Add support for parsing Nero M4A chapters (#7159)

5 files changed, 234 insertions, 2 deletions

diff --git a/model/src/main/java/de/danoeh/antennapod/model/feed/Chapter.java b/model/src/main/java/de/danoeh/antennapod/model/feed/Chapter.java
index 3683a2a44..6c6647880 100644
--- a/model/src/main/java/de/danoeh/antennapod/model/feed/Chapter.java
+++ b/model/src/main/java/de/danoeh/antennapod/model/feed/Chapter.java
@@ -4,7 +4,7 @@ import java.util.List;
 
 public class Chapter {
     private long id;
-    /** Defines starting point in milliseconds. */
+    /** The start time of the chapter in milliseconds */
     private long start;
     private String title;
     private String link;
@@ -66,7 +66,7 @@ public class Chapter {
 
     @Override
     public String toString() {
-        return "ID3Chapter [title=" + getTitle() + ", start=" + getStart() + ", url=" + getLink() + "]";
+        return "Chapter [title=" + getTitle() + ", start=" + getStart() + ", url=" + getLink() + "]";
     }
 
     public long getId() {
diff --git a/parser/media/src/main/java/de/danoeh/antennapod/parser/media/m4a/M4AChapterReader.java b/parser/media/src/main/java/de/danoeh/antennapod/parser/media/m4a/M4AChapterReader.java
new file mode 100644
index 000000000..85a163fa7
--- /dev/null
+++ b/parser/media/src/main/java/de/danoeh/antennapod/parser/media/m4a/M4AChapterReader.java
@@ -0,0 +1,161 @@
+package de.danoeh.antennapod.parser.media.m4a;
+
+import android.util.Log;
+
+import org.apache.commons.io.IOUtils;
+
+import java.io.InputStream;
+import java.nio.charset.StandardCharsets;
+import java.util.ArrayList;
+import java.util.List;
+
+import de.danoeh.antennapod.model.feed.Chapter;
+
+import java.io.IOException;
+import java.nio.ByteBuffer;
+import java.nio.ByteOrder;
+
+public class M4AChapterReader {
+    private static final String TAG = "M4AChapterReader";
+    private final List<Chapter> chapters = new ArrayList<>();
+    private final InputStream inputStream;
+    private static final int FTYP_CODE = 0x66747970; // "ftyp"
+
+    public M4AChapterReader(InputStream input) {
+        inputStream = input;
+    }
+
+    /**
+     * Read the input stream populating the chapters list
+     */
+    public void readInputStream() {
+        try {
+            isM4A(inputStream);
+            int dataSize = this.findAtom("moov.udta.chpl");
+            if (dataSize == -1) {
+                Log.d(TAG, "Nero Chapter Atom not found");
+            } else {
+                Log.d(TAG, "Nero Chapter Atom found. Data Size: " + dataSize);
+                this.parseNeroChapterAtom(dataSize);
+            }
+        } catch (Exception e) {
+            Log.d(TAG, "ERROR: " + e.getMessage());
+        }
+    }
+
+    /**
+     * Find the atom with the given name in the M4A file
+     *
+     * @param name the name of the atom to find, separated by dots
+     * @return the size of the atom (minus the 8-byte header) if found
+     * @throws IOException if an I/O error occurs or the atom is not found
+     */
+    public int findAtom(String name) throws IOException {
+        // Split the name into parts encoded as UTF-8
+        String[] parts = name.split("\\.");
+        int partIndex = 0;
+        // Initialize remaining size to track the current part's size and check if it is exceeded
+        int remainingSize = -1;
+
+        // Read the M4A file atom by atom
+        ByteBuffer buffer = ByteBuffer.allocate(8).order(ByteOrder.BIG_ENDIAN);
+        while (true) {
+            // Read the atom header
+            IOUtils.readFully(inputStream, buffer.array());
+            // Get the size of the current atom
+            int chunkSize = buffer.getInt();
+            int dataSize = chunkSize - 8;
+
+            // Get the atom type
+            String atomType = StandardCharsets.UTF_8.decode(buffer).toString();
+
+            // Reset the buffer for reading the atom data
+            buffer.clear();
+
+            // Check if the current atom matches the current part of the name
