OggAudioParser.java

/*
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License.  You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */
package org.apache.tika.parser.ogg;

import java.io.IOException;
import java.nio.BufferUnderflowException;
import java.nio.ByteBuffer;
import java.nio.charset.Charset;
import java.nio.charset.StandardCharsets;
import java.text.DecimalFormat;
import java.text.NumberFormat;
import java.util.Arrays;
import java.util.Base64;
import java.util.List;
import java.util.Locale;

import org.gagravarr.ogg.audio.OggAudioHeaders;
import org.gagravarr.ogg.audio.OggAudioInfoHeader;
import org.gagravarr.ogg.audio.OggAudioStatistics;
import org.gagravarr.ogg.audio.OggAudioStream;
import org.gagravarr.vorbis.VorbisComments;
import org.gagravarr.vorbis.VorbisStyleComments;
import org.xml.sax.SAXException;

import org.apache.tika.exception.TikaException;
import org.apache.tika.extractor.EmbeddedDocumentExtractor;
import org.apache.tika.extractor.EmbeddedDocumentUtil;
import org.apache.tika.io.TikaInputStream;
import org.apache.tika.metadata.Audio;
import org.apache.tika.metadata.HttpHeaders;
import org.apache.tika.metadata.KeyPrefix;
import org.apache.tika.metadata.Metadata;
import org.apache.tika.metadata.Property;
import org.apache.tika.metadata.TikaCoreProperties;
import org.apache.tika.metadata.XMP;
import org.apache.tika.metadata.XMPDM;
import org.apache.tika.parser.AbstractParser;
import org.apache.tika.parser.ParseContext;
import org.apache.tika.parser.audio.NumberAndTotal;
import org.apache.tika.parser.mp3.ID3Tags;
import org.apache.tika.sax.XHTMLContentHandler;

/**
 * Parent parser for the various Ogg Audio formats, such as
 * Vorbis and Opus.
 */
public abstract class OggAudioParser extends AbstractParser {
    private static final long serialVersionUID = 5168743829615945633L;

    private static final KeyPrefix VORBIS =
            KeyPrefix.file("vorbis:", "Vorbis comment field names");

    /**
     * The Vorbis comment header vendor string (encoder library identification), also
     * captured under {@link org.apache.tika.metadata.XMP#CREATOR_TOOL}; kept under its
     * own name too since some consumers look for the raw vorbis: field.
     */
    private static final Property VORBIS_VENDOR = Property.internalText("vorbis:vendor");

    // Codec bitstream/library version string (e.g. "Theora 3.2.1"); distinct from the vendor/encoder tool under XMP#CREATOR_TOOL.
    protected static final Property CODEC_VERSION = Property.internalText("ogg:codec-version");

    /**
     * Comment holding an embedded picture (e.g. cover art) as a base64
     * encoded FLAC picture block
     */
    private static final String METADATA_BLOCK_PICTURE = "metadata_block_picture";


    /**
     * Returns the first positive integer found under the given comment keys,
     * or null if there is none.
     */
    private static Integer firstPositiveInteger(VorbisStyleComments comments, String... keys) {
        for (String key : keys) {
            for (String value : comments.getComments(key)) {
                try {
                    int parsed = Integer.parseInt(value.trim());
                    if (parsed > 0) {
                        return parsed;
                    }
                } catch (NumberFormatException e) {
                    //skip unparseable values
                }
            }
        }
        return null;
    }

    protected static void extractChannelInfo(Metadata metadata, OggAudioInfoHeader info) {
        extractChannelInfo(metadata, info.getNumChannels());
    }

    protected static void extractChannelInfo(Metadata metadata, int channelCount) {
        if (channelCount == 1) {
            metadata.set(XMPDM.AUDIO_CHANNEL_TYPE, "Mono");
        } else if (channelCount == 2) {
            metadata.set(XMPDM.AUDIO_CHANNEL_TYPE, "Stereo");
        } else if (channelCount == 5) {
            metadata.set(XMPDM.AUDIO_CHANNEL_TYPE, "5.1");
        } else if (channelCount == 7) {
            metadata.set(XMPDM.AUDIO_CHANNEL_TYPE, "7.1");
        }
    }

    protected static void extractComments(Metadata metadata, XHTMLContentHandler xhtml,
            VorbisStyleComments comments, ParseContext context)
            throws IOException, TikaException, SAXException {
        // Get the specific known comments
        metadata.set(TikaCoreProperties.TITLE, comments.getTitle());
        metadata.set(TikaCoreProperties.CREATOR, comments.getArtist());
        metadata.set(XMPDM.ARTIST, comments.getArtist());
        metadata.set(XMPDM.ALBUM, comments.getAlbum());
        metadata.set(XMPDM.GENRE, comments.getGenre());
        metadata.set(XMPDM.RELEASE_DATE, comments.getDate());
        metadata.add(XMP.CREATOR_TOOL, comments.getVendor());
        metadata.add(VORBIS_VENDOR, comments.getVendor());

