matplotlib-users Mailing List for matplotlib

Well, I'm not aware of any special way, but again, I'm no guru.

Given these lists : left, height you could simply do something like :

>>> bar(left, height, width)
>>> for l,h in zip(left,height):
>>>     text(l+width/2., h + dh, str(h), horizontalalignment =3D 'center')

where dh is the distance wanted between the top of the bar and the text.


David

2006/3/13, Francis Penney <fp...@gm...>:
> Just a simple bar chart question.  How do I display the values above each
> bar?
>

>>>>> "Derek" == Derek Basch <db...@ya...> writes:

    Derek> I am trying to plot a linear regression on a logarithmic
    Derek> scale and am not quite sure how to do it.  I used examples
    Derek> that I had found online but the linear regression line
    Derek> doesn't plot the same on a logarithmic scale. Can anyone
    Derek> help me? Here is what I have so far:

This is not an issue of log versus non-log.  You should sort your
xdata before plotting the line; you just don't notice it when you plot
nonsorted data that all lie on the same line.

Just do this after your data definition

  prediction_experiment.sort()

Also, rather than

  x = [f[0] for f in prediction_experiment]
  y = [z[1] for z in prediction_experiment]
  x = array(x)
  y = array(y)

you might prefer

  x, y = map(array, zip(*prediction_experiment))

JDH

I am trying to plot a linear regression on a logarithmic scale and am not quite sure how to do it.
I used examples that I had found online but the linear regression line doesn't plot the same on a
logarithmic scale. Can anyone help me? Here is what I have so far:

