add non-const iterators to sparse vector, speed up model1 code

5 files changed, 119 insertions, 3 deletions

diff --git a/utils/ccrp_onetable.h b/utils/ccrp_onetable.h
index 1fe01b0e..abe399ea 100644
--- a/utils/ccrp_onetable.h
+++ b/utils/ccrp_onetable.h
@@ -183,7 +183,7 @@ class CCRP_OneTable {
     assert(has_discount_prior() || has_alpha_prior());
     DiscountResampler dr(*this);
     ConcentrationResampler cr(*this);
-    for (int iter = 0; iter < nloop; ++iter) {
+    for (unsigned iter = 0; iter < nloop; ++iter) {
       if (has_alpha_prior()) {
         alpha_ = slice_sampler1d(cr, alpha_, *rng, 0.0,
                                std::numeric_limits<double>::infinity(), 0.0, niterations, 100*niterations);
diff --git a/utils/corpus_tools.cc b/utils/corpus_tools.cc
index d17785af..191153a2 100644
--- a/utils/corpus_tools.cc
+++ b/utils/corpus_tools.cc
@@ -8,6 +8,26 @@
 
 using namespace std;
 
+void CorpusTools::ReadLine(const string& line,
+                           vector<WordID>* src,
+                           vector<WordID>* trg) {
+  static const WordID kDIV = TD::Convert("|||");
+  static vector<WordID> tmp;
+  src->clear();
+  trg->clear();
+  TD::ConvertSentence(line, &tmp);
+  unsigned i = 0;
+  while(i < tmp.size() && tmp[i] != kDIV) {
+    src->push_back(tmp[i]);
+    ++i;
+  }
+  if (i < tmp.size() && tmp[i] == kDIV) {
+    ++i;
+    for (; i < tmp.size() ; ++i)
+      trg->push_back(tmp[i]);
+  }
+}
+
 void CorpusTools::ReadFromFile(const string& filename,
                            vector<vector<WordID> >* src,
                            set<WordID>* src_vocab,
diff --git a/utils/corpus_tools.h b/utils/corpus_tools.h
index 97bdaa94..f6699d87 100644
--- a/utils/corpus_tools.h
+++ b/utils/corpus_tools.h
@@ -7,6 +7,10 @@
 #include "wordid.h"
 
