Support SoundCloud HLS by using a workaround

This commit tries to support SoundCloud HLS streams by parsing M3U manifests, get the last segment URL (in order to get track length) and request a segment URL equals to track's duration so it's a single URL.
Warning: This code doesn't work for now and throws a StreamExtractException

extractor/src/main/java/org/schabi/newpipe/extractor/services/soundcloud/extractors/SoundcloudStreamExtractor.java

@@ -11,9 +11,9 @@
 import org.schabi.newpipe.extractor.StreamingService;
 import org.schabi.newpipe.extractor.downloader.Downloader;
 import org.schabi.newpipe.extractor.exceptions.ContentNotAvailableException;
-import org.schabi.newpipe.extractor.exceptions.ContentNotSupportedException;
 import org.schabi.newpipe.extractor.exceptions.ExtractionException;
 import org.schabi.newpipe.extractor.exceptions.ParsingException;
+import org.schabi.newpipe.extractor.exceptions.ReCaptchaException;
 import org.schabi.newpipe.extractor.linkhandler.LinkHandler;
 import org.schabi.newpipe.extractor.localization.DateWrapper;
 import org.schabi.newpipe.extractor.services.soundcloud.SoundcloudParsingHelper;
@@ -33,11 +33,14 @@
 import java.util.Collections;
 import java.util.List;
 import java.util.Locale;
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
 
 import javax.annotation.Nonnull;
 import javax.annotation.Nullable;
 
 import static org.schabi.newpipe.extractor.utils.JsonUtils.EMPTY_STRING;
+import static org.schabi.newpipe.extractor.utils.Utils.HTTPS;
 import static org.schabi.newpipe.extractor.utils.Utils.isNullOrEmpty;
 
 public class SoundcloudStreamExtractor extends StreamExtractor {
@@ -179,7 +182,7 @@ public String getHlsUrl() {
 
     @Override
     public List<AudioStream> getAudioStreams() throws IOException, ExtractionException {
-        List<AudioStream> audioStreams = new ArrayList<>();
+        final List<AudioStream> audioStreams = new ArrayList<>();
         final Downloader dl = NewPipe.getDownloader();
 
         // Streams can be streamable and downloadable - or explicitly not.
@@ -190,54 +193,102 @@ public List<AudioStream> getAudioStreams() throws IOException, ExtractionExcepti
         try {
             final JsonArray transcodings = track.getObject("media").getArray("transcodings");
 
-            // get information about what stream formats are available
-            for (Object transcoding : transcodings) {
-
+            // Get information about what stream formats are available
+            for (final Object transcoding : transcodings) {
                 final JsonObject t = (JsonObject) transcoding;
                 String url = t.getString("url");
+                String mediaUrl = null;
+                final MediaFormat mediaFormat;
+                final int bitrate;
 
                 if (!isNullOrEmpty(url)) {
+                    if (t.getString("preset").contains("mp3")) {
+                        mediaFormat = MediaFormat.MP3;
+                        bitrate = 128;
+                    } else if (t.getString("preset").contains("opus")) {
+                        mediaFormat = MediaFormat.OPUS;
+                        bitrate = 64;
+                    } else {
+                        continue;
+                    }
 
-                    // We can only play the mp3 format, but not handle m3u playlists / streams.
-                    // what about Opus?
-                    if (t.getString("preset").contains("mp3")
-                            && t.getObject("format").getString("protocol").equals("progressive")) {
+                    // TODO: move this to a separate method to generate valid urls when needed (e.g. resuming a paused stream)
+
+                    if (t.getObject("format").getString("protocol").equals("progressive")) {
                         // This url points to the endpoint which generates a unique and short living url to the stream.
-                        // TODO: move this to a separate method to generate valid urls when needed (e.g. resuming a paused stream)
                         url += "?client_id=" + SoundcloudParsingHelper.clientId();
                         final String res = dl.get(url).responseBody();
 
                         try {
                             JsonObject mp3UrlObject = JsonParser.object().from(res);
                             // Links in this file are also only valid for a short period.
-                            audioStreams.add(new AudioStream(mp3UrlObject.getString("url"),
-                                    MediaFormat.MP3, 128));
-                        } catch (JsonParserException e) {
+                            mediaUrl = mp3UrlObject.getString("url");
+                        } catch (final JsonParserException e) {
                             throw new ParsingException("Could not parse streamable url", e);
                         }
+                    } else if (t.getObject("format").getString("protocol").equals("hls")) {
+                        // This url points to the endpoint which generates a unique and short living url to the stream.
+                        url += "?client_id=" + SoundcloudParsingHelper.clientId();
+                        final String res = dl.get(url).responseBody();
+
+                        try {
+                            final JsonObject mp3HlsUrlObject = JsonParser.object().from(res);
+                            // Links in this file are also only valid for a short period.
+                            try {
+                                mediaUrl = getUrlFromHlsManifest(mp3HlsUrlObject.getString("url"));
+                            } catch (final ParsingException e) {
+                                continue;
+                            }
+                        } catch (final JsonParserException e) {
+                            throw new ParsingException("Could not parse streamable url", e);
+                        }
+                    } else {
+                        continue;
                     }
+
+                    audioStreams.add(new AudioStream(mediaUrl, mediaFormat, bitrate));
                 }
             }
 
-        } catch (NullPointerException e) {
+        } catch (final NullPointerException e) {
             throw new ExtractionException("Could not get SoundCloud's track audio url", e);
         }
 
-        if (audioStreams.isEmpty()) {
-            throw new ContentNotSupportedException("HLS audio streams are not yet supported");
-        }
-
         return audioStreams;
     }
 
     private static String urlEncode(String value) {
         try {
             return URLEncoder.encode(value, "UTF-8");
-        } catch (UnsupportedEncodingException e) {
+        } catch (final UnsupportedEncodingException e) {
             throw new IllegalStateException(e);
         }
     }
 
+    private String getUrlFromHlsManifest(final String hlsManifestUrl) throws ParsingException {
+        final Downloader dl = NewPipe.getDownloader();
+        final String hlsManifestResponse;
+
+        try {
+            hlsManifestResponse = dl.get(hlsManifestUrl).responseBody();
+        } catch (final IOException | ReCaptchaException e) {
+            throw new ParsingException("Could not get SoundCloud HLS manifest");
+        }
+
+        final List<String> hlsRangesList = new ArrayList<String>();
+        final Matcher regex = Pattern.compile("^(https?)://[-a-zA-Z0-9+&@#/%?=~_|!:,.;]*[-a-zA-Z0-9+&@#/%=~_|]")
+                .matcher(hlsManifestResponse);
+
+        while (regex.find()) {
+            hlsRangesList.add(hlsManifestResponse.substring(regex.start(0), regex.end(0)));
+        }
+
+        final String hlsLastRangeUrl = hlsRangesList.get(hlsRangesList.size() - 1);
+        final String[] hlsLastRangeUrlArray = hlsLastRangeUrl.split("/");
+
+        return HTTPS + hlsLastRangeUrlArray[0] + "/media/0/" + hlsLastRangeUrlArray[3] + "/" + hlsLastRangeUrlArray[4];
+    }
+
     @Override
     public List<VideoStream> getVideoStreams() {
         return Collections.emptyList();
@@ -334,4 +385,4 @@ public List<StreamSegment> getStreamSegments() {
     public List<MetaInfo> getMetaInfo() {
         return Collections.emptyList();
     }
-}
+}