Enhance duration parsing in YouTube extractors to handle invalid inputs and improve Shorts detection

Author: Ecomont

extractor/src/main/java/org/schabi/newpipe/extractor/services/youtube/YoutubeParsingHelper.java

@@ -235,6 +235,10 @@ public static boolean isY2ubeURL(@Nonnull final URL url) {
      */
     public static int parseDurationString(@Nonnull final String input)
             throws ParsingException, NumberFormatException {
+        if (!input.matches(".*\\d.*") && !input.equalsIgnoreCase("SHORTS")) {
+            throw new ParsingException("Error duration string contains no digits: " + input);
+        }
+
         // If time separator : is not detected, try . instead
         final String[] splitInput = input.contains(":")
                 ? input.split(":")

extractor/src/main/java/org/schabi/newpipe/extractor/services/youtube/extractors/YoutubeStreamInfoItemExtractor.java

@@ -49,6 +49,7 @@
 import java.time.format.DateTimeFormatter;
 import java.util.List;
 import java.util.regex.Pattern;
+import java.util.stream.Collectors;
 
 public class YoutubeStreamInfoItemExtractor implements StreamInfoItemExtractor {
 
@@ -155,19 +156,24 @@ public long getDuration() throws ParsingException {
             duration = videoInfo.getString("lengthSeconds");
 
             if (isNullOrEmpty(duration)) {
-                final JsonObject timeOverlay = videoInfo.getArray("thumbnailOverlays")
+                final List<String> timeOverlays = videoInfo.getArray("thumbnailOverlays")
                         .stream()
                         .filter(JsonObject.class::isInstance)
                         .map(JsonObject.class::cast)
                         .filter(thumbnailOverlay ->
                                 thumbnailOverlay.has("thumbnailOverlayTimeStatusRenderer"))
-                        .findFirst()
-                        .orElse(null);
-
-                if (timeOverlay != null) {
-                    duration = getTextFromObject(
-                            timeOverlay.getObject("thumbnailOverlayTimeStatusRenderer")
-                                    .getObject("text"));
+                        .map(thumbnailOverlay -> getTextFromObject(
+                                thumbnailOverlay.getObject("thumbnailOverlayTimeStatusRenderer")
+                                        .getObject("text")))
+                        .filter(text -> !isNullOrEmpty(text))
+                        .collect(Collectors.toList());
+
+                for (final String timeOverlayText : timeOverlays) {
+                    try {
+                        return YoutubeParsingHelper.parseDurationString(timeOverlayText);
+                    } catch (final ParsingException ex) {
+                        // try next
+                    }
                 }
             }
 
@@ -452,24 +458,21 @@ public boolean isShortFormContent() throws ParsingException {
             }
 
             if (!isShort) {
-                final JsonObject thumbnailTimeOverlay = videoInfo.getArray("thumbnailOverlays")
-                        .stream()
-                        .filter(JsonObject.class::isInstance)
-                        .map(JsonObject.class::cast)
-                        .filter(thumbnailOverlay -> thumbnailOverlay.has(
-                                "thumbnailOverlayTimeStatusRenderer"))
-                        .map(thumbnailOverlay -> thumbnailOverlay.getObject(
-                                "thumbnailOverlayTimeStatusRenderer"))
-                        .findFirst()
-                        .orElse(null);
-
-                if (!isNullOrEmpty(thumbnailTimeOverlay)) {
-                    isShort = thumbnailTimeOverlay.getString("style", "")
-                            .equalsIgnoreCase("SHORTS")
-                            || thumbnailTimeOverlay.getObject("icon")
-                            .getString("iconType", "")
-                            .toLowerCase()
-                            .contains("shorts");
+                if (videoInfo.has("thumbnailOverlays")) {
+                    isShort = videoInfo.getArray("thumbnailOverlays")
+                            .stream()
+                            .filter(JsonObject.class::isInstance)
+                            .map(JsonObject.class::cast)
+                            .filter(thumbnailOverlay -> thumbnailOverlay.has(
+                                    "thumbnailOverlayTimeStatusRenderer"))
+                            .map(thumbnailOverlay -> thumbnailOverlay.getObject(
+                                    "thumbnailOverlayTimeStatusRenderer"))
+                            .anyMatch(timeOverlay -> timeOverlay.getString("style", "")
+                                    .equalsIgnoreCase("SHORTS")
+                                    || timeOverlay.getObject("icon")
+                                    .getString("iconType", "")
+                                    .toLowerCase()
+                                    .contains("shorts"));
                 }
             }
 

extractor/src/main/java/org/schabi/newpipe/extractor/services/youtube/extractors/YoutubeStreamInfoItemLockupExtractor.java

@@ -31,9 +31,9 @@
 /**
  * Note:
  * This extractor is currently (2025-07) only used to extract related video streams.<br>
- * The following features are currently not implemented because they have never been observed:
+ * The following features are currently not implemented:
  * <ul>
- *     <li>Shorts</li>
+ *     <li>Shorts: appear in related videos without a duration badge; getDuration() returns -1</li>
  *     <li>Paid content (Premium, members first or only)</li>
  * </ul>
  */
@@ -164,18 +164,25 @@ public long getDuration() throws ParsingException {
             .collect(Collectors.toList());
 
         if (potentialDurations.isEmpty()) {
-            throw new ParsingException("Could not get duration: No parsable durations detected");
+            return -1;
         }
 
         ParsingException parsingException = null;
         for (final String potentialDuration : potentialDurations) {
+            if (potentialDuration == null || !potentialDuration.matches(".*\\d.*")) {
+                continue;
+            }
             try {
                 return YoutubeParsingHelper.parseDurationString(potentialDuration);
             } catch (final ParsingException ex) {
                 parsingException = ex;
             }
         }
 
+        if (parsingException == null) {
+            return -1; // e.g. only "SHORTS" or "CC" badge was present, no duration available
+        }
+
         throw new ParsingException("Could not get duration", parsingException);
     }