15 @@
 #include <string>
 
 #include "kmer_utils.hpp"
+#include "small_vector.hpp"
 
 using namespace std;
 using namespace sdsl;
 
 class KmerBuilder;
 class BloomfilterFiller;
 
+
 class BF {
   friend class KmerBuilder;
   friend class BloomfilterFiller;
@@ -45,7 +47,7 @@ class BF {
   typedef uint64_t hash_t;
   typedef bit_vector bit_vector_t;
   typedef bit_vector_t::rank_1_type rank_t;
-  typedef vector<int> index_t;
+  typedef small_vector_t index_t;
   typedef vector<index_t> set_index_t;
   typedef vector<uint16_t> index_kmer_t;
   typedef bit_vector_t::select_1_type select_t;
@@ -88,7 +90,7 @@ class BF {
     for (const auto bf_idx: kmers) {
       int kmer_rank = _brank(bf_idx);
       const auto size = _set_index[kmer_rank].size();
-      if (size == 0 || _set_index[kmer_rank][size-1] != input_idx)
+      if (size == 0 || _set_index[kmer_rank].last() != input_idx)
         _set_index[kmer_rank].push_back(input_idx);
     }
   }

small_vector.hpp

@@ -0,0 +1,91 @@
+/**
+ * shark - Mapping-free filtering of useless RNA-Seq reads
+ * Copyright (C) 2020 Tamara Ceccato, Luca Denti, Yuri Pirola, Marco Previtali
+ *
+ * This file is part of shark.
+ *
+ * shark is free software: you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by
+ * the Free Software Foundation, either version 3 of the License, or
+ * (at your option) any later version.
+ *
+ * shark is distributed in the hope that it will be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+ * General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with shark; see the file LICENSE. If not, see
+ * <https://www.gnu.org/licenses/>.
+ **/
+
+#include <cstdint>
+#include <vector>
+
+struct small_vector_t {
+  union {
+    struct {
+      uint8_t flag;
+      uint8_t size;
+      uint16_t arr[3];
+    } s;
+    std::vector<uint16_t>* l;
+  } v;
+
+  small_vector_t() {
+    v.s.flag = 1u;
+    v.s.size = 0;
+  }
+
+  ~small_vector_t() {
+    if ((v.s.flag & 0x1) == 0) {
+      delete v.l;
+    }
+  }
+
+  void push_back(uint16_t x) {
+    if ((v.s.flag & 0x1) != 0) {
+      if (v.s.size < 3) v.s.arr[v.s.size++] = x;
+      else {
+        std::vector<uint16_t>* ptr = new std::vector<uint16_t>(v.s.arr, v.s.arr + 3);
+        ptr->push_back(x);
+        v.l = ptr;
+      }
+    } else {
+      v.l->push_back(x);
+    }
+  }
+
+  size_t size() const {
+    if ((v.s.flag & 0x1) != 0) {
+      return v.s.size;
+    } else {
+      return v.l->size();
+    }
+  }
+
+  uint16_t last() const {
+    if ((v.s.flag & 0x1) != 0) {
+      return v.s.arr[v.s.size - 1];
+    } else {
+      return v.l->back();
+    }
+  }
+
+  const uint16_t* begin() const {
+    if ((v.s.flag & 0x1) != 0) {
+      return v.s.arr;
+    } else {
+      return v.l->data();
+    }
+  }
+
+  const uint16_t* end() const {
+    if ((v.s.flag & 0x1) != 0) {
+      return v.s.arr + v.s.size;
+    } else {
+      return v.l->data() + v.l->size();
+    }
+  }
+
+};