diff options
author | Chris Dyer <cdyer@allegro.clab.cs.cmu.edu> | 2014-10-25 00:18:38 -0400 |
---|---|---|
committer | Chris Dyer <cdyer@allegro.clab.cs.cmu.edu> | 2014-10-25 00:18:38 -0400 |
commit | 13d3dd3a8dca40c3ce70c2b117964ca3bfa5f971 (patch) | |
tree | ed0a2fa0a830d3f1b825ea27da86f9950f79f704 /decoder/ff_conll.h | |
parent | 0968a3e56e5fb8d3afee2b1c4904d76160b0fb17 (diff) |
extract features from converted conll files
Diffstat (limited to 'decoder/ff_conll.h')
-rw-r--r-- | decoder/ff_conll.h | 45 |
1 files changed, 45 insertions, 0 deletions
diff --git a/decoder/ff_conll.h b/decoder/ff_conll.h new file mode 100644 index 00000000..b37356d8 --- /dev/null +++ b/decoder/ff_conll.h @@ -0,0 +1,45 @@ +#ifndef FF_CONLL_H_ +#define FF_CONLL_H_ + +#include <vector> +#include <boost/xpressive/xpressive.hpp> +#include "ff.h" + +using namespace boost::xpressive; +using namespace std; + +class CoNLLFeatures : public FeatureFunction { + public: + CoNLLFeatures(const string& param); + protected: + virtual void TraversalFeaturesImpl(const SentenceMetadata& smeta, + const HG::Edge& edge, + const vector<const void*>& ant_contexts, + SparseVector<double>* features, + SparseVector<double>* estimated_features, + void* context) const; + virtual void PrepareForInput(const SentenceMetadata& smeta); + virtual void ParseArgs(const string& in); + virtual string Escape(const string& x) const; + virtual void ReplaceMacroWithString(string& feature_instance, + bool token_vs_label, + int relative_location, + const string& actual_token) const; + virtual void ReplaceTokenMacroWithString(string& feature_instance, + int relative_location, + const string& actual_token) const; + virtual void ReplaceLabelMacroWithString(string& feature_instance, + int relative_location, + const string& actual_token) const; + virtual void Error(const string&) const; + + private: + vector<int> token_relative_locations, label_relative_locations; + string feature_template; + vector<WordID> current_input; + WordID kSOS, kEOS; + sregex macro_regex; + unsigned which_feat; +}; + +#endif |