support for reading chapters from ffmpeg metadata files

implements #3676

NEWS.md

@@ -4,6 +4,8 @@
 
 * translations: added a Belarusian translation of the programs & the man pages
   by prydespar (see `AUTHORS`).
+* mkvmerge, MKVToolNix GUI's chapter editor: added support for reading
+  chapters from ffmpeg metadata files. Implements #3676.
 
 
 # Version 82.0 "I'm The President" 2024-01-02

src/common/chapters/chapters.cpp

@@ -16,6 +16,7 @@
 
 #include <algorithm>
 #include <cassert>
+#include <optional>
 
 #include <QRegularExpression>
 
@@ -29,6 +30,7 @@
 #include "common/error.h"
 #include "common/iso3166.h"
 #include "common/locale.h"
+#include "common/math_fwd.h"
 #include "common/mm_io_x.h"
 #include "common/mm_file_io.h"
 #include "common/mm_proxy_io.h"
@@ -350,6 +352,146 @@ parse_simple(mm_text_io_c *in,
   return 0 == num ? kax_cptr{} : chaps;
 }
 
+bool
+probe_ffmpeg_meta(mm_text_io_c *in) {
+  std::string line;
+
+  in->setFilePointer(0);
+
+  if (!in->getline2(line))
+    return false;
+
+  mtx::string::strip(line);
+
+  return line == ";FFMETADATA1";
+}
+
+kax_cptr
+parse_ffmpeg_meta(mm_text_io_c *in,
+                  int64_t min_ts,
+                  int64_t max_ts,
+                  int64_t offset,
+                  mtx::bcp47::language_c const &language,
+                  std::string const &charset) {
+  in->setFilePointer(0);
+
+  std::string line, title;
+  std::optional<int64_t> start_scaled, end_scaled;
+  mtx_mp_rational_t time_base;
+  kax_cptr chapters;
+  libmatroska::KaxEditionEntry *edition{};
+  charset_converter_cptr cc_utf8;
+  bool in_chapter = false;
+  bool do_convert = in->get_byte_order_mark() == byte_order_mark_e::none;
+
+  if (do_convert)
+    cc_utf8 = charset_converter_c::init(charset);
+
+  auto use_language = language.is_valid()           ? language
+                    : g_default_language.is_valid() ? g_default_language
+                    :                                 mtx::bcp47::language_c::parse("eng"s);
+
+  QRegularExpression
+    start_line_re{     Q("^start *=([0-9]+)"),             QRegularExpression::CaseInsensitiveOption},
+    end_line_re{       Q("^end *=([0-9]+)"),               QRegularExpression::CaseInsensitiveOption},
+    title_line_re{     Q("^title *=(.*)"),                 QRegularExpression::CaseInsensitiveOption},
+    time_base_line_re{ Q("^timebase *=([0-9]+)/([0-9]+)"), QRegularExpression::CaseInsensitiveOption};
+  QRegularExpressionMatch matches;
+
+  auto reset_values = [&]() {
+    start_scaled.reset();
+    end_scaled.reset();
+    title.clear();
+
+    time_base = mtx::rational(1, 1);
+  };
+
+  auto add_chapter_atom = [&]() {
+    if (!start_scaled || !in_chapter) {
+      reset_values();
+      return;
+    }
+
+    auto start = mtx::to_int_rounded(*start_scaled * time_base);
+
+    if ((start < min_ts) || ((max_ts != -1) && (start > max_ts))) {
+      reset_values();
+      return;
+    }
+
+    if (!chapters) {
+      chapters   = std::make_shared<libmatroska::KaxChapters>();
+      edition = &get_child<libmatroska::KaxEditionEntry>(*chapters);
+    }
+
+    auto &atom = add_empty_child<libmatroska::KaxChapterAtom>(*edition);
+    get_child<libmatroska::KaxChapterUID>(atom).SetValue(create_unique_number(UNIQUE_CHAPTER_IDS));
+    get_child<libmatroska::KaxChapterTimeStart>(atom).SetValue(start - offset);
+
+    if (end_scaled) {
+      auto end = mtx::to_int_rounded(*end_scaled * time_base);
+      get_child<libmatroska::KaxChapterTimeEnd>(atom).SetValue(end - offset);
+    }
+
+    auto &display = get_child<libmatroska::KaxChapterDisplay>(atom);
+    get_child<libmatroska::KaxChapterString>(display).SetValueUTF8(title);
+
+    if (use_language.is_valid()) {
+      get_child<libmatroska::KaxChapterLanguage>(display).SetValue(use_language.get_closest_iso639_2_alpha_3_code());
+      if (!mtx::bcp47::language_c::is_disabled())
+        get_child<libmatroska::KaxChapLanguageIETF>(display).SetValue(use_language.format());
+      else
+        delete_children<libmatroska::KaxChapLanguageIETF>(display);
+    }
+
+    if (!g_default_country.empty())
+      get_child<libmatroska::KaxChapterCountry>(display).SetValue(g_default_country);
+
+    reset_values();
+  };
+
+  while (in->getline2(line)) {
+    if (do_convert)
+      line = cc_utf8->utf8(line);
+
+    mtx::string::strip(line);
+    if (line.empty() || (line[0] == ';') || (line[0] == '#'))
+      continue;
+
+    if (mtx::string::to_lower_ascii(line) == "[chapter]") {
+      add_chapter_atom();
+      in_chapter = true;
+
+    } else if (line[0] == '[') {
+      add_chapter_atom();
+      in_chapter = false;
+
+    } else if (!in_chapter)
+      continue;
+
+    else if ((matches = title_line_re.match(Q(line))).hasMatch())
+      title = to_utf8(matches.captured(1));
+
+    else if ((matches = start_line_re.match(Q(line))).hasMatch())
+      start_scaled = matches.captured(1).toLongLong();
+
+    else if ((matches = end_line_re.match(Q(line))).hasMatch())
+      end_scaled = matches.captured(1).toLongLong();
+
+    else if ((matches = time_base_line_re.match(Q(line))).hasMatch()) {
+      auto numerator   = matches.captured(1).toLongLong();
+      auto deniminator = matches.captured(2).toLongLong();
+
+      if ((numerator != 0) && (deniminator != 0))
+        time_base = mtx::rational(numerator, deniminator) * 1'000'000'000;
+    }
+  }
+
+  add_chapter_atom();
+
+  return chapters;
+}
+
 /** \brief Probe a file for different chapter formats and parse the file.
 
