From 54e2644ed4dcc4faf5135e60926cfdda14886fea Mon Sep 17 00:00:00 2001
From: Carl Hetherington <cth@carlh.net>
Date: Fri, 6 Dec 2024 17:46:24 +0100
Subject: Support FCP XML files containing subtitles (#2909).

---
 src/lib/content_factory.cc |   5 +++
 src/lib/decoder_factory.cc |   6 +++
 src/lib/exceptions.h       |   9 ++++
 src/lib/fcpxml.cc          |  82 ++++++++++++++++++++++++++++++++++
 src/lib/fcpxml.h           |  55 +++++++++++++++++++++++
 src/lib/fcpxml_content.cc  | 108 +++++++++++++++++++++++++++++++++++++++++++++
 src/lib/fcpxml_content.h   |  57 ++++++++++++++++++++++++
 src/lib/fcpxml_decoder.cc  | 105 +++++++++++++++++++++++++++++++++++++++++++
 src/lib/fcpxml_decoder.h   |  50 +++++++++++++++++++++
 src/lib/wscript            |   3 ++
 10 files changed, 480 insertions(+)
 create mode 100644 src/lib/fcpxml.cc
 create mode 100644 src/lib/fcpxml.h
 create mode 100644 src/lib/fcpxml_content.cc
 create mode 100644 src/lib/fcpxml_content.h
 create mode 100644 src/lib/fcpxml_decoder.cc
 create mode 100644 src/lib/fcpxml_decoder.h

(limited to 'src/lib')

diff --git a/src/lib/content_factory.cc b/src/lib/content_factory.cc
index 9f464896a..fa3f0876c 100644
--- a/src/lib/content_factory.cc
+++ b/src/lib/content_factory.cc
@@ -29,6 +29,7 @@
 #include "dcp_content.h"
 #include "dcp_subtitle_content.h"
 #include "dcpomatic_log.h"
+#include "fcpxml_content.h"
 #include "ffmpeg_audio_stream.h"
 #include "ffmpeg_content.h"
 #include "film.h"
@@ -97,6 +98,8 @@ content_factory(cxml::ConstNodePtr node, boost::optional<boost::filesystem::path
 		content = std::make_shared<VideoMXFContent>(node, film_directory, version);
 	} else if (type == "AtmosMXF") {
 		content = std::make_shared<AtmosMXFContent>(node, film_directory, version);
+	} else if (type == "FCPXML") {
+		content = std::make_shared<FCPXMLContent>(node, film_directory, version, notes);
 	}
 
 	return content;
@@ -178,6 +181,8 @@ content_factory (boost::filesystem::path path)
 				throw KDMAsContentError ();
 			}
 			single = std::make_shared<DCPSubtitleContent>(path);
+		} else if (ext == ".fcpxml") {
+			single = std::make_shared<FCPXMLContent>(path);
 		} else if (ext == ".mxf" && dcp::SMPTETextAsset::valid_mxf(path)) {
 			single = std::make_shared<DCPSubtitleContent>(path);
 		} else if (ext == ".mxf" && VideoMXFContent::valid_mxf(path)) {
diff --git a/src/lib/decoder_factory.cc b/src/lib/decoder_factory.cc
index ea0eda83d..31156ab3f 100644
--- a/src/lib/decoder_factory.cc
+++ b/src/lib/decoder_factory.cc
@@ -25,6 +25,8 @@
 #include "dcp_decoder.h"
 #include "dcp_subtitle_content.h"
 #include "dcp_subtitle_decoder.h"
+#include "fcpxml_content.h"
+#include "fcpxml_decoder.h"
 #include "ffmpeg_content.h"
 #include "ffmpeg_decoder.h"
 #include "image_content.h"
@@ -92,5 +94,9 @@ decoder_factory (shared_ptr<const Film> film, shared_ptr<const Content> content,
 		return make_shared<AtmosMXFDecoder>(film, amc);
 	}
 
+	if (auto c = dynamic_pointer_cast<const FCPXMLContent>(content)) {
+		return make_shared<FCPXMLDecoder>(film, c);
+	}
+
 	return {};
 }
diff --git a/src/lib/exceptions.h b/src/lib/exceptions.h
index 3069d792e..c1bb8f6b6 100644
--- a/src/lib/exceptions.h
+++ b/src/lib/exceptions.h
@@ -542,4 +542,13 @@ public:
 };
 
 
+class FCPXMLError : public std::runtime_error
+{
+public:
+	explicit FCPXMLError(std::string s)
