From 0172721855098ca02b207231a654dffa5e4eb1c9 Mon Sep 17 00:00:00 2001 From: redpony Date: Tue, 22 Jun 2010 05:12:27 +0000 Subject: initial checkin git-svn-id: https://ws10smt.googlecode.com/svn/trunk@2 ec762483-ff6d-05da-a07a-a48fb63a330f --- decoder/hg_io.h | 39 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 39 insertions(+) create mode 100644 decoder/hg_io.h (limited to 'decoder/hg_io.h') diff --git a/decoder/hg_io.h b/decoder/hg_io.h new file mode 100644 index 00000000..7162106e --- /dev/null +++ b/decoder/hg_io.h @@ -0,0 +1,39 @@ +#ifndef _HG_IO_H_ +#define _HG_IO_H_ + +#include + +#include "lattice.h" +class Hypergraph; + +struct HypergraphIO { + + // the format is basically a list of nodes and edges in topological order + // any edge you read, you must have already read its tail nodes + // any node you read, you must have already read its incoming edges + // this may make writing a bit more challenging if your forest is not + // topologically sorted (but that probably doesn't happen very often), + // but it makes reading much more memory efficient. + // see test_data/small.json.gz for an email encoding + static bool ReadFromJSON(std::istream* in, Hypergraph* out); + + // if remove_rules is used, the hypergraph is serialized without rule information + // (so it only contains structure and feature information) + static bool WriteToJSON(const Hypergraph& hg, bool remove_rules, std::ostream* out); + + static void WriteAsCFG(const Hypergraph& hg); + + // serialization utils + static void ReadFromPLF(const std::string& in, Hypergraph* out, int line = 0); + // return PLF string representation (undefined behavior on non-lattices) + static std::string AsPLF(const Hypergraph& hg, bool include_global_parentheses = true); + static void PLFtoLattice(const std::string& plf, Lattice* pl); + static std::string Escape(const std::string& s); // PLF helper +}; + +namespace B64 { + bool b64decode(const unsigned char* data, const size_t insize, char* out, const size_t outsize); + void b64encode(const char* data, const size_t size, std::ostream* out); +} + +#endif -- cgit v1.2.3