from pylab import *

prediction_experiment = [(313.11000000000001, 25.797999999999998), (4499.1999999999998, 25000.0),
(168830.0, 440000.0), (143090.0, 78571.399999999994), (34811.0, 78571.399999999994), (161240.0,
70967.699999999997), (1000000.0, 78571.399999999994), (0.93820000000000003, 1.4666699999999999),
(3.0781000000000001, 2.8571399999999998), (64.768000000000001, 78571.399999999994),
(42.656999999999996, 10576.9), (11.473000000000001, 193.05000000000001), (173.90000000000001,
14.666700000000001), (2815.1999999999998, 16.8933), (78387.0, 78571.399999999994), (31665.0,
78571.399999999994), (1000000.0, 4000.0), (59298.0, 2638.8499999999999), (13179.0, 25000.0),
(1496.9000000000001, 188.08199999999999), (815.99000000000001, 24.601900000000001),
(386.07999999999998, 68750.0), (273750.0, 34896.800000000003), (28435.0, 25000.0),
(9699.2000000000007, 25000.0), (13515.0, 40000.0), (19219.0, 25000.0), (379170.0, 25000.0),
(379560.0, 78571.399999999994), (427.58999999999997, 25000.0), (638.86000000000001, 25000.0),
(89046.0, 25000.0), (151440.0, 25000.0), (1602.5, 25000.0), (4242.8000000000002, 25000.0),
(118700.0, 25000.0), (11454.0, 5500.0), (4094.5, 511.62799999999999), (17730.0, 25000.0),
(754.64999999999998, 1692.3099999999999), (2183.9000000000001, 25000.0), (192330.0, 25000.0),
(241170.0, 25000.0), (1.4981, 7.2522099999999998), (9.3475999999999999, 147.07400000000001),
(14.512, 5.4083300000000003), (194.66999999999999, 152.30199999999999), (10.856999999999999,
7.9026300000000003), (5.1627999999999998, 94.151399999999995), (1000000.0, 53658.5),
(912.51999999999998, 448.98000000000002), (18511.0, 25000.0), (1870.7, 6027.3999999999996),
(2665.5999999999999, 10138.200000000001), (14708.0, 55000.0), (38.012, 222.988),
(295.61000000000001, 44898.0), (5.4006999999999996, 6.6227600000000004), (1000000.0,
78571.399999999994), (1000000.0, 2050.3299999999999), (1549.5, 25000.0), (19.260999999999999,
13.0), (16.82, 3.5536300000000001), (67.953999999999994, 83.956599999999995), (1651.2,
3384.6199999999999), (104.84999999999999, 38.357599999999998), (1435.2, 10091.700000000001),
(78.744, 27.834), (351.27999999999997, 28.004100000000001), (27.716000000000001,
68.239699999999999), (380.61000000000001, 25000.0), (423.26999999999998, 5500.0), (342660.0,
155000.0), (60756.0, 25000.0), (155060.0, 25000.0), (1000000.0, 78571.399999999994),
(228.90000000000001, 40.740699999999997), (1355.5, 25000.0), (4744.1999999999998,
1235.0799999999999), (1000000.0, 43239.5), (5800.6000000000004, 78571.399999999994),
(73.650000000000006, 15.1372), (66.897999999999996, 666.66700000000003), (16031.0, 25000.0),
(82.275000000000006, 78571.399999999994), (1000000.0, 25000.0), (3287.6999999999998, 25000.0),
(418600.0, 78571.399999999994), (126.83, 1739.1300000000001), (139.5, 37288.099999999999),
(13376.0, 25000.0), (268.54000000000002, 25000.0), (175630.0, 25000.0), (29286.0,
78571.399999999994), (405.25, 1.0), (2048.8000000000002, 25000.0), (1569.0999999999999,
63.398800000000001), (1378.5, 70967.699999999997), (16883.0, 25000.0), (3467.3000000000002,
25000.0), (138140.0, 78571.399999999994), (24807.0, 523.80999999999995), (434.27999999999997,
7.0816499999999998), (31013.0, 25000.0), (48096.0, 78571.399999999994), (1510.3, 1093.98),
(1574.4000000000001, 3283.5799999999999), (7285.5, 880.0), (9767.5, 12500.0), (1000000.0,
78571.399999999994), (6.4268000000000001, 314.0), (22.384, 36.363399999999999),
(46.899999999999999, 5.2683299999999997), (347.31, 220.0), (66.600999999999999,
165.83099999999999), (1361.5999999999999, 78571.399999999994), (436640.0, 4782.6099999999997),
(171290.0, 57032.300000000003), (52.948999999999998, 29733.299999999999), (192.00999999999999,
329.0), (1171.4000000000001, 3666.6700000000001), (474.16000000000003, 782.64300000000003),
(42.241, 1.99908), (184960.0, 25000.0), (1867.2, 15714.299999999999), (5690.5, 2480.0),
(3382.5999999999999, 1157.8900000000001), (1000000.0, 78571.399999999994), (1301.2,
372.62900000000002), (1608.8, 25000.0), (98.102999999999994, 57.258899999999997), (42922.0,
78571.399999999994), (3068.0999999999999, 1.0621700000000001), (210.25999999999999, 16176.5),
(611.63999999999999, 31884.099999999999), (211700.0, 78571.399999999994), (11631.0,
240.30600000000001), (1798.0999999999999, 654.76199999999994), (67.960999999999999,
17728.099999999999), (23457.0, 25000.0), (2092.0999999999999, 25000.0), (3984.8000000000002,
25000.0), (1000000.0, 78571.399999999994), (4816.3000000000002, 25000.0), (20144.0, 25000.0),
(1000000.0, 70967.699999999997), (34961.0, 6287.0), (276.86000000000001, 767.16999999999996),
(1000000.0, 70967.699999999997), (665.45000000000005, 1398.5999999999999), (650750.0,
78571.399999999994), (15336.0, 5500.0), (1000000.0, 78571.399999999994), (167.88,
