fix issue 1417

tesseract-ocr · Feb 7, 2015 · 4c7c960 · 4c7c960
1 parent 09b0c91
commit 4c7c960
Show file tree

Hide file tree

Showing 3 changed files with 20 additions and 1 deletion.
diff --git a/api/renderer.cpp b/api/renderer.cpp
@@ -114,6 +114,13 @@ bool TessTextRenderer::AddImageHandler(TessBaseAPI* api) {
   AppendString(utf8);
   delete[] utf8;
 
+  bool pageBreak = false;
+  api->GetBoolVariable("include_page_breaks", &pageBreak);
+  const char* pageSeparator = api->GetStringVariable("page_separator");
+  if(pageBreak) {
+    AppendString(pageSeparator);
+  }
+
   return true;
 }
 

diff --git a/ccmain/tesseractclass.cpp b/ccmain/tesseractclass.cpp
@@ -442,6 +442,12 @@ Tesseract::Tesseract()
                 this->params()),
     BOOL_MEMBER(preserve_interword_spaces, false,
                 "Preserve multiple interword spaces", this->params()),
+    BOOL_MEMBER(include_page_breaks, FALSE,
+                "Include page separator string in output text after each "
+                "image/page.", this->params()),
+    STRING_MEMBER(page_separator, "\f",
+                  "Page separator (default is form feed control character)",
+                  this->params()),
 
     // The following parameters were deprecated and removed from their original
     // locations. The parameters are temporarily kept here to give Tesseract

diff --git a/ccmain/tesseractclass.h b/ccmain/tesseractclass.h
@@ -1009,7 +1009,13 @@ class Tesseract : public Wordrec {
   double_VAR_H(textord_tabfind_aligned_gap_fraction, 0.75,
                "Fraction of height used as a minimum gap for aligned blobs.");
   INT_VAR_H(tessedit_parallelize, 0, "Run in parallel where possible");
-  BOOL_VAR_H(preserve_interword_spaces, false, "Preserve multiple interword spaces");
+  BOOL_VAR_H(preserve_interword_spaces, false,
+             "Preserve multiple interword spaces");
+  BOOL_VAR_H(include_page_breaks, false,
+             "Include page separator string in output text after each "
+             "image/page.");
+  STRING_VAR_H(page_separator, "\f",
+               "Page separator (default is form feed control character)");
 
   // The following parameters were deprecated and removed from their original
   // locations. The parameters are temporarily kept here to give Tesseract