matplotlib-users Mailing List for matplotlib

By the way, soon I will have a book out on creating interactive
applications using matplotlib. I devote an entire chapter to clearly (I
hope!) demonstrating how to use particular GUI toolkits with matplotlib. It
is currently going through the final pre-publishing steps, so I should be
announcing its release fairly soon.

Cheers!
Ben Root


On Wed, Jan 28, 2015 at 2:42 PM, Sterling Smith <sm...@fu...>
wrote:

> Did you come across
> http://matplotlib.org/examples/user_interfaces/embedding_in_qt4.html
> ?
> -Sterling
>
> On Jan 28, 2015, at 11:38AM, mo...@po... wrote:
>
> > On 2015-01-28 16:25 Thomas Caswell <tca...@gm...> wrote:
> >> It should be exactly the same as for Qt4, just importing from the Qt5
> >> version
> >
> > Sorry, but my question was to unspecific.
> >
> > I don't know how it worked with Qt4.
> >
> > I need to know I can "draw" a plot into a QWidget or something else. I
> > couldn't find an example or documentation about it.
> >
> >
> ------------------------------------------------------------------------------
> > Dive into the World of Parallel Programming. The Go Parallel Website,
> > sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> > hub for all things parallel software development, from weekly thought
> > leadership blogs to news, videos, case studies, tutorials and more. Take
> a
> > look and join the conversation now. http://goparallel.sourceforge.net/
> > _______________________________________________
> > Matplotlib-users mailing list
> > Mat...@li...
> > https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming. The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Even deprecated, mpl.finance can do some of  the things you need.
Check these tutotials, and others, here:
https://www.youtube.com/playlist?list=PLQVvvaa0QuDc2QjQOkZ4rtLYZVll_sZFZ
There is also something about live stream and memory usage concern with
matplotlib when reloading the charts, somewhere.
Regards,
Milcent


Em Fri Jan 30 2015 at 21:03:56, Benjamin Root <ben...@ou...> escreveu:

> To be clear, most of what was in finance.py were convenience functions for
> parsing through stock data from the yahoo interface, and for plotting.
> Taking a quick look, perhaps it might make sense to pull out a couple
> fundamental chart types, but most of the code are just simply convenience
> wrappers, and largely outdated now.
>
> The problem we are having is that users would file bug reports saying that
> we were doing a particular chart incorrectly, and none of us had any domain
> knowledge to know if that was the truth, false, or just a matter of opinion
> in the field. Matplotlib is also intended to be a general-purpose plotting
> library. It really shouldn't be doing much of the data preparatory work.
> You should just tell it what to plot and let it crank. If that data
> happened to have used a moving window, it wouldn't matter if it was an
> average, median, or what-have-you. Matplotlib is fairly low-level, and
> works very well that way. For example, there isn't an
> "streamed_data_plot()" function. You have to do the streaming yourself and
> update the plot.
>
> For data wrangling, I would suggest using Pandas. It interfaces quite
> nicely (mostly) with matplotlib, and it is considered the de facto tool to
> use for time series statistical analyses. Once the data is mashed into the
> form you need, then you can plot it however you like.
>
> If you really want to keep finance.py alive, then all we are looking for
> is someone knowledgable to stand up and take responsibility for it. It is
> mostly self-contained, so it is even possible to spin it off as its own
> mpl_toolkit managed separately from matplotlib.
>
> I mean, let's face it... do you really want your finances managed by a
> bunch of meteorologists and astrophysicists?  ;-P
>
> Cheers!
> Ben Root
>
>
> On Fri, Jan 30, 2015 at 1:30 PM, Thomas Caswell <tca...@gm...>
> wrote:
>
>> Boris,
>>
>> Please direct such questions to the user list in the future (I have
>> included the list on my reply).  You may need to join the list to be able
>> to post.
>>
>> The reason that mpl.finance was deprecated is that none of the current
>> core developers work in finance and hence do not have the domain expertise
>> to maintain the module.  We are currently looking for a volunteer to take
>> responsibility for that bit of code.
>>
>> I have no experience with plotting financial data and am hesitant to
>> speculate about performance, but have gotten 20-30 fps out of mpl on other
>> applications.
>>
>> Tom
>>
>> On Fri Jan 30 2015 at 1:19:47 PM tbad <tba...@gm...> wrote:
>>
>>> Hello Thomas,
>>>
>>> - You write in Github that matplotlib finance is deprecated since
>>> matplotlib 1.4. What is instead of it?
>>> - Can you please advice on what to use if i want to use matplotlib for
>>> charting stocks, volume in different formats like candlesticks, tick
>>> charts, bar charts with technical analysis add-ons like moving averages
>>> etc.? I want to have charts with intra-day real time streaming with a lot
>>> of stocks (i have a data feed provider). Or the only way is to do
>>> everything manually?
>>> - And will matplotlib be capable of streaming a lot of stocks tick by
>>> tick without delays?
>>> - Also maybe there are other python packages useful for trading purposes?
>>>
>>> Thank you for any answers beforehand,
>>> Best regards,
>>> Boris
>>>
>>
>>
>> ------------------------------------------------------------------------------
>> Dive into the World of Parallel Programming. The Go Parallel Website,
>> sponsored by Intel and developed in partnership with Slashdot Media, is
>> your
>> hub for all things parallel software development, from weekly thought
>> leadership blogs to news, videos, case studies, tutorials and more. Take a
>> look and join the conversation now. http://goparallel.sourceforge.net/
>> _______________________________________________
>> Matplotlib-users mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>>
>>
> ------------------------------------------------------------
> ------------------
> Dive into the World of Parallel Programming. The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