+		: std::runtime_error(s)
+	{}
+};
+
+
 #endif
diff --git a/src/lib/fcpxml.cc b/src/lib/fcpxml.cc
new file mode 100644
index 000000000..49798381c
--- /dev/null
+++ b/src/lib/fcpxml.cc
@@ -0,0 +1,82 @@
+/*
+    Copyright (C) 2024 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 "fcpxml.h"
+#include <dcp/raw_convert.h>
+#include <libcxml/cxml.h>
+#include <boost/algorithm/string.hpp>
+#include <boost/filesystem.hpp>
+
+
+#include "i18n.h"
+
+
+using std::map;
+using std::string;
+using std::vector;
+
+
+static dcpomatic::ContentTime
+convert_time(string const& time)
+{
+	vector<string> parts;
+	boost::algorithm::split(parts, time, boost::is_any_of("/"));
+
+	if (parts.size() != 2 || parts[1].empty() || parts[1][parts[1].length() - 1] != 's') {
+		throw FCPXMLError(String::compose("Unexpected time format %1", time));
+	}
+
+	return dcpomatic::ContentTime{dcp::raw_convert<int64_t>(parts[0]) * dcpomatic::ContentTime::HZ / dcp::raw_convert<int64_t>(parts[1])};
+}
+
+
+dcpomatic::fcpxml::Sequence
+dcpomatic::fcpxml::load(boost::filesystem::path xml_file)
+{
+	cxml::Document doc("fcpxml");
+	doc.read_file(xml_file);
+
+	auto project = doc.node_child("project");
+
+	map<string, boost::filesystem::path> assets;
+	for (auto asset: project->node_child("resources")->node_children("asset")) {
+		assets[asset->string_attribute("name")] = asset->string_attribute("src");
+	}
+
+	auto sequence = Sequence{xml_file.parent_path()};
+	for (auto video: project->node_child("sequence")->node_child("spine")->node_children("video")) {
+		auto name = video->string_attribute("name");
+		auto iter = assets.find(name);
+		if (iter == assets.end()) {
+			throw FCPXMLError(String::compose(_("Video refers to missing asset %1"), name));
+		}
+
+		auto start = convert_time(video->string_attribute("offset"));
+		sequence.video.push_back(
+			{
+				iter->second,
+				dcpomatic::ContentTimePeriod(start, start + convert_time(video->string_attribute("duration")))
+			}
+		);
+	}
+
+	return sequence;
+}
diff --git a/src/lib/fcpxml.h b/src/lib/fcpxml.h
new file mode 100644
index 000000000..00137ca32
--- /dev/null
+++ b/src/lib/fcpxml.h
@@ -0,0 +1,55 @@
+/*
+    Copyright (C) 2024 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 "dcpomatic_time.h"
+#include <vector>
+
+
+namespace dcpomatic {
+namespace fcpxml {
+
+
+class Video
+{
+public:
+	boost::filesystem::path source;  ///< filename of PNG relative to Sequence::parent
+	dcpomatic::ContentTimePeriod period;
+};
+
+
+
+class Sequence
+{
+public:
+	Sequence(boost::filesystem::path parent_)
+		: parent(parent_)
+	{}
+
+	boost::filesystem::path parent;  ///< directory containing the PNG files
+	std::vector<Video> video;
+};
+
+
+Sequence load(boost::filesystem::path xml_file);
+
+
+}
+}
diff --git a/src/lib/fcpxml_content.cc b/src/lib/fcpxml_content.cc
new file mode 100644
index 000000000..94d1197df
--- /dev/null
+++ b/src/lib/fcpxml_content.cc
@@ -0,0 +1,108 @@
+/*
+    Copyright (C) 2024 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 "fcpxml.h"
+#include "fcpxml_content.h"
+#include "text_content.h"
+#include <dcp/raw_convert.h>
+
+#include "i18n.h"
+
+
+using std::list;
+using std::make_shared;
+using std::shared_ptr;
+using std::string;
+using boost::optional;
+
+
+FCPXMLContent::FCPXMLContent(boost::filesystem::path path)
+	: Content(path)
+{
