search

xtensor-stack / xtensor

C++ tensors with broadcasting and lazy computing
BSD 3-Clause "New" or "Revised" License
3.33k stars 398 forks source link

Bug in rich rendering in Jupyter Notebook #1351

Closed agoose77 closed 5 years ago

agoose77 commented 5 years ago

It seems that there is a bug in the rich rendering of arrays. I haven't tested this for a little while, but I haven't seen any mention of this being fixed.

{
    xt::xarray<double> x = xt::arange(5*5*1);
    x.reshape({5,5, 1});
    xcpp::display(x);
}

This produces this output.

I have a numpy implementation inspired by xtensor, which doesn't have these bugs. I'm placing it here for reference: this issue is mainly just for bookkeeping.

Additionally, we could use unicode characters for the ellipses (e.g for the diagonals).

martinRenou commented 5 years ago

Thanks for reporting! I'll see if I can fix it

martinRenou commented 5 years ago

@agoose77 Can you tell me what version of xtensor you are using? I tried your code example both using master and using 0.19.1 on binder and could not see the issue

agoose77 commented 5 years ago

@martinRenou I just tried it and it seems fixed. Apologies - this was mainly something I wanted to come back and check once I had recreated my conda environment. I think it's fixed, given that this example definitely failed at one point.

martinRenou commented 5 years ago

Ok, thanks for reporting anyway :) feel free to reopen the issue if it shows up again.

agoose77 commented 5 years ago

Sorry @martinRenou, I've found a new corner case:

#include <iostream>

#include "xtensor/xarray.hpp"
#include "xtensor/xio.hpp"
#include "xtensor/xview.hpp"
#include "xcpp/xdisplay.hpp"

unsigned long n = 5;

unsigned long j = 100;

{
    xt::xarray<double> arr = xt::arange(n*j*n);
    arr.reshape({n,j,n});
    xcpp::display(arr);
}
    0.
    1.
    2.
    3.
    4.
    5.
    6.
    7.
    8.
    9.
   10.
   11.
   12.
   13.
   14.
...
...
...
...
...
...
...
  485.
  486.
  487.
  488.
  489.
  490.
  491.
  492.
  493.
  494.
  495.
  496.
  497.
  498.
  499.
  500.
  501.
  502.
  503.
  504.
  505.
  506.
  507.
  508.
  509.
  510.
  511.
  512.
  513.
  514.
...
...
...
...
...
...
...
  985.
  986.
  987.
  988.
  989.
  990.
  991.
  992.
  993.
  994.
  995.
  996.
  997.
  998.
  999.
 1000.
 1001.
 1002.
 1003.
 1004.
 1005.
 1006.
 1007.
 1008.
 1009.
 1010.
 1011.
 1012.
 1013.
 1014.
...
...
...
...
...
...
...
 1485.
 1486.
 1487.
 1488.
 1489.
 1490.
 1491.
 1492.
 1493.
 1494.
 1495.
 1496.
 1497.
 1498.
 1499.
 1500.
 1501.
 1502.
 1503.
 1504.
 1505.
 1506.
 1507.
 1508.
 1509.
 1510.
 1511.
 1512.
 1513.
 1514.
...
...
...
...
...
...
...
 1985.
 1986.
 1987.
 1988.
 1989.
 1990.
 1991.
 1992.
 1993.
 1994.
 1995.
 1996.
 1997.
 1998.
 1999.
 2000.
 2001.
 2002.
 2003.
 2004.
 2005.
 2006.
 2007.
 2008.
 2009.
 2010.
 2011.
 2012.
 2013.
 2014.
...
...
...
...
...
...
...
 2485.
 2486.
 2487.
 2488.
 2489.
 2490.
 2491.
 2492.
 2493.
 2494.
 2495.
 2496.
 2497.
 2498.
 2499.

Here's the same output in the numpy_html implementation:

0 1 2 3 4
5 6 7 8 9
10 11 12 13 14
⋮
⋮
⋮
⋮
⋮
485 486 487 488 489
490 491 492 493 494
495 496 497 498 499
500 501 502 503 504
505 506 507 508 509
510 511 512 513 514
⋮
⋮
⋮
⋮
⋮
985 986 987 988 989
990 991 992 993 994
995 996 997 998 999
1000 1001 1002 1003 1004
1005 1006 1007 1008 1009
1010 1011 1012 1013 1014
⋮
⋮
⋮
⋮
⋮
1485 1486 1487 1488 1489
1490 1491 1492 1493 1494
1495 1496 1497 1498 1499
1500 1501 1502 1503 1504
1505 1506 1507 1508 1509
1510 1511 1512 1513 1514
⋮
⋮
⋮
⋮
⋮
1985 1986 1987 1988 1989
1990 1991 1992 1993 1994
1995 1996 1997 1998 1999
2000 2001 2002 2003 2004
2005 2006 2007 2008 2009
2010 2011 2012 2013 2014
⋮
⋮
⋮
⋮
⋮
2485 2486 2487 2488 2489
2490 2491 2492 2493 2494
2495 2496 2497 2498 2499
martinRenou commented 5 years ago

That was quick :P

martinRenou commented 5 years ago

1353 should fix the issue. I looked at your implementation for numpy, your code is way cleaner than ours :P !

I will open an other PR replacing the ellipsis by unicode characters. Thanks!

agoose77 commented 5 years ago

Thanks for doing the leg work! To be honest, it's slightly tricky to get right, and I found the Python generators made the whole thing far easier :P Awesome, will keep my eyes open!