    The file \a file_name is opened and checked for supported chapter formats.
@@ -483,6 +625,11 @@ parse(mm_text_io_c *in,
         *format = format_e::cue;
       return parse_cue(in, min_ts, max_ts, offset, language, charset, tags);
 
+    } else if (probe_ffmpeg_meta(in)) {
+      if (format)
+        *format = format_e::ffmpeg_meta;
+      return parse_ffmpeg_meta(in, min_ts, max_ts, offset, language, charset);
+
     } else if (format)
       *format = format_e::xml;
 

src/common/chapters/chapters.h

@@ -54,6 +54,7 @@ enum class format_e {
   xml,
   ogg,
   cue,
+  ffmpeg_meta,
 };
 
 mtx::chapters::kax_cptr

src/mkvtoolnix-gui/chapter_editor/tool.cpp

@@ -193,7 +193,7 @@ Tool::selectFileToOpen(bool append) {
 #endif  // HAVE_DVDREAD
 
   auto fileNames = Util::getOpenFileNames(this, append ? QY("Append files in chapter editor") : QY("Open files in chapter editor"), Util::Settings::get().lastOpenDirPath(),
-                                          QY("Supported file types")           + Q(" (*.cue%1 *.mpls *.mkv *.mka *.mks *.mk3d *.txt *.webm *.xml);;").arg(ifo) +
+                                          QY("Supported file types")           + Q(" (*.cue%1 *.meta *.mpls *.mkv *.mka *.mks *.mk3d *.txt *.webm *.xml);;").arg(ifo) +
                                           QY("Matroska files")                 + Q(" (*.mkv *.mka *.mks *.mk3d);;") +
                                           QY("WebM files")                     + Q(" (*.webm);;") +
                                           QY("Blu-ray playlist files")         + Q(" (*.mpls);;") +
@@ -201,6 +201,7 @@ Tool::selectFileToOpen(bool append) {
                                           QY("XML chapter files")              + Q(" (*.xml);;") +
                                           QY("Simple OGM-style chapter files") + Q(" (*.txt);;") +
                                           QY("Cue sheet files")                + Q(" (*.cue);;") +
+                                          QY("ffmpeg metadata files")          + Q(" (*.meta);;") +
                                           QY("All files")                      + Q(" (*)"));
   if (fileNames.isEmpty())
     return;

src/mkvtoolnix-gui/merge/file_identification_thread.cpp

@@ -26,7 +26,7 @@ class FileIdentificationWorkerPrivate {
   QVector<IdentificationPack> m_toIdentify;
   QMutex m_mutex;
   QAtomicInteger<bool> m_abortPlaylistScan;
-  QRegularExpression m_simpleChaptersRE, m_xmlChaptersRE, m_xmlSegmentInfoRE, m_xmlTagsRE;
+  QRegularExpression m_simpleChaptersRE, m_ffmpegMetaChaptersRE, m_xmlChaptersRE, m_xmlSegmentInfoRE, m_xmlTagsRE;
 
