Use libsub for subrip decoding; improve default position of srt subs.

4 files changed, 41 insertions, 222 deletions

diff --git a/src/lib/subrip.cc b/src/lib/subrip.cc
index 6c9cef916..8a7423698 100644
--- a/src/lib/subrip.cc
+++ b/src/lib/subrip.cc
@@ -17,23 +17,17 @@
 
 */
 
-#include <boost/algorithm/string.hpp>
-#include <boost/lexical_cast.hpp>
 #include "subrip.h"
-#include "subrip_content.h"
-#include "subrip_subtitle.h"
 #include "cross.h"
 #include "exceptions.h"
+#include "subrip_content.h"
+#include <libsub/subrip_reader.h>
+#include <libsub/collect.h>
 
 #include "i18n.h"
 
-using std::string;
-using std::list;
 using std::vector;
-using std::cout;
 using boost::shared_ptr;
-using boost::lexical_cast;
-using boost::algorithm::trim;
 
 SubRip::SubRip (shared_ptr<const SubRipContent> content)
 {
@@ -42,179 +36,8 @@ SubRip::SubRip (shared_ptr<const SubRipContent> content)
 		throw OpenFileError (content->path (0));
 	}
 
-	enum {
-		COUNTER,
-		METADATA,
-		CONTENT
-	} state = COUNTER;
-
-	char buffer[256];
-
-	boost::optional<SubRipSubtitle> current;
-	list<string> lines;
-
-	while (!feof (f)) {
-		fgets (buffer, sizeof (buffer), f);
-		if (feof (f)) {
-			break;
-		}
-
-		string line (buffer);
-		trim_right_if (line, boost::is_any_of ("\n\r"));
-
-		switch (state) {
-		case COUNTER:
-		{
-			if (line.empty ()) {
-				/* a blank line at the start is ok */
-				break;
-			}
-
-			state = METADATA;
-			current = SubRipSubtitle ();
-		}
-		break;
-		case METADATA:
-		{
-			vector<string> p;
-			boost::algorithm::split (p, line, boost::algorithm::is_any_of (" "));
-			if (p.size() != 3 && p.size() != 7) {
-				throw SubRipError (line, _("a time/position line"), content->path (0));
-			}
-
-			current->period = ContentTimePeriod (convert_time (p[0]), convert_time (p[2]));
-
-			if (p.size() > 3) {
-				current->x1 = convert_coordinate (p[3]);
-				current->x2 = convert_coordinate (p[4]);
-				current->y1 = convert_coordinate (p[5]);
-				current->y2 = convert_coordinate (p[6]);
-			}
-			state = CONTENT;
-			break;
-		}
-		case CONTENT:
-			if (line.empty ()) {
-				state = COUNTER;
-				current->pieces = convert_content (lines);
-				_subtitles.push_back (current.get ());
-				current.reset ();
-				lines.clear ();
-			} else {
-				lines.push_back (line);
-			}
-			break;
-		}
-	}
-
-	if (state == CONTENT) {
-		current->pieces = convert_content (lines);
-		_subtitles.push_back (current.get ());
-	}
-
-	fclose (f);
-}
-
-ContentTime
-SubRip::convert_time (string t)
-{
-	ContentTime r;
-
-	vector<string> a;
-	boost::algorithm::split (a, t, boost::is_any_of (":"));
-	assert (a.size() == 3);
-	r += ContentTime::from_seconds (lexical_cast<int> (a[0]) * 60 * 60);
-	r += ContentTime::from_seconds (lexical_cast<int> (a[1]) * 60);
-
-	vector<string> b;
-	boost::algorithm::split (b, a[2], boost::is_any_of (","));
-	r += ContentTime::from_seconds (lexical_cast<int> (b[0]));
-	r += ContentTime::from_seconds (lexical_cast<double> (b[1]) / 1000);
-
-	return r;
-}
-
-int
-SubRip::convert_coordinate (string t)
-{
-	vector<string> a;
-	boost::algorithm::split (a, t, boost::is_any_of (":"));
-	assert (a.size() == 2);
-	return lexical_cast<int> (a[1]);
-}
-
-void
-SubRip::maybe_content (list<SubRipSubtitlePiece>& pieces, SubRipSubtitlePiece& p)
-{
-	if (!p.text.empty ()) {
-		pieces.push_back (p);
-		p.text.clear ();
-	}
-}
-
-list<SubRipSubtitlePiece>
-SubRip::convert_content (list<string> t)
-{
-	list<SubRipSubtitlePiece> pieces;
-	
-	SubRipSubtitlePiece p;
-
-	enum {
-		TEXT,
-		TAG
-	} state = TEXT;
-
-	string tag;
-
-	/* XXX: missing <font> support */
-	/* XXX: nesting of tags e.g. <b>foo<i>bar<b>baz</b>fred</i>jim</b> might
-	   not work, I think.
-	*/
-
-	for (list<string>::const_iterator i = t.begin(); i != t.end(); ++i) {
-		for (size_t j = 0; j < i->size(); ++j) {
-			switch (state) {
-			case TEXT:
-				if ((*i)[j] == '<' || (*i)[j] == '{') {
-					state = TAG;
-				} else {
-					p.text += (*i)[j];
-				}
-				break;
-			case TAG:
-				if ((*i)[j] == '>' || (*i)[j] == '}') {
-					if (tag == "b") {
-						maybe_content (pieces, p);
-						p.bold = true;
-					} else if (tag == "/b") {
-						maybe_content (pieces, p);
-						p.bold = false;
-					} else if (tag == "i") {
-						maybe_content (pieces, p);
-						p.italic = true;
-					} else if (tag == "/i") {
-						maybe_content (pieces, p);
-						p.italic = false;
-					} else if (tag == "u") {
-						maybe_content (pieces, p);
-						p.underline = true;
-					} else if (tag == "/u") {
-						maybe_content (pieces, p);
-						p.underline = false;
-					}
-					tag.clear ();
-					state = TEXT;
-				} else {
-					tag += (*i)[j];
-				}
-				break;
-			}
-		}
-	}
-
-	maybe_content (pieces, p);
-
-	return pieces;
+	sub::SubripReader reader (f);
+	_subtitles = sub::collect<vector<sub::Subtitle> > (reader.subtitles ());
 }
 
