Add PDF formatter for verification reports (#1823).v1.10.28

6 files changed, 411 insertions, 3 deletions

diff --git a/src/pdf_formatter.cc b/src/pdf_formatter.cc
new file mode 100644
index 00000000..5eb1772c
--- /dev/null
+++ b/src/pdf_formatter.cc
@@ -0,0 +1,286 @@
+/*
+    Copyright (C) 2025 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 "exceptions.h"
+#include "pdf_formatter.h"
+#include <fmt/format.h>
+
+
+using std::unique_ptr;
+using namespace dcp;
+
+
+int constexpr dpi = 72;
+
+/* These are in inches */
+float constexpr page_width = 8.27;
+float constexpr horizontal_margin = 0.5;
+float constexpr page_height = 11.69;
+float constexpr vertical_margin = 1.0;
+
+
+static
+void
+error_handler(HPDF_STATUS error, HPDF_STATUS detail, void*)
+{
+	throw MiscError(fmt::format("Could not create PDF {} {}", error, detail));
+}
+
+
+PDFFormatter::PDFFormatter(boost::filesystem::path file)
+	: _file(file)
+	, _pdf(HPDF_New(error_handler, nullptr))
+	, _normal_font(HPDF_GetFont(_pdf, "Helvetica", nullptr))
+	, _fixed_font(HPDF_GetFont(_pdf, "Courier", nullptr))
+	, _bold_font(HPDF_GetFont(_pdf, "Helvetica-Bold", nullptr))
+{
+	add_page();
+}
+
+
+PDFFormatter::~PDFFormatter()
+{
+	HPDF_Free(_pdf);
+}
+
+
+void
+PDFFormatter::add_page()
+{
+	_page = HPDF_AddPage(_pdf);
+	HPDF_Page_Concat(_page, 1, 0, 0, 1, horizontal_margin * dpi, (page_height - vertical_margin) * dpi);
+	_y = 0;
+}
+
+
+void
+PDFFormatter::heading(std::string const& text)
+{
+	move_down(16 * 1.4);
+	HPDF_Page_SetFontAndSize(_page, _bold_font, 20);
+	HPDF_Page_SetRGBFill(_page, 0, 0, 0);
+	HPDF_Page_BeginText(_page);
+	HPDF_Page_TextOut(_page, 0, _y, text.c_str());
+	HPDF_Page_EndText(_page);
+	move_down(20 * 1.4);
+}
+
+
+void
+PDFFormatter::subheading(std::string const& text)
+{
+	move_down(12 * 1.4);
+	HPDF_Page_SetFontAndSize(_page, _bold_font, 16);
+	HPDF_Page_SetRGBFill(_page, 0, 0, 0);
+	HPDF_Page_BeginText(_page);
+	HPDF_Page_TextOut(_page, 0, _y, text.c_str());
+	HPDF_Page_EndText(_page);
+	move_down(16 * 1.4);
+}
+
+
+unique_ptr<Formatter::Wrap>
+PDFFormatter::unordered_list()
+{
+	return unique_ptr<PDFFormatter::Wrap>(new PDFFormatter::Wrap(this));
+}
+
+
+void
+PDFFormatter::move_down(float spacing)
+{
+	_y -= spacing;
+	if (_y < ((-page_height + vertical_margin * 2) * dpi)) {
+		add_page();
+	}
+}
+
+
+void
+PDFFormatter::wrapped_text(float x, float first_line_indent, float font_size, dcp::Colour colour, std::string const& text, float width, float line_spacing)
+{
+	class Block
+	{
+	public:
+		enum class Style {
+			NORMAL,
+			FIXED
+		};
+
+		Block(std::string text_, Style style_)
+			: text(text_)
+			, style(style_)
+		{}
+
+		std::string text;
+		Style style;
+	};
+
+	std::map<std::string, Block::Style> tags = {
+		{ "[fixed]", Block::Style::FIXED },
+		{ "[/fixed]", Block::Style::NORMAL }
+	};
+
+	std::vector<Block> blocks;
+	Block block("", Block::Style::NORMAL);
+	auto work = text;
+	while (true) {
+		std::pair<std::string, Block::Style> found_tag;
+		auto found_start = std::string::npos;
+
+		for (auto const& tag: tags) {
+			auto const start = work.find(tag.first);
+			if (start < found_start) {
+				found_tag = tag;
+				found_start = start;
+			}
+		}
+
+		if (found_start == std::string::npos) {
+			break;
+		}
+
+		if (found_start > 0) {
+			block.text += work.substr(0, found_start);
+			blocks.push_back(block);
+			block = Block("", found_tag.second);
+		}
+		work = work.substr(found_start + found_tag.first.length());
+	}
+
+	if (!work.empty()) {
+		block.text += work;
+		blocks.push_back(block);
+	}
+
+	std::map<Block::Style, HPDF_Font*> fonts = {
+		{ Block::Style::NORMAL, &_normal_font },
+		{ Block::Style::FIXED, &_fixed_font }
+	};
+
+	auto px = x + first_line_indent;
+	for (auto const& block: blocks) {
+
+		int offset = 0;
+		int left = block.text.length();
+		while (left) {
+			HPDF_Page_SetFontAndSize(_page, *fonts[block.style], font_size);
+
+			HPDF_REAL text_width;
+			int fits = HPDF_Font_MeasureText(
+				*fonts[block.style],