733.33299999999997), (19191.0, 4.3558700000000004), (15329.0, 956.52200000000005), (922060.0,
25000.0), (1000000.0, 78571.399999999994), (1000000.0, 2885.8499999999999), (235890.0, 25000.0),
(1000000.0, 78571.399999999994), (29382.0, 78571.399999999994), (128.03999999999999, 2219.98),
(100.16, 708.53499999999997), (585.59000000000003, 17322.799999999999), (1000000.0,
78571.399999999994), (268.68000000000001, 20.957599999999999), (73322.0, 25000.0), (112410.0,
25000.0), (7489.1000000000004, 25000.0), (2382.4000000000001, 31.6875), (260540.0,
78571.399999999994), (518.32000000000005, 25000.0), (1000000.0, 25000.0), (1000000.0,
10051.299999999999), (1000000.0, 25000.0), (89655.0, 25000.0), (818.88, 25000.0), (759950.0,
25000.0), (46101.0, 18837.200000000001), (79.030000000000001, 5.5999999999999996),
(339.06999999999999, 5500.0), (3427.6999999999998, 25000.0), (126660.0, 4174.5699999999997),
(26801.0, 25000.0), (980.72000000000003, 293333.0), (136190.0, 25000.0), (3505.9000000000001,
742.24000000000001), (33.753, 222.51400000000001), (436760.0, 25000.0), (135.43000000000001,
25000.0), (305.18000000000001, 25000.0), (1349.0999999999999, 25000.0), (12096.0,
758.62099999999998), (1000000.0, 44898.0), (240.83000000000001, 20.0), (134.13,
50.645099999999999), (35997.0, 25000.0), (1000000.0, 78571.399999999994), (7.2370000000000001,
2.2916699999999999), (600.94000000000005, 19.6416), (13117.0, 4150.9399999999996),
(1188.0999999999999, 78571.399999999994), (799680.0, 78571.399999999994), (14954.0,
78571.399999999994), (479.23000000000002, 25000.0), (7310.8000000000002, 1050.6199999999999),
(1000000.0, 234.24199999999999), (12.324, 2.5555300000000001), (414240.0, 25000.0), (16625.0,
25000.0), (10895.0, 60.009999999999998), (241.34999999999999, 796.81299999999999), (1000000.0,
25000.0), (11247.0, 1692.3099999999999), (13030.0, 25000.0), (59775.0, 78571.399999999994),
(2506.5999999999999, 160.0), (9283.2999999999993, 25000.0), (1000000.0, 78571.399999999994),
(1000000.0, 78571.399999999994), (1000000.0, 9909.9099999999999), (932.30999999999995,
14.982100000000001), (2996.3000000000002, 1375.0), (1000000.0, 78571.399999999994),
(6337.1000000000004, 27500.0), (267.69, 787.68299999999999), (7534.5, 25000.0),
(4675.8999999999996, 27.0943), (1057.2, 3.7930999999999999), (5138.0, 25000.0), (36588.0,
58688.699999999997), (1000000.0, 78571.399999999994), (148.02000000000001, 2.6151200000000001),
(473.39999999999998, 36.369), (1000000.0, 78571.399999999994), (4123.1000000000004,
1973.0899999999999), (554500.0, 78571.399999999994), (1000000.0, 25000.0), (13519.0, 26445.0),
(533.91999999999996, 22000.0), (189.97999999999999, 3.8650000000000002), (73.228999999999999,
27848.099999999999), (697.75999999999999, 1000.0), (145420.0, 78571.399999999994),
(8319.2000000000007, 22000.0), (6829.3000000000002, 8.2332999999999998), (3313.8000000000002,
17600.0), (4224.8000000000002, 1294.1199999999999), (1705.8, 25000.0), (49627.0,
70967.699999999997), (125460.0, 78571.399999999994), (11.382999999999999, 7.9088900000000004),
(153.88, 71.488900000000001), (49.518999999999998, 25000.0), (8976.7999999999993, 25000.0),
(377.97000000000003, 607.06399999999996), (202910.0, 19674.200000000001), (395.48000000000002,
2109.3000000000002), (23.899999999999999, 29.602499999999999), (1000000.0, 25000.0), (159980.0,
550.0), (11661.0, 25000.0), (6065.0, 25000.0), (1000000.0, 11399.0), (540.85000000000002,
1571.4300000000001), (886.70000000000005, 767.35299999999995), (315.69, 62857.0), (3640.0, 427.0),
(2753.3000000000002, 25000.0), (1000000.0, 78571.399999999994), (1000000.0, 25000.0),
(83.905000000000001, 18032.799999999999), (1220.9000000000001, 29.352499999999999),
(1357.4000000000001, 63.269300000000001), (4245.1999999999998, 19550.900000000001),
(969.55999999999995, 25000.0), (1482.9000000000001, 758.62099999999998), (5051.3999999999996,
17632.599999999999), (580.74000000000001, 75.862099999999998), (79868.0, 2696.0799999999999),
(42.079999999999998, 1.0), (9.0183999999999997, 5.9118700000000004), (45.177, 3379.4200000000001),
(881.97000000000003, 84.410899999999998), (1245.2, 5138.4799999999996), (2044.4000000000001,
8429.1200000000008), (3889.6999999999998, 78571.399999999994), (292.23000000000002, 25000.0),
(51.344999999999999, 3.1993499999999999), (21.702999999999999, 25.287400000000002),
(4552.6999999999998, 25000.0), (12536.0, 197.983), (65.808999999999997, 105.723),
(2555.0999999999999, 25000.0), (539.64999999999998, 5500.0), (112.56, 6267.8299999999999),
(2950.6999999999998, 25000.0), (3805.8000000000002, 7333.3299999999999), (883.62,
209.18600000000001), (455.74000000000001, 12.0718), (294.00999999999999, 889.0),
(11.066000000000001, 31.7607)]