        //xmpDM:copyright is single-valued, so map the first comment; like
        //vendor, the raw comments also stay available under the vorbis: name
        List<String> copyrights = comments.getComments("copyright");
        if (!copyrights.isEmpty()) {
            metadata.set(XMPDM.COPYRIGHT, copyrights.get(0));
        }

        for (String comment : comments.getComments("comment")) {
            metadata.add(XMPDM.LOG_COMMENT, comment);
        }

        // Grab the rest just in case; the pictures become embedded
        //  documents instead, their raw base64 blocks help nobody
        List<String> done = Arrays.asList(
                VorbisComments.KEY_TITLE, VorbisComments.KEY_ARTIST,
                VorbisComments.KEY_ALBUM, VorbisComments.KEY_GENRE,
                VorbisComments.KEY_DATE, VorbisComments.KEY_TRACKNUMBER,
                "vendor", "comment", METADATA_BLOCK_PICTURE
        );
        // BAG: a Vorbis comment field can legitimately repeat.
        for (String key : comments.getAllComments().keySet()) {
            if (!done.contains(key)) {
                for (String value : comments.getAllComments().get(key)) {
                    metadata.add(VORBIS, key, value);
                }
            }
        }

        // Output as text too
        xhtml.element("h1", comments.getTitle());
        xhtml.element("p", comments.getArtist());

        // Album and Track number
        if (comments.getTrackNumber() != null) {
            xhtml.element("p", comments.getAlbum() + ", track " + comments.getTrackNumber());
            metadata.set(Audio.RAW_TRACK_NUMBER, comments.getTrackNumber());
            NumberAndTotal trackNumberAndTotal = NumberAndTotal.parse(comments.getTrackNumber());
            if (trackNumberAndTotal != null) {
                if (trackNumberAndTotal.number != null) {
                    metadata.set(XMPDM.TRACK_NUMBER, trackNumberAndTotal.number);
                }
                if (trackNumberAndTotal.total != null) {
                    metadata.set(Audio.TRACK_COUNT, trackNumberAndTotal.total);
                }
            }
        } else {
            xhtml.element("p", comments.getAlbum());
        }
        for (String discValue : comments.getComments("discnumber")) {
            metadata.set(Audio.RAW_DISC_NUMBER, discValue);
            NumberAndTotal discNumberAndTotal = NumberAndTotal.parse(discValue);
            if (discNumberAndTotal != null) {
                if (discNumberAndTotal.number != null) {
                    metadata.set(XMPDM.DISC_NUMBER, discNumberAndTotal.number);
                }
                if (discNumberAndTotal.total != null) {
                    metadata.set(Audio.DISC_COUNT, discNumberAndTotal.total);
                }
            }
        }
        //explicit totals win over the combined "n/total" form
        Integer trackTotal = firstPositiveInteger(comments, "tracktotal", "totaltracks");
        if (trackTotal != null) {
            metadata.set(Audio.TRACK_COUNT, trackTotal);
        }
        Integer discTotal = firstPositiveInteger(comments, "disctotal", "totaldiscs");
        if (discTotal != null) {
            metadata.set(Audio.DISC_COUNT, discTotal);
        }

        // A few other bits
        xhtml.element("p", comments.getDate());
        for (String comment : comments.getComments("comment")) {
            xhtml.element("p", comment);
        }
        xhtml.element("p", comments.getGenre());

        // Any embedded pictures, such as cover art, become
        //  embedded documents of the audio file
        extractPictures(xhtml, comments, context);
    }

    /**
     * Sends the embedded pictures, such as cover art, from the comments to
     * the embedded document extractor. The pictures are carried as base64
     * encoded FLAC picture blocks; malformed blocks are skipped silently.
     * The pictures only become embedded documents, no metadata is recorded
     * on the audio document itself.
     */
    private static void extractPictures(XHTMLContentHandler xhtml,
            VorbisStyleComments comments, ParseContext context)
            throws IOException, SAXException {
        EmbeddedDocumentExtractor extractor = null;
        for (String block : comments.getComments(METADATA_BLOCK_PICTURE)) {
            byte[] decoded;
            try {
                decoded = Base64.getMimeDecoder().decode(block);
            } catch (IllegalArgumentException e) {
                //not valid base64, skip
                continue;
            }
            if (extractor == null) {
                extractor = EmbeddedDocumentUtil.getEmbeddedDocumentExtractor(context);
            }
            extractPictureBlock(decoded, xhtml, context, extractor);
        }
    }

