Improving "priority_thread.py"

[blueskyleaf]

I am working on improving the priority_thread.py code, so that when selecting "Select priority of first un-downloaded piece to High" 2,5% of the first un-downloaded pieces are set to Highest, 2,5% of the next un-downloaded pieces are set to Higher and 2,5% of the next un-downloaded pieces after that is set to High.

That way any torrent will stream very efficiently. Because just setting High priority on 1 single piece doesn't work well, as is the default in the current code.

But I am unable to change the priority of an amount of pieces relative to the total amount of pieces in the torrent. So currently I have a set amount of pieces that is changed.
Which doesn't work so well when the torrent either have very few or very many of pieces.

Here is what I have done so far.
I have changed this:

__target_priority = 5

(...)

# lf is now the first un-downloaded, desired piece of this torrent
if (tor.handle.piece_priority(lf) < __target_priority):
tor.handle.piece_priority(lf,__target_priority)
__last_first[t] = lf

into this:

__target_priority = 7
__target_priority2 = 5
__target_priority3 = 2

(...)

# lf is now the first un-downloaded, desired piece of this torrent
if (tor.handle.piece_priority(lf) < __target_priority):
    for e in range(0, 25):
        tor.handle.piece_priority(lf+e,__target_priority)
    for e2 in range(25, 51):
        tor.handle.piece_priority(lf+e2,__target_priority2)
    for e3 in range(51, 77):
        tor.handle.piece_priority(lf+e3,__target_priority3)
__last_first[t] = lf

What is the variable for the total amount of pieces, so that I can dynamically calculate the ranges of the for loops?

[tgharib]

For an alternative, check out https://github.com/JohnDoee/deluge-streaming

It supports sequential downloading and it even supports seeking. Once you seek forward, it will sequentially download from that seek point.

[blueskyleaf]

Great, thanks!