x = [f[0] for f in prediction_experiment]
y = [z[1] for z in prediction_experiment]

x = array(x)
y = array(y)

def log10Product(x, pos):
    """The two args are the value and tick position.
    Label ticks with the product of the exponentiation"""
    return '%1i' % (x)

ax = subplot(111)

ax.set_xscale('log')
ax.set_yscale('log')

formatter = FuncFormatter(log10Product)
ax.xaxis.set_major_formatter(formatter)
ax.yaxis.set_major_formatter(formatter)
    
# the bestfit line from polyfit
m, b = polyfit(x, y, 1)  # a line is 1st order polynomial...

plot(x, y, 'bs', x, m*x+b, 'r-', linewidth=1, markersize=2)

# Must add 1 to allow the last decades label to be shown
ax.set_xlim(1e-1, max(x)+1)
ax.set_ylim(1e-1, max(y)+1)
grid(True)        
xlabel(r"Prediction", fontsize = 12)
ylabel(r"Experimental IC50 [nM]", fontsize = 12)

show()

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

This problem was resolved off list and I'm including it here for
others benefit.  If you upgrade to a newer version of numpy and are
compiling mpl from src, you will need to=20

  1) install the new numpy

  2) flush all traces of your previous matplotlib build, eg by
     removing the build dir

  3) re-install matplotlib

Le Mardi 14 Mars 2006 13:58, vous avez =E9crit=A0:
> >>>>> "manouchk" =3D=3D manouchk  <man...@gm...> writes:
>
>     manouchk> Well, I use the src.rpm to rebuild so the build is done
>     manouchk> in a cleaned directory (Ijust modified the name of the
>     manouchk> package to respect mandriva policy python-NumPy...)
>
>     manouchk> By the way I use mandriva 2005 I did build in that order
>     manouchk> numpy-0.9.6 then scipy-0.4.6 (maybe not relevant for
>     manouchk> matplotlib) and then matplotlib-0.87.1
>
>     manouchk> I see that coocker is not yet chipping version 0.87.1
>
> I'm still willing to bet the problem is caused by an unclean build.
> Please make sure you rm -rf all build dirs and the
> site-packages/matplotlib install dir for good measure.  Travis
> Oliphant and I both separately tested a matplotlib 87.1 build against
> numpy 0.9.6.  At first I got a segfault when I did not have a clean
> build.  After removing the build dirs and reinstalling, everything
> went fine and my tests passed.

you are right (at least I think)
I done the build again step by step... and the simplest example work!
I think that I stupidly(?) forgot to install Numpy-0.4.6 before compiling=
=20
matplotlib 0.87.1
sorry for the disturbance.

