--- scribo/ChangeLog | 4 ++ scribo/io/xml/save_text_lines.hh | 98 ++++++++++++++++++++------------------ 2 files changed, 55 insertions(+), 47 deletions(-) diff --git a/scribo/ChangeLog b/scribo/ChangeLog index 53358b7..5dc9919 100644 --- a/scribo/ChangeLog +++ b/scribo/ChangeLog @@ -1,5 +1,9 @@ 2010-04-13 Guillaume Lazzara <z@lrde.epita.fr> + * io/xml/save_text_lines.hh: Fix XML output. + +2010-04-13 Guillaume Lazzara <z@lrde.epita.fr> + * debug/decision_image.hh: Add the anchor::Type as argument. 2010-04-13 Guillaume Lazzara <z@lrde.epita.fr> diff --git a/scribo/io/xml/save_text_lines.hh b/scribo/io/xml/save_text_lines.hh index 22aa3d9..7f0fdf4 100644 --- a/scribo/io/xml/save_text_lines.hh +++ b/scribo/io/xml/save_text_lines.hh @@ -84,56 +84,60 @@ namespace scribo file << "<PcGts xmlns=\"http://schema.primaresearch.org/PAGE/gts/pagecontent/2009-03-16\" xmlns:xsi=\"http://www.w3.org/2001/XMLSchema-instance\" xsi:schemaLocation=\"http://schema.primaresearch.org/PAGE/gts/pagecontent/2009-03-16 http://schema.primaresearch.org/PAGE/gts/pagecontent/2009-03-16/pagecontent.xsd\" pcGtsId=\"" << input_name << "\">" << std::endl; - file << " <Metadata>" << std::endl; - file << " <Creator>LRDE</pcCreator>" << std::endl; - file << " <Created/>" << std::endl; - file << " <LastChange/>" << std::endl; - file << " <Comments>Generated by Scribo from Olena.</pcComments>" << std::endl; - file << " </Metadata>" << std::endl; - - file << " <Page image_filename=\"" << input_name - << "\" image_width=\"" << lines.component_set_().labeled_image().ncols() - << "\" image_height=\"" << lines.component_set_().labeled_image().nrows() + file << " <PcMetadata>" << std::endl; + file << " <PcCreator>LRDE</PcCreator>" << std::endl; + file << " <PcCreated/>" << std::endl; + file << " <PcLastChange/>" << std::endl; + file << " <PcComments>Generated by Scribo from Olena.</PcComments>" << std::endl; + file << " </PcMetadata>" << std::endl; + + file << " <page image_filename=\"" << input_name + << "\" image_width=\"" << lines.components().labeled_image().ncols() + << "\" image_height=\"" << lines.components().labeled_image().nrows() << "\">" << std::endl; for_all_lines(l, lines) - { - file << " <text_region id=\"" << lines(l).id() - << "\" txt_orientation=\"" << lines(l).orientation() - << "\" txt_reading_orientation=\"" << lines(l).reading_orientation() - << "\" txt_reading_direction=\"" << lines(l).reading_direction() - << "\" txt_reverse_video=\"" << (lines(l).reverse_video() ? "true" : "false") - << "\" txt_indented=\"" << (lines(l).indented() ? "true" : "false") - // EXTENSIONS - Not officially supported - << "\" baseline=\"" << lines(l).baseline() - << "\" median=\"" << lines(l).median() - << "\" x_height=\"" << lines(l).x_height() - << "\" d_height=\"" << lines(l).d_height() - << "\" a_height=\"" << lines(l).a_height() - << "\" char_space=\"" << lines(l).char_space() - << "\" char_width=\"" << lines(l).char_width() - // End of EXTENSIONS - << "\">" - << std::endl; - - file << " <coords>" << std::endl - << " <point x=\"" << lines(l).bbox().pmin().row() - << "\" y=\"" << lines(l).bbox().pmin().col() << "\"/>" - << std::endl - << " <point x=\"" << lines(l).bbox().pmin().row() - << "\" y=\"" << lines(l).bbox().pmax().col() << "\"/>" - << std::endl - << " <point x=\"" << lines(l).bbox().pmax().row() - << "\" y=\"" << lines(l).bbox().pmin().col() << "\"/>" - << std::endl - << " <point x=\"" << lines(l).bbox().pmax().row() - << "\" y=\"" << lines(l).bbox().pmax().col() << "\"/>" - << std::endl - << " </coords>" << std::endl; - } - - file << " </text_region>" << std::endl; - file << " </Page>" << std::endl; + if (lines(l).tag() != line::Merged + && lines(l).tag() != line::Ignored) + { + file << " <text_region id=\"" << lines(l).id() + << "\" txt_orientation=\"" << lines(l).orientation() + << "\" txt_reading_orientation=\"" << lines(l).reading_orientation() + << "\" txt_reading_direction=\"" << lines(l).reading_direction() + << "\" txt_reverse_video=\"" << (lines(l).reverse_video() ? "true" : "false") + << "\" txt_indented=\"" << (lines(l).indented() ? "true" : "false") + << "\" kerning=\"" << lines(l).char_space() + + // EXTENSIONS - Not officially supported + << "\" baseline=\"" << lines(l).baseline() + << "\" meanline=\"" << lines(l).meanline() + << "\" x_height=\"" << lines(l).x_height() + << "\" d_height=\"" << lines(l).d_height() + << "\" a_height=\"" << lines(l).a_height() + << "\" char_width=\"" << lines(l).char_width() + // End of EXTENSIONS + << "\">" + << std::endl; + + file << " <coords>" << std::endl + << " <point x=\"" << lines(l).bbox().pmin().col() + << "\" y=\"" << lines(l).bbox().pmin().row() << "\"/>" + << std::endl + << " <point x=\"" << lines(l).bbox().pmax().col() + << "\" y=\"" << lines(l).bbox().pmin().row() << "\"/>" + << std::endl + << " <point x=\"" << lines(l).bbox().pmax().col() + << "\" y=\"" << lines(l).bbox().pmax().row() << "\"/>" + << std::endl + << " <point x=\"" << lines(l).bbox().pmin().col() + << "\" y=\"" << lines(l).bbox().pmax().row() << "\"/>" + << std::endl + << " </coords>" << std::endl; + + file << " </text_region>" << std::endl; + } + + file << " </page>" << std::endl; file << "</PcGts>" << std::endl; -- 1.5.6.5