Add subtitle analysis so that the outline of all subtitles

in a piece of content can be overlaid onto the preview (#1233).

9 files changed, 395 insertions, 2 deletions

diff --git a/src/lib/analyse_subtitles_job.cc b/src/lib/analyse_subtitles_job.cc
new file mode 100644
index 000000000..7a1c4ab00
--- /dev/null
+++ b/src/lib/analyse_subtitles_job.cc
@@ -0,0 +1,118 @@
+/*
+    Copyright (C) 2020 Carl Hetherington <cth@carlh.net>
+
+    This file is part of DCP-o-matic.
+
+    DCP-o-matic is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    DCP-o-matic is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with DCP-o-matic.  If not, see <http://www.gnu.org/licenses/>.
+
+*/
+
+#include "analyse_subtitles_job.h"
+#include "playlist.h"
+#include "player.h"
+#include "subtitle_analysis.h"
+#include "bitmap_text.h"
+#include "render_text.h"
+#include "text_content.h"
+#include "image.h"
+#include <iostream>
+
+#include "i18n.h"
+
+using std::string;
+using boost::shared_ptr;
+using boost::weak_ptr;
+
+AnalyseSubtitlesJob::AnalyseSubtitlesJob (shared_ptr<const Film> film, shared_ptr<Content> content)
+	: Job (film)
+	, _content (content)
+	, _path (_film->subtitle_analysis_path(content))
+{
+}
+
+
+string
+AnalyseSubtitlesJob::name () const
+{
+	return _("Analysing subtitles");
+}
+
+
+string
+AnalyseSubtitlesJob::json_name () const
+{
+	return N_("analyse_subtitles");
+}
+
+
+void
+AnalyseSubtitlesJob::run ()
+{
+	shared_ptr<Playlist> playlist (new Playlist());
+	shared_ptr<Content> content = _content.lock ();
+	DCPOMATIC_ASSERT (content);
+	playlist->add (_film, content);
+
+	shared_ptr<Player> player (new Player(_film, playlist));
+	player->set_ignore_audio ();
+	player->set_fast ();
+	player->set_play_referenced ();
+	player->Text.connect (bind(&AnalyseSubtitlesJob::analyse, this, _1, _2));
+
+	set_progress_unknown ();
+
+	if (!content->text.empty()) {
+		while (!player->pass ()) {}
+	}
+
+	SubtitleAnalysis analysis (_bounding_box, content->text.front()->x_offset(), content->text.front()->y_offset());
+	analysis.write (_path);
+
+	set_progress (1);
+	set_state (FINISHED_OK);
+}
+
+
+void
+AnalyseSubtitlesJob::analyse (PlayerText text, TextType type)
+{
+	if (type != TEXT_OPEN_SUBTITLE) {
+		return;
+	}
+
+	BOOST_FOREACH (BitmapText const& i, text.bitmap) {
+		if (!_bounding_box) {
+			_bounding_box = i.rectangle;
+		} else {
+			_bounding_box->extend (i.rectangle);
+		}
+	}
+
+	if (!text.string.empty()) {
+		/* We can provide dummy values for time and frame rate here as they are only used to calculate fades */
+		dcp::Size const frame = _film->frame_size();
+		BOOST_FOREACH (PositionImage i, render_text(text.string, text.fonts, frame, dcpomatic::DCPTime(), 24)) {
+			dcpomatic::Rect<double> rect (
+					double(i.position.x) / frame.width, double(i.position.y) / frame.height,
+					double(i.image->size().width) / frame.width, double(i.image->size().height) / frame.height
+					);
+			if (!_bounding_box) {
+				_bounding_box = rect;
+			} else {
+				_bounding_box->extend (rect);
+			}
+		}
+	}
+}
+
diff --git a/src/lib/analyse_subtitles_job.h b/src/lib/analyse_subtitles_job.h
new file mode 100644
index 000000000..88b5d23dc
--- /dev/null
+++ b/src/lib/analyse_subtitles_job.h
@@ -0,0 +1,48 @@
+/*
+    Copyright (C) 2020 Carl Hetherington <cth@carlh.net>
+
+    This file is part of DCP-o-matic.
+
+    DCP-o-matic is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    DCP-o-matic is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with DCP-o-matic.  If not, see <http://www.gnu.org/licenses/>.
+
+*/
+
+#include "job.h"
+#include "types.h"
+#include "player_text.h"
+
+class Film;
+class Content;
+
+class AnalyseSubtitlesJob : public Job
+{
+public:
+	AnalyseSubtitlesJob (boost::shared_ptr<const Film> film, boost::shared_ptr<Content> content);
+
+	std::string name () const;
+	std::string json_name () const;
+	void run ();
+
+	boost::filesystem::path path () const {
+		return _path;
+	}
+
+private:
+	void analyse (PlayerText text, TextType type);
+
+	boost::weak_ptr<Content> _content;
+	boost::filesystem::path _path;
+	boost::optional<dcpomatic::Rect<double> > _bounding_box;
+};
+
diff --git a/src/lib/film.cc b/src/lib/film.cc
index a89d58a1f..a24e9aa30 100644
--- a/src/lib/film.cc
+++ b/src/lib/film.cc
@@ -54,6 +54,8 @@
 #include "cinema.h"
 #include "change_signaller.h"
 #include "check_content_change_job.h"
+#include "ffmpeg_subtitle_stream.h"
+#include "font.h"
 #include <libcxml/cxml.h>
 #include <dcp/cpl.h>
 #include <dcp/certificate_chain.h>
@@ -302,6 +304,39 @@ Film::audio_analysis_path (shared_ptr<const Playlist> playlist) const
 	return p;
 }
 