+            if (atomType.equals(parts[partIndex])) {
+                if (partIndex == parts.length - 1) {
+                    // If the current atom is the last part of the name return its size
+                    return dataSize;
+                } else {
+                    // Else move to the next part of the name
+                    partIndex++;
+                    // Update the remaining size
+                    remainingSize = dataSize;
+                }
+            } else {
+                // Do not check the remaining size of top-level atoms
+                if (partIndex > 0) {
+                    // Update the remaining size
+                    remainingSize -= dataSize;
+                    // If the remaining size is exhausted, throw an exception
+                    if (remainingSize <= 0) {
+                        throw new IOException("Part size exceeded for part \"" + parts[partIndex - 1]
+                                + "\" while searching atom. Remaining Size: " + remainingSize);
+                    }
+                }
+                // Skip the rest of the atom
+                IOUtils.skipFully(inputStream, dataSize);
+            }
+        }
+    }
+
+    /**
+     * Parse the Nero Chapter Atom in the M4A file
+     * Assumes that the current position is at the start of the Nero Chapter Atom
+     *
+     * @param chunkSize the size of the Nero Chapter Atom
+     * @throws IOException if an I/O error occurs
+     * @see <a href="https://github.com/Zeugma440/atldotnet/wiki/Focus-on-Chapter-metadata#nero-chapters">Nero Chapter</a>
+     */
+    private void parseNeroChapterAtom(long chunkSize) throws IOException {
+        // Read the Nero Chapter Atom data into a buffer
+        ByteBuffer byteBuffer = ByteBuffer.allocate((int) chunkSize).order(ByteOrder.BIG_ENDIAN);
+        IOUtils.readFully(inputStream, byteBuffer.array());
+        // Skip the 5-byte header
+        // Nero Chapter Atom consists of a 5-byte header followed by chapter data
+        // The first 4 bytes are the version and flags, the 5th byte is reserved
+        byteBuffer.position(5);
+        // Get the chapter count
+        int chapterCount = byteBuffer.getInt();
+        Log.d(TAG, "Nero Chapter Count: " + chapterCount);
+
+        // Parse each chapter
+        for (int i = 0; i < chapterCount; i++) {
+            long startTime = byteBuffer.getLong();
+            int chapterNameSize = byteBuffer.get();
+            byte[] chapterNameBytes = new byte[chapterNameSize];
+            byteBuffer.get(chapterNameBytes, 0, chapterNameSize);
+            String chapterName = new String(chapterNameBytes, StandardCharsets.UTF_8);
+
+            Chapter chapter = new Chapter();
+            chapter.setStart(startTime / 10000);
+            chapter.setTitle(chapterName);
+            chapter.setChapterId(String.valueOf(i + 1));
+            chapters.add(chapter);
+
+            Log.d(TAG, "Nero Chapter " + (i + 1) + ": " + chapter);
+        }
+    }
+
+    public List<Chapter> getChapters() {
+        return chapters;
+    }
+
+    /**
+     * Assert that the input stream is an M4A file by checking the signature
+     *
+     * @param inputStream the input stream to check
+     * @throws IOException if an I/O error occurs
+     */
+    public static void isM4A(InputStream inputStream) throws IOException {
+        ByteBuffer byteBuffer = ByteBuffer.allocate(8).order(ByteOrder.BIG_ENDIAN);
+        IOUtils.readFully(inputStream, byteBuffer.array());
+
+        int ftypSize = byteBuffer.getInt();
+        if (byteBuffer.getInt() != FTYP_CODE) {
+            throw new IOException("Not an M4A file");
+        }
+        IOUtils.skipFully(inputStream, ftypSize - 8);
+    }
+}
diff --git a/parser/media/src/test/java/de/danoeh/antennapod/parser/media/m4a/M4AChapterReaderTest.java b/parser/media/src/test/java/de/danoeh/antennapod/parser/media/m4a/M4AChapterReaderTest.java
new file mode 100644
index 000000000..9111f9727
--- /dev/null