 struct CorpusTools {
+  static void ReadLine(const std::string& line,
+                       std::vector<WordID>* src,
+                       std::vector<WordID>* trg);
+
   static void ReadFromFile(const std::string& filename,
                            std::vector<std::vector<WordID> >* src,
                            std::set<WordID>* src_vocab = NULL,
diff --git a/utils/fast_sparse_vector.h b/utils/fast_sparse_vector.h
index e86cbdc1..6e5dfb14 100644
--- a/utils/fast_sparse_vector.h
+++ b/utils/fast_sparse_vector.h
@@ -66,6 +66,60 @@ BOOST_STATIC_ASSERT(sizeof(PairIntT<float>) == sizeof(std::pair<unsigned,float>)
 template <typename T, unsigned LOCAL_MAX = (sizeof(T) == sizeof(float) ? 15u : 7u)>
 class FastSparseVector {
  public:
+  struct iterator {
+    iterator(FastSparseVector<T>& v, const bool is_end) : local_(!v.is_remote_) {
+      if (local_) {
+        local_it_ = &v.data_.local[is_end ? v.local_size_ : 0];
+      } else {
+        if (is_end)
+          remote_it_ = v.data_.rbmap->end();
+        else
+          remote_it_ = v.data_.rbmap->begin();
+      }
+    }
+    iterator(FastSparseVector<T>& v, const bool, const unsigned k) : local_(!v.is_remote_) {
+      if (local_) {
+        unsigned i = 0;
+        while(i < v.local_size_ && v.data_.local[i].first() != k) { ++i; }
+        local_it_ = &v.data_.local[i];
+      } else {
+        remote_it_ = v.data_.rbmap->find(k);
+      }
+    }
+    const bool local_;
+    PairIntT<T>* local_it_;
+    typename std::map<unsigned, T>::iterator remote_it_;
+    std::pair<const unsigned, T>& operator*() const {
+      if (local_)
+        return *reinterpret_cast<std::pair<const unsigned, T>*>(local_it_);
+      else
+        return *remote_it_;
+    }
+
+    std::pair<const unsigned, T>* operator->() const {
+      if (local_)
+        return reinterpret_cast<std::pair<const unsigned, T>*>(local_it_);
+      else
+        return &*remote_it_;
+    }
+
+    iterator& operator++() {
+      if (local_) ++local_it_; else ++remote_it_;
+      return *this;
+    }
+
+    inline bool operator==(const iterator& o) const {
+      if (o.local_ != local_) return false;
+      if (local_) {
+        return local_it_ == o.local_it_;
+      } else {
+        return remote_it_ == o.remote_it_;
+      }
+    }
+    inline bool operator!=(const iterator& o) const {
+      return !(o == *this);
+    }
+  };
   struct const_iterator {
     const_iterator(const FastSparseVector<T>& v, const bool is_end) : local_(!v.is_remote_) {
       if (local_) {
@@ -77,12 +131,21 @@ class FastSparseVector {
           remote_it_ = v.data_.rbmap->begin();
       }
     }
+    const_iterator(const FastSparseVector<T>& v, const bool, const unsigned k) : local_(!v.is_remote_) {
+      if (local_) {
+        unsigned i = 0;
+        while(i < v.local_size_ && v.data_.local[i].first() != k) { ++i; }
+        local_it_ = &v.data_.local[i];
+      } else {
+        remote_it_ = v.data_.rbmap->find(k);
+      }
+    }
     const bool local_;
     const PairIntT<T>* local_it_;
     typename std::map<unsigned, T>::const_iterator remote_it_;
     const std::pair<const unsigned, T>& operator*() const {
       if (local_)
-        return *reinterpret_cast<const std::pair<const unsigned, float>*>(local_it_);
+        return *reinterpret_cast<const std::pair<const unsigned, T>*>(local_it_);
       else
         return *remote_it_;
     }
@@ -160,6 +223,9 @@ class FastSparseVector {
   bool nonzero(unsigned k) const {
     return static_cast<bool>(value(k));
   }
+  inline T& operator[](unsigned k) {
+    return get_or_create_bin(k);
+  }
   inline void set_value(unsigned k, const T& v) {
     get_or_create_bin(k) = v;
   }
@@ -283,6 +349,18 @@ class FastSparseVector {
     }
     return o;
   }
+  iterator find(unsigned k) {
+    return iterator(*this, false, k);
+  }
+  iterator begin() {
+    return iterator(*this, false);
+  }
+  iterator end() {
+    return iterator(*this, true);
+  }
+  const_iterator find(unsigned k) const {
+    return const_iterator(*this, false, k);
+  }
   const_iterator begin() const {
     return const_iterator(*this, false);
   }
diff --git a/utils/sampler.h b/utils/sampler.h
index b237c716..3e4a4086 100644
--- a/utils/sampler.h
+++ b/utils/sampler.h
@@ -12,6 +12,7 @@
 #include <boost/random/mersenne_twister.hpp>
 #include <boost/random/uniform_real.hpp>
 #include <boost/random/variate_generator.hpp>
+#include <boost/random/gamma_distribution.hpp>
 #include <boost/random/normal_distribution.hpp>
 #include <boost/random/poisson_distribution.hpp>
 #include <boost/random/uniform_int.hpp>
@@ -76,6 +77,18 @@ struct RandomNumberGenerator {
     return boost::poisson_distribution<int>(lambda)(m_random);
   }
 
+  double NextGamma(double shape, double scale = 1.0) {
+    boost::gamma_distribution<> gamma(shape);
+    boost::variate_generator<boost::mt19937&,boost::gamma_distribution<> > vg(m_generator, gamma);
+    return vg() * scale;
+  }
+
+  double NextBeta(double alpha, double beta) {
+    double x = NextGamma(alpha);
+    double y = NextGamma(beta);
+    return x / (x + y);
+  }
+
   bool AcceptMetropolisHastings(const prob_t& p_cur,
                                 const prob_t& p_prev,
                                 const prob_t& q_cur,
@@ -123,11 +136,12 @@ size_t RandomNumberGenerator<RNG>::SelectSample(const SampleSet<F>& ss, double T
   const bool anneal = (T != 1.0);
   F sum = F(0);
   if (anneal) {
-    for (int i = 0; i < ss.m_scores.size(); ++i)
+    for (unsigned i = 0; i < ss.m_scores.size(); ++i)
       sum += pow(ss.m_scores[i], annealing_factor);  // p^(1/T)
   } else {
     sum = std::accumulate(ss.m_scores.begin(), ss.m_scores.end(), F(0));
   }
+  //std::cerr << "SUM: " << sum << std::endl;
   //for (size_t i = 0; i < ss.m_scores.size(); ++i) std::cerr << ss.m_scores[i] << ",";
   //std::cerr << std::endl;