From f9859ad4116733e145d7b8eb31c3cc9318ff7564 Mon Sep 17 00:00:00 2001
From: "graehl@gmail.com"
 <graehl@gmail.com@ec762483-ff6d-05da-a07a-a48fb63a330f>
Date: Mon, 2 Aug 2010 07:57:23 +0000
Subject: fake tdict names for non-ids, push viterbi cost to root in hg, store
 as feature.  type erased fsa feature via virtual interface.  made
 lexical_cast assume C locale for speed.

git-svn-id: https://ws10smt.googlecode.com/svn/trunk@465 ec762483-ff6d-05da-a07a-a48fb63a330f
---
 decoder/ff_fsa_data.h | 105 ++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 105 insertions(+)
 create mode 100755 decoder/ff_fsa_data.h

(limited to 'decoder/ff_fsa_data.h')

diff --git a/decoder/ff_fsa_data.h b/decoder/ff_fsa_data.h
new file mode 100755
index 00000000..66d2cca8
--- /dev/null
+++ b/decoder/ff_fsa_data.h
@@ -0,0 +1,105 @@
+#ifndef FF_FSA_DATA_H
+#define FF_FSA_DATA_H
+
+#include <stdint.h> //C99
+#include <sstream>
+#include "sentences.h"
+#include "feature_accum.h"
+#include "value_array.h"
+
+typedef ValueArray<uint8_t> Bytes;
+
+// stuff I see no reason to have virtual.
+struct FsaFeatureFunctionData
+{
+  FsaFeatureFunctionData(int statesz=0,Sentence const& end_sentence_phrase=Sentence()) : ssz(statesz),start(statesz),h_start(statesz),end_phrase_(end_sentence_phrase) {
+    debug_=true;
+  }
+
+  std::string name_;
+  std::string name() const {
+    return name_;
+  }
+  typedef SparseFeatureAccumulator Accum;
+  bool debug_;
+  bool debug() const { return debug_; }
+  void state_copy(void *to,void const*from) const {
+    if (ssz)
+      std::memcpy(to,from,ssz);
+  }
+  void state_zero(void *st) const { // you should call this if you don't know the state yet and want it to be hashed/compared properly
+    std::memset(st,0,ssz);
+  }
+  Features features() const {
+    return features_;
+  }
+  int n_features() const {
+    return features_.size();
+  }
+  int state_bytes() const { return ssz; } // or override this
+  void const* start_state() const {
+    return start.begin();
+  }
+  void const * heuristic_start_state() const {
+    return h_start.begin();
+  }
+  Sentence const& end_phrase() const { return end_phrase_; }
+  template <class T>
+  static inline T* state_as(void *p) { return (T*)p; }
+  template <class T>
+  static inline T const* state_as(void const* p) { return (T*)p; }
+  std::string describe_features(FeatureVector const& feats) {
+    std::ostringstream o;
+    o<<feats;
+    return o.str();
+  }
+  void print_state(std::ostream &o,void const*state) const {
+    char const* i=(char const*)state;
+    char const* e=i+ssz;
+    for (;i!=e;++i)
+      print_hex_byte(o,*i);
+  }
+
+protected:
+  Features features_;
+  int ssz; // don't forget to set this. default 0 (it may depend on params of course)
+  Bytes start,h_start; // start state and estimated-features (heuristic) start state.  set these.  default empty.
+  Sentence end_phrase_; // words appended for final traversal (final state cost is assessed using Scan) e.g. "</s>" for lm.
+  void set_state_bytes(int sb=0) {
+    if (start.size()!=sb) start.resize(sb);
+    if (h_start.size()!=sb) h_start.resize(sb);
+    ssz=sb;
+  }
+  void set_end_phrase(WordID single) {
+    end_phrase_=singleton_sentence(single);
+  }
+
+  inline void static to_state(void *state,char const* begin,char const* end) {
+    std::memcpy(state,begin,end-begin);
+  }
+  inline void static to_state(void *state,char const* begin,int n) {
+    std::memcpy(state,begin,n);
+  }
+  template <class T>
+  inline void static to_state(void *state,T const* begin,int n=1) {
+    to_state(state,(char const*)begin,n*sizeof(T));
+  }
+  template <class T>
+  inline void static to_state(void *state,T const* begin,T const* end) {
+    to_state(state,(char const*)begin,(char const*)end);
+  }
+  inline static char hexdigit(int i) {
+    int j=i-10;
+    return j>=0?'a'+j:'0'+i;
+  }
+  inline static void print_hex_byte(std::ostream &o,unsigned c) {
+    o<<hexdigit(c>>4);
+    o<<hexdigit(c&0x0f);
+  }
+  inline static void Add(Featval v,SingleFeatureAccumulator *a) {
+    a->Add(v);
+  }
+
+};
+
+#endif
-- 
cgit v1.2.3