Fixed issue 669

git-svn-id: https://tesseract-ocr.googlecode.com/svn/trunk@743 d0cd1f9f-072b-0410-8dd7-cf729c803f20
jimregan · Sep 21, 2012 · c7cef53 · c7cef53
1 parent d71045f
commit c7cef53
Show file tree

Hide file tree

Showing 2 changed files with 34 additions and 26 deletions.
diff --git a/classify/shapetable.cpp b/classify/shapetable.cpp
@@ -29,20 +29,17 @@
 namespace tesseract {
 
 // Writes to the given file. Returns false in case of error.
-bool UnicharAndFonts::Serialize(FILE* fp) {
-  inT32 uni_id = unichar_id;
-  if (fwrite(&uni_id, sizeof(uni_id), 1, fp) != 1) return false;
+bool UnicharAndFonts::Serialize(FILE* fp) const {
+  if (fwrite(&unichar_id, sizeof(unichar_id), 1, fp) != 1) return false;
   if (!font_ids.Serialize(fp)) return false;
   return true;
 }
 // Reads from the given file. Returns false in case of error.
 // If swap is true, assumes a big/little-endian swap is needed.
 bool UnicharAndFonts::DeSerialize(bool swap, FILE* fp) {
-  inT32 uni_id;
-  if (fread(&uni_id, sizeof(uni_id), 1, fp) != 1) return false;
+  if (fread(&unichar_id, sizeof(unichar_id), 1, fp) != 1) return false;
   if (swap)
-    ReverseN(&uni_id, sizeof(uni_id));
-  unichar_id = uni_id;
+    ReverseN(&unichar_id, sizeof(unichar_id));
   if (!font_ids.DeSerialize(swap, fp)) return false;
   return true;
 }
@@ -55,17 +52,20 @@ int UnicharAndFonts::SortByUnicharId(const void* v1, const void* v2) {
 }
 
 // Writes to the given file. Returns false in case of error.
-bool Shape::Serialize(FILE* fp) {
-  if (fwrite(&unichars_sorted_, 1, 1, fp) != 1)
+bool Shape::Serialize(FILE* fp) const {
+  uinT8 sorted = unichars_sorted_;
+  if (fwrite(&sorted, sizeof(sorted), 1, fp) != 1)
     return false;
   if (!unichars_.SerializeClasses(fp)) return false;
   return true;
 }
 // Reads from the given file. Returns false in case of error.
 // If swap is true, assumes a big/little-endian swap is needed.
 bool Shape::DeSerialize(bool swap, FILE* fp) {
-  if (fread(&unichars_sorted_, 1, 1, fp) != 1)
+  uinT8 sorted;
+  if (fread(&sorted, sizeof(sorted), 1, fp) != 1)
     return false;
+  unichars_sorted_ = sorted != 0;
   if (!unichars_.DeSerializeClasses(swap, fp)) return false;
   return true;
 }
@@ -197,14 +197,26 @@ STRING ShapeTable::DebugStr(int shape_id) const {
   const Shape& shape = GetShape(shape_id);
   STRING result;
   result.add_str_int("Shape", shape_id);
+  if (shape.size() > 100) {
+    result.add_str_int(" Num unichars=", shape.size());
+    return result;
+  }
   for (int c = 0; c < shape.size(); ++c) {
     result.add_str_int(" c_id=", shape[c].unichar_id);
     result += "=";
     result += unicharset_->id_to_unichar(shape[c].unichar_id);
-    result.add_str_int(", ", shape[c].font_ids.size());
-    result += " fonts =";
-    for (int f = 0; f < shape[c].font_ids.size(); ++f) {
-      result.add_str_int(" ", shape[c].font_ids[f]);
+    if (shape.size() < 10) {
+      result.add_str_int(", ", shape[c].font_ids.size());
+      result += " fonts =";
+      int num_fonts = shape[c].font_ids.size();
+      if (num_fonts > 10) {
+        result.add_str_int(" ", shape[c].font_ids[0]);
+        result.add_str_int(" ... ", shape[c].font_ids[num_fonts - 1]);
+      } else {
+        for (int f = 0; f < num_fonts; ++f) {
+          result.add_str_int(" ", shape[c].font_ids[f]);
+        }
+      }
     }
   }
   return result;
@@ -327,12 +339,12 @@ int ShapeTable::BuildFromShape(const Shape& shape,
 }
 
 // Returns true if the shapes are already merged.
-bool ShapeTable::AlreadyMerged(int shape_id1, int shape_id2) {
+bool ShapeTable::AlreadyMerged(int shape_id1, int shape_id2) const {
   return MasterDestinationIndex(shape_id1) == MasterDestinationIndex(shape_id2);
 }
 
 // Returns true if any shape contains multiple unichars.
-bool ShapeTable::AnyMultipleUnichars() {
+bool ShapeTable::AnyMultipleUnichars() const {
   int num_shapes = NumShapes();
   for (int s1 = 0; s1 < num_shapes; ++s1) {
     if (MasterDestinationIndex(s1) != s1) continue;
@@ -408,10 +420,6 @@ void ShapeTable::MergeShapes(int shape_id1, int shape_id2) {
   shape_table_[master_id2]->set_destination_index(master_id1);
   // Add all the shapes of master_id2 to master_id1.
   shape_table_[master_id1]->AddShape(*shape_table_[master_id2]);
-  tprintf("Merged shape %d->%d, %d->%d, now with %d unichars: %s\n",
-          shape_id1, master_id1, shape_id2, master_id2,
-          shape_table_[master_id1]->size(),
-          DebugStr(master_id1).string());
 }
 
 // Returns the destination of this shape, (if merged), taking into account

diff --git a/classify/shapetable.h b/classify/shapetable.h
@@ -41,16 +41,16 @@ struct UnicharAndFonts {
   }
 
   // Writes to the given file. Returns false in case of error.
-  bool Serialize(FILE* fp);
+  bool Serialize(FILE* fp) const;
   // Reads from the given file. Returns false in case of error.
   // If swap is true, assumes a big/little-endian swap is needed.
   bool DeSerialize(bool swap, FILE* fp);
 
   // Sort function to sort a pair of UnicharAndFonts by unichar_id.
   static int SortByUnicharId(const void* v1, const void* v2);
 
-  GenericVector<int> font_ids;
-  int unichar_id;
+  GenericVector<inT32> font_ids;
+  inT32 unichar_id;
 };
 
 // A Shape is a collection of unichar-ids and a list of fonts associated with
@@ -64,7 +64,7 @@ class Shape {
   Shape() : destination_index_(-1) {}
 
   // Writes to the given file. Returns false in case of error.
-  bool Serialize(FILE* fp);
+  bool Serialize(FILE* fp) const;
   // Reads from the given file. Returns false in case of error.
   // If swap is true, assumes a big/little-endian swap is needed.
   bool DeSerialize(bool swap, FILE* fp);
@@ -188,9 +188,9 @@ class ShapeTable {
   int BuildFromShape(const Shape& shape, const ShapeTable& master_shapes);
 
   // Returns true if the shapes are already merged.
-  bool AlreadyMerged(int shape_id1, int shape_id2);
+  bool AlreadyMerged(int shape_id1, int shape_id2) const;
   // Returns true if any shape contains multiple unichars.
-  bool AnyMultipleUnichars();
+  bool AnyMultipleUnichars() const;
   // Returns the maximum number of unichars over all shapes.
   int MaxNumUnichars() const;
   // Merges shapes with a common unichar over the [start, end) interval.