 ContentTime
@@ -224,5 +47,5 @@ SubRip::length () const
 		return ContentTime ();
 	}
 
-	return _subtitles.back().period.to;
+	return ContentTime::from_seconds (_subtitles.back().to.metric().get().all_as_seconds ());
 }
diff --git a/src/lib/subrip.h b/src/lib/subrip.h
index 7603a101d..14bc360c0 100644
--- a/src/lib/subrip.h
+++ b/src/lib/subrip.h
@@ -21,6 +21,7 @@
 #define DCPOMATIC_SUBRIP_H
 
 #include "subrip_subtitle.h"
+#include <libsub/subtitle.h>
 
 class SubRipContent;
 class subrip_time_test;
@@ -36,18 +37,7 @@ public:
 	ContentTime length () const;
 
 protected:
-	std::vector<SubRipSubtitle> _subtitles;
-	
-private:
-	friend struct subrip_time_test;
-	friend struct subrip_coordinate_test;
-	friend struct subrip_content_test;
-	friend struct subrip_parse_test;
-	
-	static ContentTime convert_time (std::string);
-	static int convert_coordinate (std::string);
-	static std::list<SubRipSubtitlePiece> convert_content (std::list<std::string>);
-	static void maybe_content (std::list<SubRipSubtitlePiece> &, SubRipSubtitlePiece &);
+	std::vector<sub::Subtitle> _subtitles;
 };
 
 #endif
diff --git a/src/lib/subrip_decoder.cc b/src/lib/subrip_decoder.cc
index 2ae285b5e..411e7542d 100644
--- a/src/lib/subrip_decoder.cc
+++ b/src/lib/subrip_decoder.cc
@@ -39,11 +39,9 @@ SubRipDecoder::seek (ContentTime time, bool accurate)
 	SubtitleDecoder::seek (time, accurate);
 	
