From 0172721855098ca02b207231a654dffa5e4eb1c9 Mon Sep 17 00:00:00 2001
From: redpony <redpony@ec762483-ff6d-05da-a07a-a48fb63a330f>
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
---
 extools/sentence_pair.h | 39 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 39 insertions(+)
 create mode 100644 extools/sentence_pair.h

(limited to 'extools/sentence_pair.h')

diff --git a/extools/sentence_pair.h b/extools/sentence_pair.h
new file mode 100644
index 00000000..d78be359
--- /dev/null
+++ b/extools/sentence_pair.h
@@ -0,0 +1,39 @@
+#ifndef _SENTENCE_PAIR_H_
+#define _SENTENCE_PAIR_H_
+
+#include <utility>
+#include <vector>
+#include "wordid.h"
+#include "array2d.h"
+
+// represents a parallel sentence with a word alignment and category
+// annotations over subspans (currently in terms of f)
+// you should read one using ParseInputLine and then use the public
+// member variables to query things about it
+struct AnnotatedParallelSentence {
+  // read annotated parallel sentence from string
+  void ParseInputLine(const char* buf);
+
+  std::vector<WordID> f, e;  // words in f and e
+
+  // word alignment information
+  std::vector<int> e_aligned, f_aligned; // counts the number of times column/row x is aligned
+  Array2D<bool> aligned;
+  std::vector<std::vector<std::pair<short, short> > > aligns_by_fword;
+
+  // span type information
+  Array2D<std::vector<WordID> > span_types;  // span_types(i,j) is the list of category
+                               // types for a span (i,j) in the TARGET language.
+
+  int f_len, e_len;
+
+  static int ReadAlignmentPoint(const char* buf, int start, int end, bool permit_col, short* a, short* b);
+
+ private:
+  void Reset();
+  void AllocateForAlignment();
+  void ParseAlignmentPoint(const char* buf, int start, int end);
+  void ParseSpanLabel(const char* buf, int start, int end);
+};
+
+#endif
-- 
cgit v1.2.3