    /**
     * Parses one FLAC picture block and sends the picture it holds to the
     * embedded document extractor. Native FLAC PICTURE metadata blocks use
     * the very same structure, so {@link FlacParser} shares this method.
     * Malformed or truncated blocks are skipped silently.
     */
    static void extractPictureBlock(byte[] block, XHTMLContentHandler xhtml,
            ParseContext context, EmbeddedDocumentExtractor extractor)
            throws IOException, SAXException {
        // The picture block holds a 32 bit BE picture type, the mime
        // type, the description, the image geometry and the picture
        // data, with mime type, description and data length prefixed
        int pictureType;
        String mimeType;
        String description;
        byte[] picture;
        try {
            ByteBuffer buffer = ByteBuffer.wrap(block);
            pictureType = buffer.getInt();
            mimeType = getPrefixedString(buffer, StandardCharsets.ISO_8859_1);
            if (mimeType == null || "-->".equals(mimeType)) {
                // Malformed, or a link to a picture rather than an
                // embedded one
                return;
            }
            description = getPrefixedString(buffer, StandardCharsets.UTF_8);
            if (description == null) {
                return;
            }
            // Width, height, color depth and number of colors
            buffer.position(buffer.position() + 16);
            int dataLength = buffer.getInt();
            if (dataLength <= 0 || dataLength > buffer.remaining()) {
                return;
            }
            picture = new byte[dataLength];
            buffer.get(picture);
        } catch (BufferUnderflowException | IllegalArgumentException e) {
            //truncated picture block, skip
            return;
        }

        Metadata pictureMetadata = Metadata.newInstance(context);
        pictureMetadata.set(TikaCoreProperties.EMBEDDED_RESOURCE_TYPE,
                TikaCoreProperties.EmbeddedResourceType.INLINE.toString());
        if (!mimeType.isEmpty()) {
            pictureMetadata.set(HttpHeaders.CONTENT_TYPE, mimeType);
        }
        if (!description.isEmpty()) {
            pictureMetadata.set(TikaCoreProperties.TITLE, description);
        }
        //the FLAC picture block reuses the ID3v2 APIC picture types
        if (pictureType >= 0 && pictureType < ID3Tags.PICTURE_TYPES.length) {
            pictureMetadata.set(TikaCoreProperties.DESCRIPTION,
                    ID3Tags.PICTURE_TYPES[pictureType]);
        }
        if (extractor.shouldParseEmbedded(pictureMetadata, context)) {
            try (TikaInputStream pictureStream = TikaInputStream.get(picture)) {
                extractor.parseEmbedded(pictureStream, xhtml, pictureMetadata, context, true);
            }
        }
    }

    /**
     * Reads a 32 bit length prefixed string from the buffer, or null if the
     * declared length is invalid for the remaining data.
     */
    private static String getPrefixedString(ByteBuffer buffer, Charset charset) {
        int length = buffer.getInt();
        if (length < 0 || length > buffer.remaining()) {
            return null;
        }
        byte[] bytes = new byte[length];
        buffer.get(bytes);
        return new String(bytes, charset);
    }

    protected static void extractDuration(Metadata metadata, XHTMLContentHandler xhtml,
            OggAudioHeaders headers, OggAudioStream audio) throws IOException, SAXException {
        // Have the statistics calculated
        OggAudioStatistics stats = new OggAudioStatistics(headers, audio);
        stats.calculate();

        // Record the duration, if available
        extractDuration(metadata, xhtml, stats.getDurationSeconds());
    }

    protected static void extractDuration(Metadata metadata, XHTMLContentHandler xhtml,
            double duration) throws SAXException {
        // Record the duration, if available
        if (duration > 0) {
            // Save as metadata to the nearest .01 seconds.
            // DecimalFormat is not thread-safe and these parsers are shared across
            // threads, so create a new one per call (see MP4Parser).
            DecimalFormat durationFormat =
                    (DecimalFormat) NumberFormat.getNumberInstance(Locale.ROOT);
            durationFormat.applyPattern("0.0#");
            metadata.add(XMPDM.DURATION, durationFormat.format(duration));

            // Output as Hours / Minutes / Seconds / Parts
            String durationStr = formatDuration(duration);
            xhtml.element("p", durationStr);
        }
    }

    private static String formatDuration(double durationSeconds) {
        long totalSeconds = (long) durationSeconds;
        long hours = totalSeconds / 3600;
        long minutes = (totalSeconds % 3600) / 60;
        long seconds = totalSeconds % 60;
        double fraction = durationSeconds - totalSeconds;

        if (hours > 0) {
            return String.format(Locale.ROOT, "%d:%02d:%02d", hours, minutes, seconds);
        } else {
            return String.format(Locale.ROOT, "%d:%02d", minutes, seconds);
        }
    }

}