+				reinterpret_cast<unsigned char const*>(block.text.substr(offset).c_str()),
+				block.text.length() - offset,
+				width - px,
+				font_size,
+				0,
+				0,
+				true,
+				&text_width);
+
+			if (fits == 0) {
+				/* Try again without word-wrap */
+				fits = HPDF_Font_MeasureText(
+					*fonts[block.style],
+					reinterpret_cast<unsigned char const*>(block.text.substr(offset).c_str()),
+					block.text.length() - offset,
+					width - px,
+					font_size,
+					0,
+					0,
+					false,
+					&text_width);
+			}
+
+			HPDF_Page_SetRGBFill(_page, colour.r / 255.0, colour.g / 255.0, colour.b / 255.0);
+			HPDF_Page_BeginText(_page);
+			HPDF_Page_TextOut(_page, px, _y, block.text.substr(offset, fits).c_str());
+			HPDF_Page_EndText(_page);
+			if (fits < static_cast<int>(block.text.length()) - offset) {
+				px = x;
+				move_down(line_spacing);
+			} else {
+				px += text_width;
+			}
+
+			offset += fits;
+			left -= fits;
+		}
+	}
+
+	move_down(line_spacing);
+}
+
+
+void
+PDFFormatter::list_item(std::string const& text, boost::optional<std::string> type)
+{
+	float const indent = 16 * _indent;
+	float constexpr dot_radius = 1.5;
+	float constexpr font_size = 10;
+
+	HPDF_Page_Circle(_page, indent + dot_radius, _y + font_size / 3, dot_radius);
+	HPDF_Page_Fill(_page);
+
+	dcp::Colour colour(0, 0, 0);
+	if (type && *type == "ok") {
+		colour = dcp::Colour(0, 153, 0);
+	} else if (type && *type == "warning") {
+		colour = dcp::Colour(255, 127, 102);
+	} else if (type && (*type == "error" || *type == "bv21-error")) {
+		colour = dcp::Colour(153, 0, 0);
+	}
+
+	wrapped_text(indent, dot_radius * 6, font_size, colour, text, (page_width - horizontal_margin * 2) * dpi, font_size * 1.2);
+}
+
+
+std::function<std::string (std::string)>
+PDFFormatter::process_string()
+{
+	return [](std::string s) {
+		return s;
+	};
+}
+
+
+std::function<std::string (std::string)>
+PDFFormatter::fixed_width()
+{
+	return [](std::string s) {
+		return String::compose("[fixed]%1[/fixed]", s);
+	};
+}
+
+
+void
+PDFFormatter::finish()
+{
+	HPDF_SaveToFile(_pdf, _file.string().c_str());
+}
+
+
+void
+PDFFormatter::indent(int amount)
+{
+	_indent += amount;
+}
+
diff --git a/src/pdf_formatter.h b/src/pdf_formatter.h
new file mode 100644
index 00000000..b7d0213d
--- /dev/null
+++ b/src/pdf_formatter.h
@@ -0,0 +1,81 @@
+/*
+    Copyright (C) 2025 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 "verify_report.h"
+#include <hpdf.h>
+
+
+namespace dcp {
+
+
+class PDFFormatter : public Formatter
+{
+public:
+	PDFFormatter(boost::filesystem::path file);
+	~PDFFormatter();
+
+	class Wrap : public Formatter::Wrap
+	{
+	public:
+		Wrap(PDFFormatter* formatter)
+			: _formatter(formatter)
+		{
+			_formatter->indent(1);
+		}
+
+		~Wrap()
+		{
+			_formatter->indent(-1);
+		}
+
+	private:
+		PDFFormatter* _formatter;
+	};
+
+	void heading(std::string const& text) override;
+	void subheading(std::string const& text) override;
+	std::unique_ptr<Formatter::Wrap> unordered_list() override;
+	void list_item(std::string const& text, boost::optional<std::string> type = {}) override;
+	std::function<std::string (std::string)> process_string() override;
+	std::function<std::string (std::string)> fixed_width() override;
+	void finish() override;
+
+	void indent(int amount);
+
+private:
+	void add_page();
+	void move_down(float spacing);
+	void wrapped_text(float x, float first_line_indent, float font_size, dcp::Colour colour, std::string const& text, float width, float line_spacing);
+
+	boost::filesystem::path _file;
+	HPDF_Doc _pdf;
+	HPDF_Page _page;
+	float _y;
+	HPDF_Font _normal_font;
+	HPDF_Font _fixed_font;
+	HPDF_Font _bold_font;
+
+	int _indent = 0;
+};
+
+
+}
+
diff --git a/src/wscript b/src/wscript
index 3e3b75b8..6176ab53 100644
--- a/src/wscript
+++ b/src/wscript
@@ -265,6 +265,11 @@ def build(bld):
  
     uselib = 'BOOST_FILESYSTEM BOOST_SIGNALS2 BOOST_DATETIME OPENSSL SIGC++ LIBXML++ OPENJPEG CXML XMLSEC1 ASDCPLIB_DCPOMATIC XERCES AVCODEC AVUTIL FMT FAST_FLOAT'
 