+
+boost::filesystem::path
+Film::subtitle_analysis_path (shared_ptr<const Content> content) const
+{
+	boost::filesystem::path p = dir ("analysis");
+
+	Digester digester;
+	digester.add (content->digest());
+
+	if (!content->text.empty()) {
+		shared_ptr<TextContent> tc = content->text.front();
+		digester.add (tc->x_scale());
+		digester.add (tc->y_scale());
+		BOOST_FOREACH (shared_ptr<dcpomatic::Font> i, tc->fonts()) {
+			digester.add (i->id());
+		}
+		if (tc->effect()) {
+			digester.add (tc->effect().get());
+		}
+		digester.add (tc->line_spacing());
+		digester.add (tc->outline_width());
+	}
+
+	shared_ptr<const FFmpegContent> fc = dynamic_pointer_cast<const FFmpegContent>(content);
+	if (fc) {
+		digester.add (fc->subtitle_stream()->identifier());
+	}
+
+	p /= digester.get ();
+	return p;
+}
+
+
 /** Add suitable Jobs to the JobManager to create a DCP for this Film.
  *  @param gui true if this is being called from a GUI tool.
  *  @param check true to check the content in the project for changes before making the DCP.
diff --git a/src/lib/film.h b/src/lib/film.h
index 86e9be6d9..6cce07c17 100644
--- a/src/lib/film.h
+++ b/src/lib/film.h
@@ -106,6 +106,7 @@ public:
 	boost::filesystem::path internal_video_asset_filename (dcpomatic::DCPTimePeriod p) const;
 
 	boost::filesystem::path audio_analysis_path (boost::shared_ptr<const Playlist>) const;
+	boost::filesystem::path subtitle_analysis_path (boost::shared_ptr<const Content>) const;
 
 	void send_dcp_to_tms ();
 	void make_dcp (bool gui = false, bool check = true);
diff --git a/src/lib/job_manager.cc b/src/lib/job_manager.cc
index d95f95a24..c40178f41 100644
--- a/src/lib/job_manager.cc
+++ b/src/lib/job_manager.cc
@@ -1,5 +1,5 @@
 /*
-    Copyright (C) 2012-2018 Carl Hetherington <cth@carlh.net>
+    Copyright (C) 2012-2020 Carl Hetherington <cth@carlh.net>
 
     This file is part of DCP-o-matic.
 
