1 files changed, 52 insertions, 0 deletions
diff --git a/dtrain/ksampler.h b/dtrain/ksampler.h
new file mode 100644
index 00000000..0783f98b
--- /dev/null
+++ b/dtrain/ksampler.h
@@ -0,0 +1,52 @@
+#ifndef _DTRAIN_KSAMPLER_H_
+#define _DTRAIN_KSAMPLER_H_
+
+#include "hg_sampler.h" // cdec
+#include "kbestget.h"
+#include "score.h"
+
+namespace dtrain
+{
+
+
+struct KSampler : public HypSampler
+{
+  const unsigned k_;
+  vector<ScoredHyp> s_;
+  MT19937* prng_;
+  score_t (*scorer)(NgramCounts&, const unsigned, const unsigned, unsigned, vector<score_t>);
+  unsigned src_len_;
+
+  explicit KSampler(const unsigned k, MT19937* prng) :
+    k_(k), prng_(prng) {}
+
+  virtual void
+  NotifyTranslationForest(const SentenceMetadata& smeta, Hypergraph* hg)
+  {
+    src_len_ = smeta.GetSourceLength();
+    ScoredSamples(*hg);
+  }
+
+  vector<ScoredHyp>* GetSamples() { return &s_; }
+
+  void ScoredSamples(const Hypergraph& forest) {
+    s_.clear();
+    std::vector<HypergraphSampler::Hypothesis> samples;
+    HypergraphSampler::sample_hypotheses(forest, k_, prng_, &samples);
+    for (unsigned i = 0; i < k_; ++i) {
+      ScoredHyp h;
+      h.w = samples[i].words;
+      h.f = samples[i].fmap;
+      h.model = log(samples[i].model_score);
+      h.rank = i;
+      h.score = scorer_->Score(h.w, *ref_, i, src_len_);
+      s_.push_back(h);
+    }
+  }
+};
+
+
+} // namespace
+
+#endif
+