+    if bld.env.LIBDCP_HAVE_HARU:
+        source += "pdf_formatter.cc "
+        headers += "pdf_formatter.h "
+        uselib += ' HARU'
+
     # Main library
     if bld.env.STATIC:
         obj = bld(features='cxx cxxstlib')
diff --git a/tools/dcpverify.cc b/tools/dcpverify.cc
index bff3fde3..8966c3ca 100644
--- a/tools/dcpverify.cc
+++ b/tools/dcpverify.cc
@@ -36,6 +36,9 @@
 #include "compose.hpp"
 #include "filesystem.h"
 #include "html_formatter.h"
+#ifdef LIBDCP_HAVE_HARU
+#include "pdf_formatter.h"
+#endif
 #include "raw_convert.h"
 #include "text_formatter.h"
 #include "verify.h"
@@ -72,7 +75,12 @@ help (string n)
 	     << "  --no-asset-hash-check                        don't check asset hashes\n"
 	     << "  --asset-hash-check-maximum-size <size-in-MB> only check hashes for assets smaller than this size (in MB)\n"
 	     << "  --no-picture-details-check                   don't check details of picture assets (J2K bitstream etc.)\n"
-	     << "  -o <filename>                                write HTML report to filename\n"
+	     << "  -o <filename>                                write report to filename "
+#ifdef LIBDCP_HAVE_HARU
+	" (.txt, .htm, .html or .pdf)\n"
+#else
+	" (.txt, .htm or .html)\n"
+#endif
 	     << "  -q, --quiet                                  don't report progress\n";
 }
 
@@ -228,8 +236,18 @@ main (int argc, char* argv[])
 	}
 
 	if (report_filename) {
-		dcp::HTMLFormatter formatter(*report_filename);
-		dcp::verify_report({ result }, formatter);
+		if (report_filename->extension() == ".htm" || report_filename->extension() == ".html") {
+			dcp::HTMLFormatter formatter(*report_filename);
+			dcp::verify_report({ result }, formatter);
+#ifdef LIBDCP_HAVE_HARU
+		} else if (report_filename->extension() == ".pdf") {
+			dcp::PDFFormatter formatter(*report_filename);
+			dcp::verify_report({ result }, formatter);
+#endif
+		} else {
+			dcp::TextFormatter formatter(*report_filename);
+			dcp::verify_report({ result }, formatter);
+		}
 	}
 
 	exit (failed ? EXIT_FAILURE : EXIT_SUCCESS);
diff --git a/tools/wscript b/tools/wscript
index a5fcf545..10192f50 100644
--- a/tools/wscript
+++ b/tools/wscript
@@ -33,6 +33,8 @@
 
 def build(bld):
     uselib = 'OPENJPEG CXML OPENMP ASDCPLIB_DCPOMATIC BOOST_FILESYSTEM LIBXML++ XMLSEC1 OPENSSL XERCES DL MAGICK AVCODEC AVUTIL FMT FAST_FLOAT'
+    if bld.env.LIBDCP_HAVE_HARU:
+        uselib += ' HARU'
 
     for f in ['diff', 'info', 'verify']:
         obj = bld(features='cxx cxxprogram')
diff --git a/wscript b/wscript
index c162011b..052bb014 100644
--- a/wscript
+++ b/wscript
@@ -271,6 +271,20 @@ def configure(conf):
     conf.check_cfg(package='fmt', args='--cflags --libs', uselib_store='FMT', mandatory=True)
     conf.check_cxx(header_name="fast_float/fast_float.h", uselib_store='FAST_FLOAT', mandatory=True)
 
+    haru = conf.check_cc(fragment="""
+                         #include "hpdf.h"\n
+                         int main() { HPDF_Doc pdf; return 0; }\n
+                         """,
+                         msg='Checking for haru library',
+                         libpath='/usr/local/lib',
+                         lib=['hpdf'],
+                         uselib_store='HARU',
+                         define_name='LIBDCP_HAVE_HARU',
+                         mandatory=False)
+
+    if haru:
+        conf.env.append_value('LIBDCP_HAVE_HARU', '1')
+
     if not conf.env.DISABLE_TESTS:
         conf.recurse('test')
         if conf.options.enable_gcov:
@@ -288,6 +302,8 @@ def build(bld):
         boost_lib_suffix = ''
 
     libs="-L${libdir} -ldcp%s -lcxml -lboost_system%s" % (bld.env.API_VERSION, boost_lib_suffix)
+    if bld.env.HAVE_HARU:
+        libs += " -lhpdf"
     if bld.env.TARGET_LINUX:
         libs += " -ldl"