+	text.push_back(make_shared<TextContent>(this, TextType::OPEN_SUBTITLE, TextType::OPEN_SUBTITLE));
+}
+
+
+FCPXMLContent::FCPXMLContent(cxml::ConstNodePtr node, optional<boost::filesystem::path> film_directory, int version, list<string>& notes)
+	: Content(node, film_directory)
+{
+	text = TextContent::from_xml(this, node, version, notes);
+}
+
+
+
+void
+FCPXMLContent::examine(shared_ptr<const Film> film, shared_ptr<Job> job)
+{
+	Content::examine(film, job);
+
+	auto sequence = dcpomatic::fcpxml::load(path(0));
+
+	boost::mutex::scoped_lock lm(_mutex);
+	only_text()->set_use(true);
+	if (!sequence.video.empty()) {
+		_length = sequence.video.back().period.to;
+	}
+}
+
+
+dcpomatic::DCPTime
+FCPXMLContent::full_length(shared_ptr<const Film> film) const
+{
+	FrameRateChange const frc(film, shared_from_this());
+	return { _length, frc };
+}
+
+
+dcpomatic::DCPTime
+FCPXMLContent::approximate_length() const
+{
+	return { _length, {} };
+}
+
+
+string
+FCPXMLContent::summary() const
+{
+	return path_summary() + " " + _("[subtitles]");
+}
+
+
+string
+FCPXMLContent::technical_summary() const
+{
+	return Content::technical_summary() + " - " + _("FCP XML subtitles");
+}
+
+
+void
+FCPXMLContent::as_xml(xmlpp::Element* element, bool with_paths, PathBehaviour path_behaviour, optional<boost::filesystem::path> film_directory) const
+{
+	cxml::add_child(element, "Type", "FCPXML");
+	Content::as_xml(element, with_paths, path_behaviour, film_directory);
+
+	if (only_text()) {
+		only_text()->as_xml(element);
+	}
+
+	cxml::add_child(element, "Length", dcp::raw_convert<string>(_length.get()));
+}
+
diff --git a/src/lib/fcpxml_content.h b/src/lib/fcpxml_content.h
new file mode 100644
index 000000000..701fd6d02
--- /dev/null
+++ b/src/lib/fcpxml_content.h
@@ -0,0 +1,57 @@
+/*
+    Copyright (C) 2024 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/>.
+
+*/
+
+
+#ifndef DCPOMATIC_FCPXML_CONTENT_H
+#define DCPOMATIC_FCPXML_CONTENT_H
+
+
+#include "content.h"
+
+
+
+class FCPXMLContent : public Content
+{
+public:
+	FCPXMLContent(boost::filesystem::path path);
+	FCPXMLContent(cxml::ConstNodePtr node, boost::optional<boost::filesystem::path> film_directory, int version, std::list<std::string>& notes);
+
+	std::shared_ptr<FCPXMLContent> shared_from_this() {
+		return std::dynamic_pointer_cast<FCPXMLContent>(Content::shared_from_this());
+	}
+
+	std::shared_ptr<const FCPXMLContent> shared_from_this() const {
+		return std::dynamic_pointer_cast<const FCPXMLContent>(Content::shared_from_this());
+	}
+
+	void examine(std::shared_ptr<const Film> film, std::shared_ptr<Job>) override;
+	std::string summary() const override;
+	std::string technical_summary() const override;
+	void as_xml(xmlpp::Element*, bool with_paths, PathBehaviour path_behaviour, boost::optional<boost::filesystem::path> film_directory) const override;
+	dcpomatic::DCPTime full_length(std::shared_ptr<const Film> film) const override;
+	dcpomatic::DCPTime approximate_length() const override;
+
+private:
+	dcpomatic::ContentTime _length;
+};
+
+
+
+#endif
diff --git a/src/lib/fcpxml_decoder.cc b/src/lib/fcpxml_decoder.cc
new file mode 100644
index 000000000..edb19e639
--- /dev/null
+++ b/src/lib/fcpxml_decoder.cc
@@ -0,0 +1,105 @@
+/*
+    Copyright (C) 2024 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 "fcpxml_content.h"
+#include "fcpxml_decoder.h"
+#include "ffmpeg_image_proxy.h"
