Add font info to hocr output - fix issue 1219

git-svn-id: https://tesseract-ocr.googlecode.com/svn/trunk@1132 d0cd1f9f-072b-0410-8dd7-cf729c803f20
allen8807 · Aug 3, 2014 · 1156098 · 1156098
1 parent 19ddc89
commit 1156098
Show file tree

Hide file tree

Showing 8 changed files with 45 additions and 10 deletions.
diff --git a/api/baseapi.cpp b/api/baseapi.cpp
@@ -1356,6 +1356,8 @@ char* TessBaseAPI::GetHOCRText(int page_number) {
 
   int lcnt = 1, bcnt = 1, pcnt = 1, wcnt = 1;
   int page_id = page_number + 1;  // hOCR uses 1-based page numbers.
+  bool font_info = false;
+  GetBoolVariable("hocr_font_info", &font_info);
 
   STRING hocr_str("");
 
@@ -1428,12 +1430,23 @@ char* TessBaseAPI::GetHOCRText(int page_number) {
     hocr_str.add_str_int("<span class='ocrx_word' id='word_", page_id);
     hocr_str.add_str_int("_", wcnt);
     int left, top, right, bottom;
+    bool bold, italic, underlined, monospace, serif, smallcaps;
+    int pointsize, font_id;
+    const char *font_name;
     res_it->BoundingBox(RIL_WORD, &left, &top, &right, &bottom);
+    font_name = res_it->WordFontAttributes(&bold, &italic, &underlined,
+                                           &monospace, &serif, &smallcaps,
+                                           &pointsize, &font_id);
     hocr_str.add_str_int("' title='bbox ", left);
     hocr_str.add_str_int(" ", top);
     hocr_str.add_str_int(" ", right);
     hocr_str.add_str_int(" ", bottom);
     hocr_str.add_str_int("; x_wconf ", res_it->Confidence(RIL_WORD));
+    if (font_info) {
+      hocr_str += "; x_font ";
+      HOcrEscape(font_name, hocr_str);
+      hocr_str.add_str_int("; x_fsize ", pointsize);
+    }
     hocr_str += "'";
     if (res_it->WordRecognitionLanguage()) {
       hocr_str += " lang='";
@@ -1447,12 +1460,6 @@ char* TessBaseAPI::GetHOCRText(int page_number) {
         break;
     }
     hocr_str += ">";
-    bool bold, italic, underlined, monospace, serif, smallcaps;
-    int pointsize, font_id;
-    // TODO(rays): Is hOCR interested in the font name?
-    (void) res_it->WordFontAttributes(&bold, &italic, &underlined,
-                                      &monospace, &serif, &smallcaps,
-                                      &pointsize, &font_id);
     bool last_word_in_line = res_it->IsAtFinalElement(RIL_TEXTLINE, RIL_WORD);
     bool last_word_in_para = res_it->IsAtFinalElement(RIL_PARA, RIL_WORD);
     bool last_word_in_block = res_it->IsAtFinalElement(RIL_BLOCK, RIL_WORD);

diff --git a/api/capi.cpp b/api/capi.cpp
@@ -42,6 +42,11 @@ TESS_API TessResultRenderer* TESS_CALL TessHOcrRendererCreate(const char* output
     return new TessHOcrRenderer(outputbase);
 }
 
+TESS_API TessResultRenderer* TESS_CALL TessHOcrRendererCreate2(const char* outputbase, BOOL font_info)
+{
+    return new TessHOcrRenderer(outputbase, font_info);
+}
+
 TESS_API TessResultRenderer* TESS_CALL TessPDFRendererCreate(const char* outputbase, const char* datadir)
 {
     return new TessPDFRenderer(outputbase, datadir);

diff --git a/api/capi.h b/api/capi.h
@@ -100,6 +100,7 @@ TESS_API void  TESS_CALL TessDeleteBlockList(BLOCK_LIST* block_list);
 /* Renderer API */
 TESS_API TessResultRenderer* TESS_CALL TessTextRendererCreate(const char* outputbase);
 TESS_API TessResultRenderer* TESS_CALL TessHOcrRendererCreate(const char* outputbase);
+TESS_API TessResultRenderer* TESS_CALL TessHOcrRendererCreate2(const char* outputbase, BOOL font_info);
 TESS_API TessResultRenderer* TESS_CALL TessPDFRendererCreate(const char* outputbase, const char* datadir);
 TESS_API TessResultRenderer* TESS_CALL TessUnlvRendererCreate(const char* outputbase);
 TESS_API TessResultRenderer* TESS_CALL TessBoxTextRendererCreate(const char* outputbase);
@@ -161,7 +162,7 @@ TESS_API int   TESS_CALL TessBaseAPIInit1(TessBaseAPI* handle, const char* datap
 TESS_API int   TESS_CALL TessBaseAPIInit2(TessBaseAPI* handle, const char* datapath, const char* language, TessOcrEngineMode oem);
 TESS_API int   TESS_CALL TessBaseAPIInit3(TessBaseAPI* handle, const char* datapath, const char* language);
 