To be clear, most of what was in finance.py were convenience functions for
parsing through stock data from the yahoo interface, and for plotting.
Taking a quick look, perhaps it might make sense to pull out a couple
fundamental chart types, but most of the code are just simply convenience
wrappers, and largely outdated now.

The problem we are having is that users would file bug reports saying that
we were doing a particular chart incorrectly, and none of us had any domain
knowledge to know if that was the truth, false, or just a matter of opinion
in the field. Matplotlib is also intended to be a general-purpose plotting
library. It really shouldn't be doing much of the data preparatory work.
You should just tell it what to plot and let it crank. If that data
happened to have used a moving window, it wouldn't matter if it was an
average, median, or what-have-you. Matplotlib is fairly low-level, and
works very well that way. For example, there isn't an
"streamed_data_plot()" function. You have to do the streaming yourself and
update the plot.

For data wrangling, I would suggest using Pandas. It interfaces quite
nicely (mostly) with matplotlib, and it is considered the de facto tool to
use for time series statistical analyses. Once the data is mashed into the
form you need, then you can plot it however you like.

If you really want to keep finance.py alive, then all we are looking for is
someone knowledgable to stand up and take responsibility for it. It is
mostly self-contained, so it is even possible to spin it off as its own
mpl_toolkit managed separately from matplotlib.

I mean, let's face it... do you really want your finances managed by a
bunch of meteorologists and astrophysicists?  ;-P

Cheers!
Ben Root


On Fri, Jan 30, 2015 at 1:30 PM, Thomas Caswell <tca...@gm...> wrote:

> Boris,
>
> Please direct such questions to the user list in the future (I have
> included the list on my reply).  You may need to join the list to be able
> to post.
>
> The reason that mpl.finance was deprecated is that none of the current
> core developers work in finance and hence do not have the domain expertise
> to maintain the module.  We are currently looking for a volunteer to take
> responsibility for that bit of code.
>
> I have no experience with plotting financial data and am hesitant to
> speculate about performance, but have gotten 20-30 fps out of mpl on other
> applications.
>
> Tom
>
> On Fri Jan 30 2015 at 1:19:47 PM tbad <tba...@gm...> wrote:
>
>> Hello Thomas,
>>
>> - You write in Github that matplotlib finance is deprecated since
>> matplotlib 1.4. What is instead of it?
>> - Can you please advice on what to use if i want to use matplotlib for
>> charting stocks, volume in different formats like candlesticks, tick
>> charts, bar charts with technical analysis add-ons like moving averages
>> etc.? I want to have charts with intra-day real time streaming with a lot
>> of stocks (i have a data feed provider). Or the only way is to do
>> everything manually?
>> - And will matplotlib be capable of streaming a lot of stocks tick by
>> tick without delays?
>> - Also maybe there are other python packages useful for trading purposes?
>>
>> Thank you for any answers beforehand,
>> Best regards,
>> Boris
>>
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming. The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>

Boris,

Please direct such questions to the user list in the future (I have
included the list on my reply).  You may need to join the list to be able
to post.

The reason that mpl.finance was deprecated is that none of the current core
developers work in finance and hence do not have the domain expertise to
maintain the module.  We are currently looking for a volunteer to take
responsibility for that bit of code.

I have no experience with plotting financial data and am hesitant to
speculate about performance, but have gotten 20-30 fps out of mpl on other
applications.

Tom

On Fri Jan 30 2015 at 1:19:47 PM tbad <tba...@gm...> wrote:

> Hello Thomas,
>
> - You write in Github that matplotlib finance is deprecated since
> matplotlib 1.4. What is instead of it?
> - Can you please advice on what to use if i want to use matplotlib for
> charting stocks, volume in different formats like candlesticks, tick
> charts, bar charts with technical analysis add-ons like moving averages
> etc.? I want to have charts with intra-day real time streaming with a lot
> of stocks (i have a data feed provider). Or the only way is to do
> everything manually?
> - And will matplotlib be capable of streaming a lot of stocks tick by tick
> without delays?
> - Also maybe there are other python packages useful for trading purposes?
>
> Thank you for any answers beforehand,
> Best regards,
> Boris
>

Unfortunately, no. I have made several attempts at implementing this, with
the last attempt during the sprints at SciPy 2014. I got tantalizingly
close, but it fell apart as I tried to bring all the pieces back together.
The mplot3d code is actually very old and doesn't use the transforms system
very well, nor does the transform system extend nicely to 3d.