@@ -26,6 +26,7 @@
 #include "job.h"
 #include "cross.h"
 #include "analyse_audio_job.h"
+#include "analyse_subtitles_job.h"
 #include "film.h"
 #include <boost/thread.hpp>
 #include <boost/foreach.hpp>
@@ -250,6 +251,42 @@ JobManager::analyse_audio (
 	emit (boost::bind (boost::ref (JobAdded), weak_ptr<Job> (job)));
 }
 
+
+void
+JobManager::analyse_subtitles (
+	shared_ptr<const Film> film,
+	shared_ptr<Content> content,
+	boost::signals2::connection& connection,
+	function<void()> ready
+	)
+{
+	{
+		boost::mutex::scoped_lock lm (_mutex);
+
+		BOOST_FOREACH (shared_ptr<Job> i, _jobs) {
+			shared_ptr<AnalyseSubtitlesJob> a = dynamic_pointer_cast<AnalyseSubtitlesJob> (i);
+			if (a && a->path() == film->subtitle_analysis_path(content)) {
+				i->when_finished (connection, ready);
+				return;
+			}
+		}
+	}
+
+	shared_ptr<AnalyseSubtitlesJob> job;
+
+	{
+		boost::mutex::scoped_lock lm (_mutex);
+
+		job.reset (new AnalyseSubtitlesJob(film, content));
+		connection = job->Finished.connect (ready);
+		_jobs.push_back (job);
+		_empty_condition.notify_all ();
+	}
+
+	emit (boost::bind (boost::ref (JobAdded), weak_ptr<Job> (job)));
+}
+
+
 void
 JobManager::increase_priority (shared_ptr<Job> job)
 {
diff --git a/src/lib/job_manager.h b/src/lib/job_manager.h
index d4287f874..dd7a28db2 100644
--- a/src/lib/job_manager.h
+++ b/src/lib/job_manager.h
@@ -1,5 +1,5 @@
 /*
-    Copyright (C) 2012-2018 Carl Hetherington <cth@carlh.net>
+    Copyright (C) 2012-2020 Carl Hetherington <cth@carlh.net>
 
     This file is part of DCP-o-matic.
 
@@ -32,6 +32,7 @@
 class Job;
 class Film;
 class Playlist;
+class Content;
 struct threed_test7;
 
 extern bool wait_for_jobs ();
@@ -64,6 +65,13 @@ public:
 		boost::function<void()> ready
 		);
 
