mx_array_to_chunks: Increment the start index of chunk data by previous offset

[jaglinux]

No description provided.

[jaglinux]

@vegaluisjose Thanks for the code, really helps in understanding vdb on mlx framework.
For query_vdb.py, I see that the response is always same for any queries because the values of chunks are all same. Since the start index is not incrementing, whatever might be the score, the resulting index returns the same chunk.

[stronk7]

Well spotted, @jaglinux . I can confirm that, without that change, the context passed is, 99% of times 100% useless. I just applied for it here and now and my results are way better, heh.

Ciao :-)