If anybody wants to re-attack this problem, I would suggest by starting
with creating some 3d->2d projection classes to be added to the transforms
system, and get unit tests for these. From there, axis3d and art3d needs to
be piece-by-piece remade to properly build on the trasnforms stack.

Sorry for the lack of happier news. The only work-around that I can think
of is to apply the log scaling yourself to the data prior to plotting.

Ben Root


On Thu, Jan 29, 2015 at 5:30 PM, John Ladasky <joh...@sb...>
wrote:

>  The following minimal code example illustrates a problem I'm having.
>
> import matplotlib.pyplot as plt
> from mpl_toolkits.mplot3d import Axes3D
>
> # succeeds
> fig = plt.figure()
> ax = fig.add_subplot(111, projection='3d')
> plt.show()
>
> # succeeds
> fig = plt.figure()
> ax = fig.add_subplot(111)
> ax.set_xscale("log")
> plt.show()
>
> # fails
> fig = plt.figure()
> ax = fig.add_subplot(111, projection='3d')
> ax.set_xscale("log")
> plt.show()
>
>
> Here is the traceback:
>
> Exception in Tkinter callback
> Traceback (most recent call last):
>   File "/usr/lib/python3.4/tkinter/__init__.py", line 1490, in __call__
>     return self.func(*args)
>   File
> "/usr/lib/python3/dist-packages/matplotlib/backends/backend_tkagg.py", line
> 276, in resize
>     self.show()
>   File
> "/usr/lib/python3/dist-packages/matplotlib/backends/backend_tkagg.py", line
> 348, in draw
>     FigureCanvasAgg.draw(self)
>   File
> "/usr/lib/python3/dist-packages/matplotlib/backends/backend_agg.py", line
> 451, in draw
>     self.figure.draw(self.renderer)
>   File "/usr/lib/python3/dist-packages/matplotlib/artist.py", line 56, in
> draw_wrapper
>     draw(artist, renderer, *args, **kwargs)
>   File "/usr/lib/python3/dist-packages/matplotlib/figure.py", line 1035,
> in draw
>     func(*args)
>   File "/usr/lib/python3/dist-packages/mpl_toolkits/mplot3d/axes3d.py",
> line 267, in draw
>     ax.draw(renderer)
>   File "/usr/lib/python3/dist-packages/mpl_toolkits/mplot3d/axis3d.py",
> line 404, in draw
>     tick.draw(renderer)
>   File "/usr/lib/python3/dist-packages/matplotlib/artist.py", line 56, in
> draw_wrapper
>     draw(artist, renderer, *args, **kwargs)
>   File "/usr/lib/python3/dist-packages/matplotlib/axis.py", line 242, in
> draw
>     self.label1.draw(renderer)
>   File "/usr/lib/python3/dist-packages/matplotlib/artist.py", line 56, in
> draw_wrapper
>     draw(artist, renderer, *args, **kwargs)
>   File "/usr/lib/python3/dist-packages/matplotlib/text.py", line 599, in
> draw
>     ismath=ismath, mtext=self)
>   File
> "/usr/lib/python3/dist-packages/matplotlib/backends/backend_agg.py", line
> 169, in draw_text
>     return self.draw_mathtext(gc, x, y, s, prop, angle)
>   File
> "/usr/lib/python3/dist-packages/matplotlib/backends/backend_agg.py", line
> 158, in draw_mathtext
>     x = np.round(x + ox + xd)
>   File "/usr/lib/python3/dist-packages/numpy/core/fromnumeric.py", line
> 2629, in round_
>     return round(decimals, out)
>   File "/usr/lib/python3/dist-packages/numpy/ma/core.py", line 4855, in
> round
>     result._mask = self._mask
> AttributeError: 'numpy.float64' object has no attribute '_mask'
>
> I am hoping to use logarithmic axes in a 3D plot.  I haven't found
> anything in the documentation which says that I am restricted to linear
> scaling.  My system configuration is: Python 3.4, Matplotlib 1.3.1, Ubuntu
> Linux 14.04 64-bit.  My system is selecting the TKAgg backend.
>
> Any advice is appreciated, thanks!
>
>
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming. The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
>

The following minimal code example illustrates a problem I'm having.