-TESS_API int TESS_CALL TessBaseAPIInit4(TessBaseAPI* handle, const char* datapath, const char* language, TessOcrEngineMode mode, 
+TESS_API int TESS_CALL TessBaseAPIInit4(TessBaseAPI* handle, const char* datapath, const char* language, TessOcrEngineMode mode,
     char** configs, int configs_size,
     char** vars_vec, char** vars_values, size_t vars_vec_size,
     BOOL set_only_non_debug_params);

diff --git a/api/renderer.cpp b/api/renderer.cpp
@@ -122,6 +122,12 @@ bool TessTextRenderer::AddImageHandler(TessBaseAPI* api) {
  **********************************************************************/
 TessHOcrRenderer::TessHOcrRenderer(const char *outputbase)
     : TessResultRenderer(outputbase, "hocr") {
+    font_info_ = false;
+}
+
+TessHOcrRenderer::TessHOcrRenderer(const char *outputbase, bool font_info)
+    : TessResultRenderer(outputbase, "hocr") {
+    font_info_ = font_info;
 }
 
 bool TessHOcrRenderer::BeginDocumentHandler() {
@@ -139,7 +145,12 @@ bool TessHOcrRenderer::BeginDocumentHandler() {
       "  <meta name='ocr-system' content='tesseract " TESSERACT_VERSION_STR
               "' />\n"
       "  <meta name='ocr-capabilities' content='ocr_page ocr_carea ocr_par"
-      " ocr_line ocrx_word'/>\n"
+      " ocr_line ocrx_word");
+  if (font_info_)
+    AppendString(
+      " ocrp_lang ocrp_dir ocrp_font ocrp_fsize ocrp_wconf");
+  AppendString(
+      "'/>\n"
       "</head>\n<body>\n");
 
   return true;

diff --git a/api/renderer.h b/api/renderer.h
@@ -150,12 +150,16 @@ class TESS_API TessTextRenderer : public TessResultRenderer {
  */
 class TESS_API TessHOcrRenderer : public TessResultRenderer {
  public:
+  explicit TessHOcrRenderer(const char *outputbase, bool font_info);
   explicit TessHOcrRenderer(const char *outputbase);
 
 protected:
   virtual bool BeginDocumentHandler();
   virtual bool AddImageHandler(TessBaseAPI* api);
   virtual bool EndDocumentHandler();
+
+private:
+  bool font_info_;              // whether to print font information
 };
 
 /**

diff --git a/api/tesseractmain.cpp b/api/tesseractmain.cpp
@@ -290,8 +290,11 @@ int main(int argc, char **argv) {
   tesseract::TessResultRenderer* renderer = NULL;
   bool b;
   api.GetBoolVariable("tessedit_create_hocr", &b);
-  if (b)
-    renderer = new tesseract::TessHOcrRenderer(outputbase);
+  if (b) {
+    bool font_info;
+    api.GetBoolVariable("hocr_font_info", &font_info);
+    renderer = new tesseract::TessHOcrRenderer(outputbase, font_info);
+  }
 
   api.GetBoolVariable("tessedit_create_pdf", &b);
   if (b && renderer == NULL)

diff --git a/ccmain/tesseractclass.cpp b/ccmain/tesseractclass.cpp
@@ -208,6 +208,8 @@ Tesseract::Tesseract()
                   "good_quality_doc gte good char limit", this->params()),
     BOOL_MEMBER(unlv_tilde_crunching, true,
                 "Mark v.bad words for tilde crunch", this->params()),
+    BOOL_MEMBER(hocr_font_info, false,
+                "Add font info to hocr output", this->params()),
     BOOL_MEMBER(crunch_early_merge_tess_fails, true,
                 "Before word crunch?", this->params()),
     BOOL_MEMBER(crunch_early_convert_bad_unlv_chs, false,

diff --git a/ccmain/tesseractclass.h b/ccmain/tesseractclass.h
@@ -836,6 +836,8 @@ class Tesseract : public Wordrec {
                "good_quality_doc gte good char limit");
   BOOL_VAR_H(unlv_tilde_crunching, true,
              "Mark v.bad words for tilde crunch");
+  BOOL_VAR_H(hocr_font_info, false,
+             "Add font info to hocr output");
   BOOL_VAR_H(crunch_early_merge_tess_fails, true, "Before word crunch?");
   BOOL_VAR_H(crunch_early_convert_bad_unlv_chs, false, "Take out ~^ early?");
   double_VAR_H(crunch_terrible_rating, 80.0, "crunch rating lt this");