+++ b/parser/media/src/test/java/de/danoeh/antennapod/parser/media/m4a/M4AChapterReaderTest.java
@@ -0,0 +1,41 @@
+package de.danoeh.antennapod.parser.media.m4a;
+
+import de.danoeh.antennapod.model.feed.Chapter;
+import org.junit.Test;
+import org.junit.runner.RunWith;
+import org.robolectric.RobolectricTestRunner;
+
+import java.io.IOException;
+import java.io.InputStream;
+import java.util.List;
+
+import static org.junit.Assert.assertEquals;
+
+@RunWith(RobolectricTestRunner.class)
+public class M4AChapterReaderTest {
+
+    @Test
+    public void testFiles() throws IOException {
+        testFile();
+    }
+
+    public void testFile() throws IOException {
+        InputStream inputStream = getClass().getClassLoader()
+                .getResource("nero-chapters.m4a").openStream();
+        M4AChapterReader reader = new M4AChapterReader(inputStream);
+        reader.readInputStream();
+        List<Chapter> chapters = reader.getChapters();
+
+        assertEquals(4, chapters.size());
+
+        assertEquals(0, chapters.get(0).getStart());
+        assertEquals(3000, chapters.get(1).getStart());
+        assertEquals(6000, chapters.get(2).getStart());
+        assertEquals(9000, chapters.get(3).getStart());
+
+        assertEquals("Chapter 1 - ❤️😊", chapters.get(0).getTitle());
+        assertEquals("Chapter 2 - ßöÄ", chapters.get(1).getTitle());
+        assertEquals("Chapter 3 - 爱", chapters.get(2).getTitle());
+        assertEquals("Chapter 4", chapters.get(3).getTitle());
+    }
+}
diff --git a/parser/media/src/test/resources/nero-chapters.m4a b/parser/media/src/test/resources/nero-chapters.m4a
new file mode 100644
index 000000000..99de2b661
--- /dev/null
+++ b/parser/media/src/test/resources/nero-chapters.m4a
diff --git a/ui/chapters/src/main/java/de/danoeh/antennapod/ui/chapters/ChapterUtils.java b/ui/chapters/src/main/java/de/danoeh/antennapod/ui/chapters/ChapterUtils.java
index 5554890ed..59317473a 100644
--- a/ui/chapters/src/main/java/de/danoeh/antennapod/ui/chapters/ChapterUtils.java
+++ b/ui/chapters/src/main/java/de/danoeh/antennapod/ui/chapters/ChapterUtils.java
@@ -16,6 +16,7 @@ import de.danoeh.antennapod.parser.media.id3.ID3ReaderException;
 import de.danoeh.antennapod.model.playback.Playable;
 import de.danoeh.antennapod.parser.media.vorbis.VorbisCommentChapterReader;
 import de.danoeh.antennapod.parser.media.vorbis.VorbisCommentReaderException;
+import de.danoeh.antennapod.parser.media.m4a.M4AChapterReader;
 import okhttp3.CacheControl;
 import okhttp3.Request;
 import okhttp3.Response;
@@ -106,6 +107,19 @@ public class ChapterUtils {
         } catch (IOException | VorbisCommentReaderException e) {
             Log.e(TAG, "Unable to load vorbis chapters: " + e.getMessage());
         }
+
+        try (CountingInputStream in = openStream(playable, context)) {
+            List<Chapter> chapters = readM4AChaptersFromInputStream(in);
+            if (!chapters.isEmpty()) {
+                Log.i(TAG, "Chapters loaded");
+                return chapters;
+            }
+        } catch (InterruptedIOException e) {
+            throw e;
+        } catch (IOException e) {
+            Log.e(TAG, "Unable to open stream " + e.getMessage());
+        }
+
         return null;
     }
 
@@ -195,6 +209,22 @@ public class ChapterUtils {
         return Collections.emptyList();
     }
 
+    @NonNull
+    private static List<Chapter> readM4AChaptersFromInputStream(InputStream input) {
+        M4AChapterReader reader = new M4AChapterReader(new BufferedInputStream(input));
+        reader.readInputStream();
+        List<Chapter> chapters = reader.getChapters();
+        if (chapters == null) {
+            return Collections.emptyList();
+        }
+        Collections.sort(chapters, new ChapterStartTimeComparator());
+        enumerateEmptyChapterTitles(chapters);
+        if (chaptersValid(chapters)) {
+            return chapters;
+        }
+        return Collections.emptyList();
+    }
+
     /**
      * Makes sure that chapter does a title and an item attribute.
      */