ESC-XX datasets changed fundamentally?

[ThetomekK]

Hi there,

did the ESC datasets changed over time?
make_file.py 's logic excpects '.ogg' audio files scattered across different folders!?
This does'nt correspond to:

{FOLD}-{CLIP_ID}-{TAKE}-{TARGET}.wav

{FOLD} - index of the cross-validation fold,
{CLIP_ID} - ID of the original Freesound clip,
{TAKE} - letter disambiguating between different fragments from the same Freesound clip,
{TARGET} - class in numeric format [0, 49].

Am i right? Are any adoptions planned to work with new dataset structure or did anybody it before?