From 76f79f600a76fe2bd3d5387e4c468286005473fe Mon Sep 17 00:00:00 2001 From: Max Bain Date: Sat, 28 Jan 2023 19:04:19 +0000 Subject: [PATCH] fix short seg timestamps bug --- whisperx/alignment.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/whisperx/alignment.py b/whisperx/alignment.py index 6d2ad6f..e4950f9 100644 --- a/whisperx/alignment.py +++ b/whisperx/alignment.py @@ -314,7 +314,7 @@ def align( word_segments_arr[["segment-idx", "subsegment-idx", "word-idx"]] = per_word_grp["local-char-idx"].min().reset_index()[["segment-idx", "subsegment-idx", "word-idx"]].astype(int) segments_arr = {} segments_arr["start"] = per_subseg_grp["start"].min().reset_index()["start"] - segments_arr["end"] = per_subseg_grp["end"].min().reset_index()["end"] + segments_arr["end"] = per_subseg_grp["end"].max().reset_index()["end"] segments_arr = pd.DataFrame(segments_arr) segments_arr[["segment-idx", "subsegment-idx-start"]] = per_subseg_grp["start"].min().reset_index()[["segment-idx", "subsegment-idx"]] segments_arr["subsegment-idx-end"] = segments_arr["subsegment-idx-start"] + 1