   explicit FileIdentificationWorkerPrivate()
   {
@@ -39,11 +39,12 @@ FileIdentificationWorker::FileIdentificationWorker(QObject *parent)
   : QObject{parent}
   , p_ptr{new FileIdentificationWorkerPrivate{}}
 {
-  auto p                = p_func();
+  auto p                    = p_func();
-  p->m_simpleChaptersRE = QRegularExpression{R"(^CHAPTER\d{2}=[\s\S]*CHAPTER\d{2}NAME=)"};
+  p->m_simpleChaptersRE     = QRegularExpression{R"(^CHAPTER\d{2}=[\s\S]*CHAPTER\d{2}NAME=)"};
-  p->m_xmlChaptersRE    = QRegularExpression{R"(^(<!--.*?-->\s*)*<\?xml[^>]+version[\s\S]*?\?>[\s\S]*?<Chapters>)"};
+  p->m_xmlChaptersRE        = QRegularExpression{R"(^(<!--.*?-->\s*)*<\?xml[^>]+version[\s\S]*?\?>[\s\S]*?<Chapters>)"};
-  p->m_xmlSegmentInfoRE = QRegularExpression{R"(^(<!--.*?-->\s*)*<\?xml[^>]+version[\s\S]*?\?>[\s\S]*?<Info>)"};
+  p->m_xmlSegmentInfoRE     = QRegularExpression{R"(^(<!--.*?-->\s*)*<\?xml[^>]+version[\s\S]*?\?>[\s\S]*?<Info>)"};
-  p->m_xmlTagsRE        = QRegularExpression{R"(^(<!--.*?-->\s*)*<\?xml[^>]+version[\s\S]*?\?>[\s\S]*?<Tags>)"};
+  p->m_xmlTagsRE            = QRegularExpression{R"(^(<!--.*?-->\s*)*<\?xml[^>]+version[\s\S]*?\?>[\s\S]*?<Tags>)"};
+  p->m_ffmpegMetaChaptersRE = QRegularExpression{R"(;FFMETADATA1)"};
 }
 
 FileIdentificationWorker::~FileIdentificationWorker() {
@@ -179,7 +180,7 @@ FileIdentificationWorker::determineIfFileThatShouldBeSelectedElsewhere(QString c
 
   auto content = QString::fromUtf8(bytes);
 
-  if (content.contains(p->m_simpleChaptersRE) || content.contains(p->m_xmlChaptersRE))
+  if (content.contains(p->m_simpleChaptersRE) || content.contains(p->m_ffmpegMetaChaptersRE) || content.contains(p->m_xmlChaptersRE))
     return IdentificationPack::FileType::Chapters;
 
   else if (content.contains(p->m_xmlSegmentInfoRE))

tests/results.txt

@@ -609,3 +609,4 @@ T_0761hevc_dolby_vision_dual_layer_single_track_annex_b:4a40b17c0d60a1c82c109975
 T_0762dovi_combining_bl_and_el:880014a766cf6b5929cf180bea71d046-2eeed24f6b3e3f2d6da0e8d25896089b-88193925316d397c23c6a3a6cb2ab740-ef854e7bae7c85ad46b1a0d71d09d3fc:passed:20231129-202129:0.387596229
 T_0763vp9_alpha_channel_data:ac329c9d810d9b19fb3ffcde6f12af2a-OK:passed:20231203-190856:0.059109676
 T_0764ui_locale_be_BY:a44c54eadfb4c8fbdc104b75aa1de1c1-72b98d331b58a0f95e10159fca191b52:passed:20240120-191944:0.043782405
+T_0765ffmpeg_metadata_chapters:f16630c4019413c98b75b959a5697391-6b2b843310e80367b5fe5aaa8a5d51c4:passed:20240310-145016:0.047790171

tests/test-0765ffmpeg_metadata_chapters.rb

@@ -0,0 +1,8 @@
+#!/usr/bin/ruby -w
+
+# T_765ffmpeg_metadata_chapters
+describe "mkvmerge / reading chapters from ffmpeg metadata files"
+
+Dir.glob("data/chapters/ffmpeg-metadata/*.meta").sort.each do |file_name|
+  test_merge "data/subtitles/srt/ven.srt", :args => "--chapters #{file_name}"
+end