asciichart icon indicating copy to clipboard operation
asciichart copied to clipboard

Height cannot be exactly configured

Open neighthan opened this issue 6 years ago • 6 comments

Though there is a height parameter for the plotting function, this only roughly controls the height. It would be nicer to be able to exactly set the height. This is even more of an issue because the height isn't always off of the requested height by the same value, so it's not trivial to correct for this. Here are a couple of examples:

Requested height: 9; actual height: 10

$ pplot 1 10 --height 9
   10.00  ┤╭ 
    9.00  ┤│ 
    8.00  ┤│ 
    7.00  ┤│ 
    6.00  ┤│ 
    5.00  ┤│ 
    4.00  ┤│ 
    3.00  ┤│ 
    2.00  ┤│ 
    1.00  ┼╯ 

Requested height: 8; actual height: 10

$ pplot 1 10 --height 8
   10.00  ┼╭ 
    9.00  ┤│ 
    8.00  ┤│ 
    7.00  ┤│ 
    6.00  ┤│ 
    5.00  ┤│ 
    4.00  ┤│ 
    3.00  ┤│ 
    2.00  ┤╯ 
    1.00  ┼  

The issue is with the floor and ceil operators that are used in computing the actual number of rows.

Here's a quick fix for this in the Python version:

    target_height = cfg['height'] if 'height' in cfg else interval
    actual_height = float("inf")

    i = 0
    while actual_height > target_height:
        ratio = (target_height - i) / interval
        min2 = floor(minimum * ratio)
        max2 = ceil(maximum * ratio)
        actual_height = max2 - min2 + 1
        i += 1

    assert actual_height == target_height

I just keep decreasing the height value used by 1 until the actual height matches the target height. This has always given me an actual height that's exactly the target height in the quick tests I've done, but there's probably a better way to do this. I also considered decreasing max2 or increasing min2 by 1 directly, but then we'd also need to set a new value for ratio, at least to keep the current behavior, and I'm not sure what the appropriate way to do that would be.

neighthan avatar Jul 01 '19 17:07 neighthan

Okay, the hacky fix mentioned above does not work in all cases; sometimes the actual_height will end up less than target_height. So, a more principled fix really is needed. (pplot 1 10 5 3 9 7 10 0.1 is an example command where that fix doesn't work).

neighthan avatar Jul 01 '19 17:07 neighthan

@neighthan thx for your involvement in this! It might take some time to get to it, but I will add a fix asap!

kroitor avatar Jul 01 '19 18:07 kroitor

@kroitor any news on a fix?

hartwork avatar Nov 05 '19 22:11 hartwork

@hartwork this is addressed in the linked PR #29, hope to merge it soon.

kroitor avatar Nov 05 '19 22:11 kroitor

It looks like the last action and comment on #29 is from July. Would be great to get that fixed. Thanks!

hartwork avatar Nov 05 '19 22:11 hartwork

@hartwork it's been a bit of a delay due to my other work, but i'll do my best to resolve this in the near future.

kroitor avatar Nov 05 '19 23:11 kroitor