mirror/tesseract
1
0
Fork 0
mirror of https://github.com/tesseract-ocr/tesseract.git synced 2024-12-23 06:57:50 +08:00
Code Issues Actions 9 Packages Projects Releases Wiki Activity

Fix bug with linebreaking in hOCR

Browse Source 
The hOCR output could incorrectly close span, p, and div tags
early. Oops, my bad.
This commit is contained in:
Nick White 2016-06-29 09:22:48 +01:00
parent d71133a769
commit 78ae2cc073
1 changed files with 4 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
api/baseapi.cpp
View File

@ -1584,6 +1584,10 @@ char* TessBaseAPI::GetHOCRText(struct ETEXT_DESC* monitor, int page_number) {
} }
hocr_str += "'>"; hocr_str += "'>";
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);
if (bold) hocr_str += "<strong>"; if (bold) hocr_str += "<strong>";
if (italic) hocr_str += "<em>"; if (italic) hocr_str += "<em>";
@ -1614,9 +1618,6 @@ char* TessBaseAPI::GetHOCRText(struct ETEXT_DESC* monitor, int page_number) {
if (bold) hocr_str += "</strong>"; if (bold) hocr_str += "</strong>";
hocr_str += "</span> "; hocr_str += "</span> ";
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);
wcnt++; wcnt++;
// Close any ending block/paragraph/textline. // Close any ending block/paragraph/textline.
if (last_word_in_line) { if (last_word_in_line) {