+#include "guess_crop.h"
+#include "image.h"
+#include "rect.h"
+#include "text_decoder.h"
+#include <dcp/array_data.h>
+
+
+
+using std::make_shared;
+using std::shared_ptr;
+using std::weak_ptr;
+
+
+FCPXMLDecoder::FCPXMLDecoder(weak_ptr<const Film> film, shared_ptr<const FCPXMLContent> content)
+	: Decoder(film)
+	, _fcpxml_content(content)
+	, _sequence(dcpomatic::fcpxml::load(content->path(0)))
+{
+	text.push_back(make_shared<TextDecoder>(this, content->only_text()));
+	update_position();
+}
+
+
+bool
+FCPXMLDecoder::pass()
+{
+	if (_next >= static_cast<int>(_sequence.video.size())) {
+		return true;
+	}
+
+	auto const png_data = dcp::ArrayData(_sequence.parent / _sequence.video[_next].source);
+	auto const full_image = FFmpegImageProxy(png_data).image(Image::Alignment::PADDED).image;
+	auto const crop = guess_crop_by_alpha(full_image);
+	auto const cropped_image = full_image->crop(crop);
+
+	auto rectangle = dcpomatic::Rect<double>{
+		static_cast<double>(crop.left) / full_image->size().width,
+		static_cast<double>(crop.top) / full_image->size().height,
+		static_cast<double>(cropped_image->size().width) / full_image->size().width,
+		static_cast<double>(cropped_image->size().height) / full_image->size().height
+	};
+
+	only_text()->emit_bitmap(_sequence.video[_next].period, cropped_image, rectangle);
+
+	++_next;
+
+	update_position();
+	return false;
+}
+
+
+void
+FCPXMLDecoder::seek(dcpomatic::ContentTime time, bool accurate)
+{
+	/* It's worth back-tracking a little here as decoding is cheap and it's nice if we don't miss
+	   too many subtitles when seeking.
+	*/
+	time -= dcpomatic::ContentTime::from_seconds(5);
+	if (time < dcpomatic::ContentTime()) {
+		time = {};
+	}
+
+	Decoder::seek(time, accurate);
+
+	_next = 0;
+	while (_next < static_cast<int>(_sequence.video.size()) && _sequence.video[_next].period.from < time) {
+		++_next;
+	}
+
+	update_position();
+}
+
+
+void
+FCPXMLDecoder::update_position()
+{
+	if (_next < static_cast<int>(_sequence.video.size())) {
+		only_text()->maybe_set_position(_sequence.video[_next].period.from);
+	}
+}
+
diff --git a/src/lib/fcpxml_decoder.h b/src/lib/fcpxml_decoder.h
new file mode 100644
index 000000000..782693256
--- /dev/null
+++ b/src/lib/fcpxml_decoder.h
@@ -0,0 +1,50 @@
+/*
+    Copyright (C) 2024 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/>.
+
+*/
+
+
+#ifndef DCPOMATIC_FCPXML_DECODER_H
+#define DCPOMATIC_FCPXML_DECODER_H
+
+
+#include "decoder.h"
+#include "fcpxml.h"
+
+
+class FCPXMLContent;
+
+
+class FCPXMLDecoder : public Decoder
+{
+public:
+	FCPXMLDecoder(std::weak_ptr<const Film> film, std::shared_ptr<const FCPXMLContent> content);
+
+	bool pass() override;
+	void seek(dcpomatic::ContentTime time, bool accurate) override;
+
+private:
+	void update_position();
+
+	std::shared_ptr<const FCPXMLContent> _fcpxml_content;
+	dcpomatic::fcpxml::Sequence _sequence;
+	int _next = 0;
+};
+
+
+#endif
diff --git a/src/lib/wscript b/src/lib/wscript
index dfe3ce487..8a4b17dbc 100644
--- a/src/lib/wscript
+++ b/src/lib/wscript
@@ -99,6 +99,9 @@ sources = """
           examine_ffmpeg_subtitles_job.cc
           exceptions.cc
           export_config.cc
+          fcpxml.cc
+          fcpxml_content.cc
+          fcpxml_decoder.cc
           frame_info.cc
           file_group.cc
           file_log.cc
-- 
cgit v1.2.3