> You'll know you get a clean build if it takes a long time <wink>

it was not too long (few minutes?)

> JDH

Le Mardi 14 Mars 2006 11:21, vous avez =E9crit :
> >>>>> "manouchk" =3D=3D manouchk  <man...@gm...> writes:
>
>     manouchk> Simply reverting Numpy from version 0.9.6 to version
>     manouchk> 0.9.4 solved the problem!  Does someone understand
>     manouchk> something to that?
>
> If you are building from source, you need to do a clean rebuild after
> upgrading numpy
>
>  > cd matplotlib_src
>  > sudo rm -rf build
>  > sudo python setup.py install
>
> If you still encounter problems after that, be sure to let us know.
>
> JDH

Well, I use the src.rpm to rebuild so the build is done in a cleaned direct=
ory=20
(Ijust modified the name of the package to respect mandriva policy=20
python-NumPy...)

By the way I use mandriva 2005
I did build in that order=20
numpy-0.9.6
then scipy-0.4.6 (maybe not relevant for matplotlib)
and then matplotlib-0.87.1

>>>>> "John" == John Hunter <jdh...@ac...> writes:

>>>>> "Yogesh" == Yogesh Wadadekar <wad...@st...> writes:
    Yogesh> Hi,

    Yogesh> I encounter problems with the solution below. My test case
    Yogesh> from yesterday was really not a 'test case' because both
    Yogesh> images had near identical values. Thus, the auto vmin,vmax
    Yogesh> settings were the same in both images.

    John> I think the difference we are seeing may due to the fact
    John> that the norm attributes are not set until the figure is
    John> drawn.  So in a script with interactive off, the vmin and
    John> vmax attrs are not updated from None to their True values.
    John> You can fix this either by working in interactive mode or by
    John> forcing a draw

Here is a more elegant solution: you don't need to force a draw, you
just need to force an autoscale (which draw does)

    from pylab import imshow, rand, show

    im = imshow(rand(10,10))
    im.autoscale()
    print im.norm.vmin, im.norm.vmax

Now you can pass the vmin and vmax attrs to your new image and expect
them to work.  FYI, in your case before when you were passing None,
imshow interprets None to mean "autoscale".  

JDH

>>>>> "Yogesh" == Yogesh Wadadekar <wad...@st...> writes:

    Yogesh> Hi,

    Yogesh> I encounter problems with the solution below. My test case
    Yogesh> from yesterday was really not a 'test case' because both
    Yogesh> images had near identical values. Thus, the auto vmin,vmax
    Yogesh> settings were the same in both images.

I think the difference we are seeing may due to the fact that the norm
attributes are not set until the figure is drawn.  So in a script with
interactive off, the vmin and vmax attrs are not updated from None to
their True values.  You can fix this either by working in interactive
mode or by forcing a draw

  from pylab import imshow, draw, rand, show

  im = imshow(rand(10,10))
  #draw()
  print im.norm.vmin, im.norm.vmax
  show()

Try the script above with draw commented and uncommented.

JDH

Hi,

I encounter problems with the solution below. My test case from yesterday
was really not a 'test case' because both images had near identical
values. Thus, the auto vmin,vmax settings were the same in both images.

The problem seems to be with the type of im.norm.vmin

>   In [3]: im = imshow(rand(20,20))

 In [1]: im = imshow(rand(20,20))

>   In [4]: im.norm.vmin
>   Out[4]: 0.001056874287314713

In [2]: im.norm.vmin

In [3]: type(im.norm.vmin)
Out[3]: <type 'NoneType'>

>
>   In [5]: im.norm.vmax
>   Out[5]: 0.99817508459091187

In [4]: im.norm.vmax

In [5]: type(im.norm.vmax)
Out[5]: <type 'NoneType'>

>
>   In [6]: newim = imshow(rand(30,30), vmin=im.norm.vmin, vmax=im.norm.vmax)

In [7]: newim = imshow(rand(30,30), vmin=im.norm.vmin, vmax=im.norm.vmax)

This works but im.norm.vmin and im.norm.vmax are not used!

