Rename currentPts into lastDecodedAvFramePts (#511)

Author: NicolasHug

src/torchcodec/decoders/_core/VideoDecoder.cpp

@@ -661,9 +661,10 @@ VideoDecoder::getFramesInRange(int64_t start, int64_t stop, int64_t step) {
 VideoDecoder::FrameOutput VideoDecoder::getFramePlayedAt(double seconds) {
   StreamInfo& streamInfo = streamInfos_[activeStreamIndex_];
   double frameStartTime =
-      ptsToSeconds(streamInfo.currentPts, streamInfo.timeBase);
+      ptsToSeconds(streamInfo.lastDecodedAvFramePts, streamInfo.timeBase);
   double frameEndTime = ptsToSeconds(
-      streamInfo.currentPts + streamInfo.currentDuration, streamInfo.timeBase);
+      streamInfo.lastDecodedAvFramePts + streamInfo.lastDecodedAvFrameDuration,
+      streamInfo.timeBase);
   if (seconds >= frameStartTime && seconds < frameEndTime) {
     // We are in the same frame as the one we just returned. However, since we
     // don't cache it locally, we have to rewind back.
@@ -824,8 +825,8 @@ void VideoDecoder::setCursorPtsInSeconds(double seconds) {
 Videos have I frames and non-I frames (P and B frames). Non-I frames need data
 from the previous I frame to be decoded.
 
-Imagine the cursor is at a random frame with PTS=x and we wish to seek to a
-user-specified PTS=y.
+Imagine the cursor is at a random frame with PTS=lastDecodedAvFramePts (x for
+brevity) and we wish to seek to a user-specified PTS=y.
 
 If y < x, we don't have a choice but to seek backwards to the highest I frame
 before y.
@@ -845,23 +846,24 @@ I    P     P    P    I    P    P    P    I    P    P    I    P    P    I    P
 
 (2) is more efficient than (1) if there is an I frame between x and y.
 */
-bool VideoDecoder::canWeAvoidSeeking(int64_t currentPts, int64_t targetPts)
-    const {
-  if (targetPts < currentPts) {
+bool VideoDecoder::canWeAvoidSeeking(int64_t targetPts) const {
+  int64_t lastDecodedAvFramePts =
+      streamInfos_.at(activeStreamIndex_).lastDecodedAvFramePts;
+  if (targetPts < lastDecodedAvFramePts) {
     // We can never skip a seek if we are seeking backwards.
     return false;
   }
-  if (currentPts == targetPts) {
+  if (lastDecodedAvFramePts == targetPts) {
     // We are seeking to the exact same frame as we are currently at. Without
     // caching we have to rewind back and decode the frame again.
     // TODO: https://github.com/pytorch-labs/torchcodec/issues/84 we could
     // implement caching.
     return false;
   }
   // We are seeking forwards.
-  // We can only skip a seek if both currentPts and targetPts share the same
-  // keyframe.
-  int currentKeyFrameIndex = getKeyFrameIndexForPts(currentPts);
+  // We can only skip a seek if both lastDecodedAvFramePts and targetPts share
+  // the same keyframe.
+  int currentKeyFrameIndex = getKeyFrameIndexForPts(lastDecodedAvFramePts);
   int targetKeyFrameIndex = getKeyFrameIndexForPts(targetPts);
   return currentKeyFrameIndex >= 0 && targetKeyFrameIndex >= 0 &&
       currentKeyFrameIndex == targetKeyFrameIndex;
@@ -879,7 +881,7 @@ void VideoDecoder::maybeSeekToBeforeDesiredPts() {
   decodeStats_.numSeeksAttempted++;
 
   int64_t desiredPtsForStream = *desiredPtsSeconds_ * streamInfo.timeBase.den;
-  if (canWeAvoidSeeking(streamInfo.currentPts, desiredPtsForStream)) {
+  if (canWeAvoidSeeking(desiredPtsForStream)) {
     decodeStats_.numSeeksSkipped++;
     return;
   }
@@ -1032,8 +1034,8 @@ VideoDecoder::AVFrameStream VideoDecoder::decodeAVFrame(
   // haven't received as frames. Eventually we will either hit AVERROR_EOF from
   // av_receive_frame() or the user will have seeked to a different location in
   // the file and that will flush the decoder.
-  streamInfo.currentPts = avFrame->pts;
-  streamInfo.currentDuration = getDuration(avFrame);
+  streamInfo.lastDecodedAvFramePts = avFrame->pts;
+  streamInfo.lastDecodedAvFrameDuration = getDuration(avFrame);
 
   return AVFrameStream(std::move(avFrame), activeStreamIndex_);
 }

src/torchcodec/decoders/_core/VideoDecoder.h

@@ -332,8 +332,8 @@ class VideoDecoder {
 
     // The current position of the cursor in the stream, and associated frame
     // duration.
-    int64_t currentPts = 0;
-    int64_t currentDuration = 0;
+    int64_t lastDecodedAvFramePts = 0;
+    int64_t lastDecodedAvFrameDuration = 0;
     // The desired position of the cursor in the stream. We send frames >=
     // this pts to the user when they request a frame.
     // We update this field if the user requested a seek. This typically
@@ -361,7 +361,7 @@ class VideoDecoder {
   // DECODING APIS AND RELATED UTILS
   // --------------------------------------------------------------------------
 
-  bool canWeAvoidSeeking(int64_t currentPts, int64_t targetPts) const;
+  bool canWeAvoidSeeking(int64_t targetPts) const;
 
   void maybeSeekToBeforeDesiredPts();