+	void analyse_subtitles (
+		boost::shared_ptr<const Film> film,
+		boost::shared_ptr<Content> content,
+		boost::signals2::connection& connection,
+		boost::function<void()> ready
+		);
+
 	boost::signals2::signal<void (boost::weak_ptr<Job>)> JobAdded;
 	boost::signals2::signal<void ()> JobsReordered;
 	boost::signals2::signal<void (boost::optional<std::string>, boost::optional<std::string>)> ActiveJobsChanged;
diff --git a/src/lib/subtitle_analysis.cc b/src/lib/subtitle_analysis.cc
new file mode 100644
index 000000000..91fec90b0
--- /dev/null
+++ b/src/lib/subtitle_analysis.cc
@@ -0,0 +1,75 @@
+/*
+    Copyright (C) 2020 Carl Hetherington <cth@carlh.net>
+
+    This file is part of DCP-o-matic.
+
+    DCP-o-matic is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    DCP-o-matic is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with DCP-o-matic.  If not, see <http://www.gnu.org/licenses/>.
+
+*/
+
+#include "subtitle_analysis.h"
+#include "exceptions.h"
+#include <libcxml/cxml.h>
+#include <dcp/raw_convert.h>
+#include <libxml++/libxml++.h>
+
+using std::string;
+using dcp::raw_convert;
+using boost::shared_ptr;
+
+int const SubtitleAnalysis::_current_state_version = 1;
+
+
+SubtitleAnalysis::SubtitleAnalysis (boost::filesystem::path path)
+{
+	cxml::Document f ("SubtitleAnalysis");
+
+	f.read_file (path);
+
+	if (f.optional_number_child<int>("Version").get_value_or(1) < _current_state_version) {
+		/* Too old.  Throw an exception so that this analysis is re-run. */
+		throw OldFormatError ("Audio analysis file is too old");
+	}
+
+	cxml::NodePtr bounding_box = f.optional_node_child("BoundingBox");
+	if (bounding_box) {
+		_bounding_box = dcpomatic::Rect<double> ();
+		_bounding_box->x = bounding_box->number_child<double>("X");
+		_bounding_box->y = bounding_box->number_child<double>("Y");
+		_bounding_box->width = bounding_box->number_child<double>("Width");
+		_bounding_box->height = bounding_box->number_child<double>("Height");
+	}
+}
+
+
+void
+SubtitleAnalysis::write (boost::filesystem::path path) const
+{
+	shared_ptr<xmlpp::Document> doc (new xmlpp::Document);
+	xmlpp::Element* root = doc->create_root_node ("SubtitleAnalysis");
+
+	root->add_child("Version")->add_child_text (raw_convert<string>(_current_state_version));
+
+	if (_bounding_box) {
+		xmlpp::Element* bounding_box = root->add_child("BoundingBox");
+		bounding_box->add_child("X")->add_child_text(raw_convert<string>(_bounding_box->x));
+		bounding_box->add_child("Y")->add_child_text(raw_convert<string>(_bounding_box->y));
+		bounding_box->add_child("Width")->add_child_text(raw_convert<string>(_bounding_box->width));
+		bounding_box->add_child("Height")->add_child_text(raw_convert<string>(_bounding_box->height));
+	}
+
+	doc->write_to_file_formatted (path.string());
+}
+
+
diff --git a/src/lib/subtitle_analysis.h b/src/lib/subtitle_analysis.h
new file mode 100644
index 000000000..47eb52411
--- /dev/null
+++ b/src/lib/subtitle_analysis.h
@@ -0,0 +1,69 @@
+/*
+    Copyright (C) 2020 Carl Hetherington <cth@carlh.net>
+
+    This file is part of DCP-o-matic.
+
+    DCP-o-matic is free software; you can redistribute it and/or modify
+    it under the terms of the GNU General Public License as published by
+    the Free Software Foundation; either version 2 of the License, or
+    (at your option) any later version.
+
+    DCP-o-matic is distributed in the hope that it will be useful,
+    but WITHOUT ANY WARRANTY; without even the implied warranty of
+    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+    GNU General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with DCP-o-matic.  If not, see <http://www.gnu.org/licenses/>.
+
+*/
+
+#include "rect.h"
+#include <boost/noncopyable.hpp>
+#include <boost/filesystem.hpp>
+
+
+/** @class SubtitleAnalysis
+ *  @brief Class to store the results of a SubtitleAnalysisJob.
+ */
+
+class SubtitleAnalysis : public boost::noncopyable
+{
+public:
+	explicit SubtitleAnalysis (boost::filesystem::path path);
+
+	SubtitleAnalysis (
+			boost::optional<dcpomatic::Rect<double> > bounding_box,
+			double analysis_x_offset_,
+			double analysis_y_offset_
+			)
+		: _bounding_box (bounding_box)
+		, _analysis_x_offset (analysis_x_offset_)
+		, _analysis_y_offset (analysis_y_offset_)
+	{}
+
+	void write (boost::filesystem::path path) const;
+
+	boost::optional<dcpomatic::Rect<double> > bounding_box () const {
+		return _bounding_box;
+	}
+
+	double analysis_x_offset () const {
+		return _analysis_x_offset;
+	}
+
+	double analysis_y_offset () const {
+		return _analysis_x_offset;
+	}
+
+private:
+	/** Smallest box which surrounds all subtitles in our content,
+	 *  expressed as a proportion of screen size (i.e. 0 is left hand side/top,
+	 *  1 is right hand side/bottom), or empty if no subtitles were found.
+	 */
+	boost::optional<dcpomatic::Rect<double> > _bounding_box;
+	double _analysis_x_offset;
+	double _analysis_y_offset;
+
+	static int const _current_state_version;
+};
diff --git a/src/lib/wscript b/src/lib/wscript
index 17f96d61e..802020527 100644
--- a/src/lib/wscript
+++ b/src/lib/wscript
@@ -23,6 +23,7 @@ import i18n
 sources = """
           active_text.cc
           analyse_audio_job.cc
+          analyse_subtitles_job.cc
           analytics.cc
           atmos_mxf_content.cc
           atomicity_checker.cc
@@ -157,6 +158,7 @@ sources = """
           string_text_file.cc
           string_text_file_content.cc
           string_text_file_decoder.cc
+          subtitle_analysis.cc
           subtitle_encoder.cc
           text_ring_buffers.cc
           timer.cc