I am using matplotlib-0.87.1 and Python 2.3.4 on Fedora Core 3.

Yogesh

>>>>> "manouchk" == manouchk  <man...@gm...> writes:

    manouchk> Simply reverting Numpy from version 0.9.6 to version
    manouchk> 0.9.4 solved the problem!  Does someone understand
    manouchk> something to that?

If you are building from source, you need to do a clean rebuild after
upgrading numpy

 > cd matplotlib_src
 > sudo rm -rf build
 > sudo python setup.py install

If you still encounter problems after that, be sure to let us know.

JDH

Simply reverting Numpy from version 0.9.6 to version 0.9.4 solved the probl=
em!=20
Does someone understand something to that?

Le Mardi 14 Mars 2006 10:19, manouchk a =E9crit=A0:
> Hi,
>
> I was usin matplotlib 0.86.2 and upgraded to 0.87.1 but it segfaults when
> launching simple_plot.py from 0.87 examples :
>
> python -v -i simple_plot.py
<
> it ends by :
> ...
> # /usr/lib/python2.4/site-packages/matplotlib/_transforms.pyc
> matches /usr/lib/python2.4/site-packages/matplotlib/_transforms.py
> import matplotlib._transforms # precompiled
> from /usr/lib/python2.4/site-packages/matplotlib/_transforms.pyc
> dlopen("/usr/lib/python2.4/site-packages/matplotlib/_ns_transforms.so", 2=
);
> Segmentation fault
>
>
> I don't understand the problem ? Is it a known problem of version of
> 0.87.1? Could I provide more specific informations?
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by xPML, a groundbreaking scripting langua=
ge
> that extends applications into web and mobile media. Attend the live
> webcast and join the prime developer group breaking into this new coding
> territory!
> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D110944&bid=3D241720&dat=
=3D121642
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Hi,

I was usin matplotlib 0.86.2 and upgraded to 0.87.1 but it segfaults when 
launching simple_plot.py from 0.87 examples :

python -v -i simple_plot.py

it ends by :
...
# /usr/lib/python2.4/site-packages/matplotlib/_transforms.pyc 
matches /usr/lib/python2.4/site-packages/matplotlib/_transforms.py
import matplotlib._transforms # precompiled 
from /usr/lib/python2.4/site-packages/matplotlib/_transforms.pyc
dlopen("/usr/lib/python2.4/site-packages/matplotlib/_ns_transforms.so", 2);
Segmentation fault


I don't understand the problem ? Is it a known problem of version of 0.87.1? 
Could I provide more specific informations?

How hard would it be to write a "shrinkwrap" function that resized  
the figure to fit the graphics
plus a little bit a white space border.

I find that I have to twiddle a lot to get the plots to fit nicely in  
the figure. In almost every case
a shrink wrap function that resized the figure to exactly match the  
maximum extents of the plots plus legends plus labels
plus a little bit of border would be just right.

Is there a straightforward way to find out the max and min extent of  
everything in the figure?

What I am thinking of, is that the initial figure size would be just  
a suggestion to bootstrap the plot sizes.
Once all the plotting was done, one could call shrinkwrap and the  
figure would offset and shrink(expand) to just fit the contents (with  
a programmable amount of white space all around).



**********************************************************************
Samuel M. Smith Ph.D.
2966 Fort Hill Road
Eagle Mountain, Utah 84043
801-768-2768 voice
801-768-2769 fax
**********************************************************************
"The greatest source of failure and unhappiness in the world is
giving up what we want most for what we want at the moment"
**********************************************************************