 	_next = 0;
-	list<SubRipSubtitlePiece>::const_iterator i = _subtitles[_next].pieces.begin();
-	while (i != _subtitles[_next].pieces.end() && _subtitles[_next].period.from < time) {
-		++i;
+	while (_next < _subtitles.size() && ContentTime::from_seconds (_subtitles[_next].from.metric().get().all_as_seconds ()) < time) {
+		++_next;
 	}
-	
 }
 
 bool
@@ -56,25 +54,27 @@ SubRipDecoder::pass ()
 	/* XXX: we are ignoring positioning specified in the file */
 	
 	list<dcp::SubtitleString> out;
-	for (list<SubRipSubtitlePiece>::const_iterator i = _subtitles[_next].pieces.begin(); i != _subtitles[_next].pieces.end(); ++i) {
-		out.push_back (
-			dcp::SubtitleString (
-				"Arial",
-				i->italic,
-				dcp::Color (255, 255, 255),
-				/* .srt files don't specify size, so this is an arbitrary value */
-				48,
-				dcp::Time (rint (_subtitles[_next].period.from.seconds() * 250)),
-				dcp::Time (rint (_subtitles[_next].period.to.seconds() * 250)),
-				0.2,
-				dcp::BOTTOM,
-				i->text,
-				dcp::NONE,
-				dcp::Color (255, 255, 255),
-				0,
-				0
-				)
-			);
+	for (list<sub::Line>::const_iterator i = _subtitles[_next].lines.begin(); i != _subtitles[_next].lines.end(); ++i) {
+		for (list<sub::Block>::const_iterator j = i->blocks.begin(); j != i->blocks.end(); ++j) {
+			out.push_back (
+				dcp::SubtitleString (
+					"Arial",
+					j->italic,
+					dcp::Color (255, 255, 255),
+					/* .srt files don't specify size, so this is an arbitrary value */
+					48,
+					dcp::Time (rint (_subtitles[_next].from.metric().get().all_as_milliseconds() / 4)),
+					dcp::Time (rint (_subtitles[_next].to.metric().get().all_as_milliseconds() / 4)),
+					i->vertical_position.line.get() * (1.5 / 22) + 0.8,
+					dcp::TOP,
+					j->text,
+					dcp::NONE,
+					dcp::Color (255, 255, 255),
+					0,
+					0
+					)
+				);
+		}
 	}
 
 	text_subtitle (out);
@@ -89,9 +89,15 @@ SubRipDecoder::subtitles_during (ContentTimePeriod p, bool starting) const
 
 	list<ContentTimePeriod> d;
 
-	for (vector<SubRipSubtitle>::const_iterator i = _subtitles.begin(); i != _subtitles.end(); ++i) {
-		if ((starting && p.contains (i->period.from)) || (!starting && p.overlaps (i->period))) {
-			d.push_back (i->period);
+	for (vector<sub::Subtitle>::const_iterator i = _subtitles.begin(); i != _subtitles.end(); ++i) {
+
+		ContentTimePeriod t (
+			ContentTime::from_seconds (i->from.metric().get().all_as_seconds()),
+			ContentTime::from_seconds (i->to.metric().get().all_as_seconds())
+			);
+		
+		if ((starting && p.contains (t.from)) || (!starting && p.overlaps (t))) {
+			d.push_back (t);
 		}
 	}
 
diff --git a/src/lib/wscript b/src/lib/wscript
index 4e62206f1..d62c22bae 100644
--- a/src/lib/wscript
+++ b/src/lib/wscript
@@ -106,7 +106,7 @@ def build(bld):
                  AVCODEC AVUTIL AVFORMAT AVFILTER SWSCALE SWRESAMPLE 
                  BOOST_FILESYSTEM BOOST_THREAD BOOST_DATETIME BOOST_SIGNALS2
                  SNDFILE OPENJPEG POSTPROC TIFF MAGICK SSH DCP CXML GLIB LZMA XML++
-                 CURL ZIP QUICKMAIL PANGOMM CAIROMM XMLSEC
+                 CURL ZIP QUICKMAIL PANGOMM CAIROMM XMLSEC SUB
                  """
 
     if bld.env.TARGET_OSX: