matplotlib-users

[Matplotlib-users] imshow for .png- low resultion

[Asma R. <asm...@gm...>]

Hi
I am trying to merge a heat map(matplotlib) with a tree(.png), but the .png
does not plot as needed or for that matter cannot be seen easily. Here is
my code:

####### CODE
 cmap = mpl.cm.hot
    if outlier==0:
       threshold=1
    else:
       threshold=outlier-0.01

    fig = plt.figure(figsize=(25,25))
    plt.suptitle(file_handle.replace('.csv',''),fontsize=22)
    cmap.set_over('green')
    cmap.set_under('grey')
    gs=gridspec.GridSpec(1, 2,height_ratios=[1,1,-2,2]
,width_ratios=[1,1,-2,2])
    phyl_ax=plt.subplot(gs[0],frameon=False)
    ht_ax=plt.subplot(gs[1])
    ##gs.tight_layout(fig, rect=[1,1,2,2])
    ht_ax.set_xlim(0,35)
    ht_ax.set_ylim(0,35)
    ##ht_ax.grid(color='white',linestyle='-',linewidth=1)
    fig.subplots_adjust(hspace=0)
    cb_ax,kw =mpl.colorbar.make_axes(ht_ax, shrink=0.65)


    plt.setp(phyl_ax.get_xticklabels(),visible=False)
    plt.setp(phyl_ax.get_yticklabels(),visible=False)
    plt.setp(ht_ax.get_xticklabels(),visible=True)
    plt.setp(ht_ax.get_yticklabels(),visible=True)
    plt.setp(phyl_ax.get_xticklines(),visible=False)
    plt.setp(phyl_ax.get_yticklines(),visible=False)
    plt.setp(ht_ax.get_xticklines(),visible=True)
    plt.setp(ht_ax.get_yticklines(),visible=True)


    img = ht_ax.imshow(data, cmap=cmap, interpolation='none',vmax=threshold)
    cb = mpl.colorbar.ColorbarBase(ax=cb_ax,cmap=cmap,
extend='neither',**kw)
    cb.cmap.set_over('green')
    image_file =
cbook.get_sample_data('/home/asmariyaz/Desktop/mytree.png',asfileobj=False)
    image = plt.imread(image_file)

How can I make the resolution of the .png image better OR for that matter
is there a better solution to merge these together?

I am attaching the plot I am getting with this code- (WrongTree.pdf)

Thanks
Asma

Re: [Matplotlib-users] imshow for .png- low resultion

[Paul H. <pmh...@gm...>]

Olga Botnivik is doing some work with these types of figures in her fork of
the seaborn project.
Example here: http://nbviewer.ipython.org/gist/olgabot/8341784
Link to the PR in github: https://github.com/mwaskom/seaborn/pull/73

Those might be a good place to start.



On Wed, Mar 5, 2014 at 12:19 PM, Asma Riyaz <asm...@gm...> wrote:

> Hi
> I am trying to merge a heat map(matplotlib) with a tree(.png), but the
> .png does not plot as needed or for that matter cannot be seen easily. Here
> is my code:
>
> ####### CODE
>  cmap = mpl.cm.hot
>     if outlier==0:
>        threshold=1
>     else:
>        threshold=outlier-0.01
>
>     fig = plt.figure(figsize=(25,25))
>     plt.suptitle(file_handle.replace('.csv',''),fontsize=22)
>     cmap.set_over('green')
>     cmap.set_under('grey')
>     gs=gridspec.GridSpec(1, 2,height_ratios=[1,1,-2,2]
> ,width_ratios=[1,1,-2,2])
>     phyl_ax=plt.subplot(gs[0],frameon=False)
>     ht_ax=plt.subplot(gs[1])
>     ##gs.tight_layout(fig, rect=[1,1,2,2])
>     ht_ax.set_xlim(0,35)
>     ht_ax.set_ylim(0,35)
>     ##ht_ax.grid(color='white',linestyle='-',linewidth=1)
>     fig.subplots_adjust(hspace=0)
>     cb_ax,kw =mpl.colorbar.make_axes(ht_ax, shrink=0.65)
>
>
>     plt.setp(phyl_ax.get_xticklabels(),visible=False)
>     plt.setp(phyl_ax.get_yticklabels(),visible=False)
>     plt.setp(ht_ax.get_xticklabels(),visible=True)
>     plt.setp(ht_ax.get_yticklabels(),visible=True)
>     plt.setp(phyl_ax.get_xticklines(),visible=False)
>     plt.setp(phyl_ax.get_yticklines(),visible=False)
>     plt.setp(ht_ax.get_xticklines(),visible=True)
>     plt.setp(ht_ax.get_yticklines(),visible=True)
>
>
>     img = ht_ax.imshow(data, cmap=cmap,
> interpolation='none',vmax=threshold)
>     cb = mpl.colorbar.ColorbarBase(ax=cb_ax,cmap=cmap,
> extend='neither',**kw)
>     cb.cmap.set_over('green')
>     image_file =
> cbook.get_sample_data('/home/asmariyaz/Desktop/mytree.png',asfileobj=False)
>     image = plt.imread(image_file)
>
> How can I make the resolution of the .png image better OR for that matter
> is there a better solution to merge these together?
>
> I am attaching the plot I am getting with this code- (WrongTree.pdf)
>
> Thanks
> Asma
>
>
> ------------------------------------------------------------------------------
> Subversion Kills Productivity. Get off Subversion & Make the Move to
> Perforce.
> With Perforce, you get hassle-free workflows. Merge that actually works.
> Faster operations. Version large binaries.  Built-in WAN optimization and
> the
> freedom to use Git, Perforce or both. Make the move to Perforce.
>
> http://pubads.g.doubleclick.net/gampad/clk?id=122218951&iu=/4140/ostg.clktrk
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>

Re: [Matplotlib-users] imshow for .png- low resultion

[Asma R. <asm...@gm...>]

Thank you for replying!!

The tree I generate is not based on clustering of matrices but rather on
sequence data and hence the tree generation is different in this case. I am
still looking into the seaborn project if it can lead me anywhere.

Thanks
Asma


On Wed, Mar 5, 2014 at 3:34 PM, Paul Hobson <pmh...@gm...> wrote:

> Olga Botnivik is doing some work with these types of figures in her fork
> of the seaborn project.
> Example here: http://nbviewer.ipython.org/gist/olgabot/8341784
> Link to the PR in github: https://github.com/mwaskom/seaborn/pull/73
>
> Those might be a good place to start.
>
>
>
> On Wed, Mar 5, 2014 at 12:19 PM, Asma Riyaz <asm...@gm...> wrote:
>
>> Hi
>> I am trying to merge a heat map(matplotlib) with a tree(.png), but the
>> .png does not plot as needed or for that matter cannot be seen easily. Here
>> is my code:
>>
>> ####### CODE
>>  cmap = mpl.cm.hot
>>     if outlier==0:
>>        threshold=1
>>     else:
>>        threshold=outlier-0.01
>>
>>     fig = plt.figure(figsize=(25,25))
>>     plt.suptitle(file_handle.replace('.csv',''),fontsize=22)
>>     cmap.set_over('green')
>>     cmap.set_under('grey')
>>     gs=gridspec.GridSpec(1, 2,height_ratios=[1,1,-2,2]
>> ,width_ratios=[1,1,-2,2])
>>     phyl_ax=plt.subplot(gs[0],frameon=False)
>>     ht_ax=plt.subplot(gs[1])
>>     ##gs.tight_layout(fig, rect=[1,1,2,2])
>>     ht_ax.set_xlim(0,35)
>>     ht_ax.set_ylim(0,35)
>>     ##ht_ax.grid(color='white',linestyle='-',linewidth=1)
>>     fig.subplots_adjust(hspace=0)
>>     cb_ax,kw =mpl.colorbar.make_axes(ht_ax, shrink=0.65)
>>
>>
>>     plt.setp(phyl_ax.get_xticklabels(),visible=False)
>>     plt.setp(phyl_ax.get_yticklabels(),visible=False)
>>     plt.setp(ht_ax.get_xticklabels(),visible=True)
>>     plt.setp(ht_ax.get_yticklabels(),visible=True)
>>     plt.setp(phyl_ax.get_xticklines(),visible=False)
>>     plt.setp(phyl_ax.get_yticklines(),visible=False)
>>     plt.setp(ht_ax.get_xticklines(),visible=True)
>>     plt.setp(ht_ax.get_yticklines(),visible=True)
>>
>>
>>     img = ht_ax.imshow(data, cmap=cmap,
>> interpolation='none',vmax=threshold)
>>     cb = mpl.colorbar.ColorbarBase(ax=cb_ax,cmap=cmap,
>> extend='neither',**kw)
>>     cb.cmap.set_over('green')
>>     image_file =
>> cbook.get_sample_data('/home/asmariyaz/Desktop/mytree.png',asfileobj=False)
>>     image = plt.imread(image_file)
>>
>> How can I make the resolution of the .png image better OR for that matter
>> is there a better solution to merge these together?
>>
>> I am attaching the plot I am getting with this code- (WrongTree.pdf)
>>
>> Thanks
>> Asma
>>
>>
>> ------------------------------------------------------------------------------
>> Subversion Kills Productivity. Get off Subversion & Make the Move to
>> Perforce.
>> With Perforce, you get hassle-free workflows. Merge that actually works.
>> Faster operations. Version large binaries.  Built-in WAN optimization and
>> the
>> freedom to use Git, Perforce or both. Make the move to Perforce.
>>
>> http://pubads.g.doubleclick.net/gampad/clk?id=122218951&iu=/4140/ostg.clktrk
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>
>>
>

Re: [Matplotlib-users] imshow for .png- low resultion

[Pierre H. <pie...@cr...>]

Hi Asma,
Le 05/03/2014 21:19, Asma Riyaz a écrit :
> I am trying to merge a heat map(matplotlib) with a tree(.png), but the 
> .png does not plot as needed or for that matter cannot be seen easily. 
> Here is my code:
>
> ####### CODE
>  [....]
>    img = ht_ax.imshow(data, cmap=cmap, 
> interpolation='none',vmax=threshold)
>    [....]
>
> How can I make the resolution of the .png image better OR for that 
> matter is there a better solution to merge these together?
>
I see the imshow call to plot the matrix but not the one you use to plot 
the png image. Could it be that the poor image rendering is due to a bad 
choice for the interpolation param ?

best,
Pierre

Re: [Matplotlib-users] imshow for .png- low resultion

[Asma R. <asm...@gm...>]

 Sorry I somehow forgot to paste the imshow call to plot the .png
image(pasted the entire code again), I am trying the different
interpolation parameters in the mean time as suggested by Pierre

   data=np.array(full_len)
    cmap = mpl.cm.hot
    if outlier==0:
       threshold=1
    else:
       threshold=outlier-0.01

    fig = plt.figure(figsize=(25,25))
    plt.suptitle(file_handle.replace('.csv',''),fontsize=22)
    cmap.set_over('green')
    cmap.set_under('grey')
    gs=gridspec.GridSpec(1, 2,height_ratios=[1,1,-2,2]
,width_ratios=[1,1,-2,2])
    gs.tight_layout(fig,)
    phyl_ax=plt.subplot(gs[0]) ##,frameon=False)##
    ht_ax=plt.subplot(gs[1],sharey=phyl_ax)
    ##gs.tight_layout(fig, rect=[1,1,2,2])
    ht_ax.set_xlim(0,35)
    ht_ax.set_ylim(0,35)
    ##ht_ax.grid(color='white',linestyle='-',linewidth=1)
    ##fig.subplots_adjust(hspace=0)
    cb_ax,kw =mpl.colorbar.make_axes(ht_ax, shrink=0.65)

    plt.setp(phyl_ax.get_xticklabels(),visible=False)
    plt.setp(phyl_ax.get_yticklabels(),visible=False)
    plt.setp(ht_ax.get_xticklabels(),visible=True)
    plt.setp(ht_ax.get_yticklabels(),visible=True)
    plt.setp(phyl_ax.get_xticklines(),visible=False)
    plt.setp(phyl_ax.get_yticklines(),visible=False)
    plt.setp(ht_ax.get_xticklines(),visible=True)
    plt.setp(ht_ax.get_yticklines(),visible=True)


    img = ht_ax.imshow(data, cmap=cmap, interpolation='none',vmax=threshold)
    cb = mpl.colorbar.ColorbarBase(ax=cb_ax,cmap=cmap,
extend='neither',**kw)
    cb.cmap.set_over('green')
    ##image_file =
cbook.get_sample_data('/home/asmariyaz/Desktop/mytree.png',asfileobj=False)
    ##image = plt.imread(image_file)
    img= mpimg.imread('/home/asmariyaz/Desktop/mytree.png')
    phyl_ax.imshow(img,interpolation='nearest')

heatmap_file=fig.savefig('/home/asmariyaz/Desktop/heatmap/'+file_handle+'.pdf',bbox_inches='tight')


On Wed, Mar 5, 2014 at 3:27 PM, Pierre Haessig <pie...@cr...>wrote:

>  Hi Asma,
> Le 05/03/2014 21:19, Asma Riyaz a écrit :
>
>  I am trying to merge a heat map(matplotlib) with a tree(.png), but the
> .png does not plot as needed or for that matter cannot be seen easily. Here
> is my code:
>
>  ####### CODE
>   [....]
>
>    img = ht_ax.imshow(data, cmap=cmap, interpolation='none',vmax=threshold)
>    [....]
>
>  How can I make the resolution of the .png image better OR for that
> matter is there a better solution to merge these together?
>
>   I see the imshow call to plot the matrix but not the one you use to
> plot the png image. Could it be that the poor image rendering is due to a
> bad choice for the interpolation param ?
>
> best,
> Pierre
>
>
> ------------------------------------------------------------------------------
> Subversion Kills Productivity. Get off Subversion & Make the Move to
> Perforce.
> With Perforce, you get hassle-free workflows. Merge that actually works.
> Faster operations. Version large binaries.  Built-in WAN optimization and
> the
> freedom to use Git, Perforce or both. Make the move to Perforce.
>
> http://pubads.g.doubleclick.net/gampad/clk?id=122218951&iu=/4140/ostg.clktrk
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>

Re: [Matplotlib-users] imshow for .png- low resultion

[Eric F. <ef...@ha...>]

On 2014/03/05 11:37 AM, Asma Riyaz wrote:
>   Sorry I somehow forgot to paste the imshow call to plot the .png
> image(pasted the entire code again), I am trying the different
> interpolation parameters in the mean time as suggested by Pierre
>
> data=np.array(full_len)
>      cmap = mpl.cm.hot
>      if outlier==0:
>         threshold=1
>      else:
>         threshold=outlier-0.01
>
>      fig = plt.figure(figsize=(25,25))
>      plt.suptitle(file_handle.replace('.csv',''),fontsize=22)
>      cmap.set_over('green')
>      cmap.set_under('grey')
>      gs=gridspec.GridSpec(1, 2,height_ratios=[1,1,-2,2]
> ,width_ratios=[1,1,-2,2])
>      gs.tight_layout(fig,)
>      phyl_ax=plt.subplot(gs[0]) ##,frameon=False)##
>      ht_ax=plt.subplot(gs[1],sharey=phyl_ax)
>      ##gs.tight_layout(fig, rect=[1,1,2,2])
>      ht_ax.set_xlim(0,35)
>      ht_ax.set_ylim(0,35)
>      ##ht_ax.grid(color='white',linestyle='-',linewidth=1)
>      ##fig.subplots_adjust(hspace=0)
>      cb_ax,kw =mpl.colorbar.make_axes(ht_ax, shrink=0.65)
>
>      plt.setp(phyl_ax.get_xticklabels(),visible=False)
>      plt.setp(phyl_ax.get_yticklabels(),visible=False)
>      plt.setp(ht_ax.get_xticklabels(),visible=True)
>      plt.setp(ht_ax.get_yticklabels(),visible=True)
>      plt.setp(phyl_ax.get_xticklines(),visible=False)
>      plt.setp(phyl_ax.get_yticklines(),visible=False)
>      plt.setp(ht_ax.get_xticklines(),visible=True)
>      plt.setp(ht_ax.get_yticklines(),visible=True)
>
>
>      img = ht_ax.imshow(data, cmap=cmap,
> interpolation='none',vmax=threshold)
>      cb = mpl.colorbar.ColorbarBase(ax=cb_ax,cmap=cmap,
> extend='neither',**kw)
>      cb.cmap.set_over('green')
>      ##image_file =
> cbook.get_sample_data('/home/asmariyaz/Desktop/mytree.png',asfileobj=False)
>      ##image = plt.imread(image_file)
>      img= mpimg.imread('/home/asmariyaz/Desktop/mytree.png')
>      phyl_ax.imshow(img,interpolation='nearest')
>
> heatmap_file=fig.savefig('/home/asmariyaz/Desktop/heatmap/'+file_handle+'.pdf',bbox_inches='tight')

Try giving savefig a dpi kwarg to increase the resolution with which the 
image is saved in the pdf?  Turn off anti-aliasing when you make the png?

Eric

Re: [Matplotlib-users] imshow for .png- low resultion

[Asma R. <asm...@gm...>]

I am sorry but I don't understand the meaning of anti-aliasing? Do you mean
on this line
---->fig.savefig('/home/asmariyaz/Desktop/heatmap/'+file_handle+'.pdf',bbox_inches='tight',dpi=100),
where I am giving it a new path+name? I made the following changes:
aspect='auto', interpolation='none' and I have attached the result


On Wed, Mar 5, 2014 at 4:49 PM, Eric Firing <ef...@ha...> wrote:

> On 2014/03/05 11:37 AM, Asma Riyaz wrote:
> >   Sorry I somehow forgot to paste the imshow call to plot the .png
> > image(pasted the entire code again), I am trying the different
> > interpolation parameters in the mean time as suggested by Pierre
> >
> > data=np.array(full_len)
> >      cmap = mpl.cm.hot
> >      if outlier==0:
> >         threshold=1
> >      else:
> >         threshold=outlier-0.01
> >
> >      fig = plt.figure(figsize=(25,25))
> >      plt.suptitle(file_handle.replace('.csv',''),fontsize=22)
> >      cmap.set_over('green')
> >      cmap.set_under('grey')
> >      gs=gridspec.GridSpec(1, 2,height_ratios=[1,1,-2,2]
> > ,width_ratios=[1,1,-2,2])
> >      gs.tight_layout(fig,)
> >      phyl_ax=plt.subplot(gs[0]) ##,frameon=False)##
> >      ht_ax=plt.subplot(gs[1],sharey=phyl_ax)
> >      ##gs.tight_layout(fig, rect=[1,1,2,2])
> >      ht_ax.set_xlim(0,35)
> >      ht_ax.set_ylim(0,35)
> >      ##ht_ax.grid(color='white',linestyle='-',linewidth=1)
> >      ##fig.subplots_adjust(hspace=0)
> >      cb_ax,kw =mpl.colorbar.make_axes(ht_ax, shrink=0.65)
> >
> >      plt.setp(phyl_ax.get_xticklabels(),visible=False)
> >      plt.setp(phyl_ax.get_yticklabels(),visible=False)
> >      plt.setp(ht_ax.get_xticklabels(),visible=True)
> >      plt.setp(ht_ax.get_yticklabels(),visible=True)
> >      plt.setp(phyl_ax.get_xticklines(),visible=False)
> >      plt.setp(phyl_ax.get_yticklines(),visible=False)
> >      plt.setp(ht_ax.get_xticklines(),visible=True)
> >      plt.setp(ht_ax.get_yticklines(),visible=True)
> >
> >
> >      img = ht_ax.imshow(data, cmap=cmap,
> > interpolation='none',vmax=threshold)
> >      cb = mpl.colorbar.ColorbarBase(ax=cb_ax,cmap=cmap,
> > extend='neither',**kw)
> >      cb.cmap.set_over('green')
> >      ##image_file =
> >
> cbook.get_sample_data('/home/asmariyaz/Desktop/mytree.png',asfileobj=False)
> >      ##image = plt.imread(image_file)
> >      img= mpimg.imread('/home/asmariyaz/Desktop/mytree.png')
> >      phyl_ax.imshow(img,interpolation='nearest')
> >
> >
> heatmap_file=fig.savefig('/home/asmariyaz/Desktop/heatmap/'+file_handle+'.pdf',bbox_inches='tight')
>
> Try giving savefig a dpi kwarg to increase the resolution with which the
> image is saved in the pdf?  Turn off anti-aliasing when you make the png?
>
> Eric
>
>
>
>
> ------------------------------------------------------------------------------
> Subversion Kills Productivity. Get off Subversion & Make the Move to
> Perforce.
> With Perforce, you get hassle-free workflows. Merge that actually works.
> Faster operations. Version large binaries.  Built-in WAN optimization and
> the
> freedom to use Git, Perforce or both. Make the move to Perforce.
>
> http://pubads.g.doubleclick.net/gampad/clk?id=122218951&iu=/4140/ostg.clktrk
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Re: [Matplotlib-users] imshow for .png- low resultion

[Pierre H. <pie...@cr...>]

Le 05/03/2014 22:37, Asma Riyaz a écrit :
>     img= mpimg.imread('/home/asmariyaz/Desktop/mytree.png')
>     phyl_ax.imshow(img,interpolation='nearest')
Ok, so here you could try replace 'nearest' by 'bilinear' or 'bicubic'.
I believe those are the most common choices for image resampling
(because when you plot an image and then save it, there is a resampling
involved).
(http://matplotlib.org/api/pyplot_api.html#matplotlib.pyplot.imshow for
the other options)

Of course, it's also worth playing the dpi argument of savefig, as
suggested by Eric.

best,
Pierre

Re: [Matplotlib-users] imshow for .png- low resultion

[Asma R. <asm...@gm...>]

Yes the interpolation had to be set to nearest or bicubic or bilinear for
it to function properly!!
Thanks a lot!!
Asma


On Thu, Mar 6, 2014 at 4:48 AM, Pierre Haessig <pie...@cr...>wrote:

>  Le 05/03/2014 22:37, Asma Riyaz a écrit :
>
>     img= mpimg.imread('/home/asmariyaz/Desktop/mytree.png')
>     phyl_ax.imshow(img,interpolation='nearest')
>
> Ok, so here you could try replace 'nearest' by 'bilinear' or 'bicubic'. I
> believe those are the most common choices for image resampling (because
> when you plot an image and then save it, there is a resampling involved).
> (http://matplotlib.org/api/pyplot_api.html#matplotlib.pyplot.imshow for
> the other options)
>
> Of course, it's also worth playing the dpi argument of savefig, as
> suggested by Eric.
>
> best,
> Pierre
>
>
> ------------------------------------------------------------------------------
> Subversion Kills Productivity. Get off Subversion & Make the Move to
> Perforce.
> With Perforce, you get hassle-free workflows. Merge that actually works.
> Faster operations. Version large binaries.  Built-in WAN optimization and
> the
> freedom to use Git, Perforce or both. Make the move to Perforce.
>
> http://pubads.g.doubleclick.net/gampad/clk?id=122218951&iu=/4140/ostg.clktrk
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>