    import matplotlib.pyplot as plt
    from mpl_toolkits.mplot3d import Axes3D

    # succeeds
    fig = plt.figure()
    ax = fig.add_subplot(111, projection='3d')
    plt.show()

    # succeeds
    fig = plt.figure()
    ax = fig.add_subplot(111)
    ax.set_xscale("log")
    plt.show()

    # fails
    fig = plt.figure()
    ax = fig.add_subplot(111, projection='3d')
    ax.set_xscale("log")
    plt.show()


Here is the traceback:

    Exception in Tkinter callback
    Traceback (most recent call last):
       File "/usr/lib/python3.4/tkinter/__init__.py", line 1490, in __call__
         return self.func(*args)
       File
    "/usr/lib/python3/dist-packages/matplotlib/backends/backend_tkagg.py",
    line 276, in resize
         self.show()
       File
    "/usr/lib/python3/dist-packages/matplotlib/backends/backend_tkagg.py",
    line 348, in draw
         FigureCanvasAgg.draw(self)
       File
    "/usr/lib/python3/dist-packages/matplotlib/backends/backend_agg.py",
    line 451, in draw
         self.figure.draw(self.renderer)
       File "/usr/lib/python3/dist-packages/matplotlib/artist.py", line
    56, in draw_wrapper
         draw(artist, renderer, *args, **kwargs)
       File "/usr/lib/python3/dist-packages/matplotlib/figure.py", line
    1035, in draw
         func(*args)
       File
    "/usr/lib/python3/dist-packages/mpl_toolkits/mplot3d/axes3d.py",
    line 267, in draw
         ax.draw(renderer)
       File
    "/usr/lib/python3/dist-packages/mpl_toolkits/mplot3d/axis3d.py",
    line 404, in draw
         tick.draw(renderer)
       File "/usr/lib/python3/dist-packages/matplotlib/artist.py", line
    56, in draw_wrapper
         draw(artist, renderer, *args, **kwargs)
       File "/usr/lib/python3/dist-packages/matplotlib/axis.py", line
    242, in draw
         self.label1.draw(renderer)
       File "/usr/lib/python3/dist-packages/matplotlib/artist.py", line
    56, in draw_wrapper
         draw(artist, renderer, *args, **kwargs)
       File "/usr/lib/python3/dist-packages/matplotlib/text.py", line
    599, in draw
         ismath=ismath, mtext=self)
       File
    "/usr/lib/python3/dist-packages/matplotlib/backends/backend_agg.py",
    line 169, in draw_text
         return self.draw_mathtext(gc, x, y, s, prop, angle)
       File
    "/usr/lib/python3/dist-packages/matplotlib/backends/backend_agg.py",
    line 158, in draw_mathtext
         x = np.round(x + ox + xd)
       File "/usr/lib/python3/dist-packages/numpy/core/fromnumeric.py",
    line 2629, in round_
         return round(decimals, out)
       File "/usr/lib/python3/dist-packages/numpy/ma/core.py", line
    4855, in round
         result._mask = self._mask
    AttributeError: 'numpy.float64' object has no attribute '_mask'

I am hoping to use logarithmic axes in a 3D plot.  I haven't found 
anything in the documentation which says that I am restricted to linear 
scaling.  My system configuration is: Python 3.4, Matplotlib 1.3.1, 
Ubuntu Linux 14.04 64-bit.  My system is selecting the TKAgg backend.

Any advice is appreciated, thanks!

Did you come across
http://matplotlib.org/examples/user_interfaces/embedding_in_qt4.html
?
-Sterling

On Jan 28, 2015, at 11:38AM, mo...@po... wrote:

> On 2015-01-28 16:25 Thomas Caswell <tca...@gm...> wrote:
>> It should be exactly the same as for Qt4, just importing from the Qt5
>> version
> 
> Sorry, but my question was to unspecific.
> 
> I don't know how it worked with Qt4.
> 
> I need to know I can "draw" a plot into a QWidget or something else. I
> couldn't find an example or documentation about it.
> 
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming. The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

At the top of www.matplotlib.org, there is an "examples" link.

Scroll down a bit, there's a section titled "user interface examples"
http://matplotlib.org/examples/user_interfaces/index.html

The QT4 example works on my machine currently, so it's reasonably up to
date.
-p

On Wed, Jan 28, 2015 at 11:38 AM, <mo...@po...> wrote:

> On 2015-01-28 16:25 Thomas Caswell <tca...@gm...> wrote:
> > It should be exactly the same as for Qt4, just importing from the Qt5
> > version
>
> Sorry, but my question was to unspecific.
>
> I don't know how it worked with Qt4.
>
> I need to know I can "draw" a plot into a QWidget or something else. I
> couldn't find an example or documentation about it.
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming. The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

http://matplotlib.org/examples/user_interfaces/embedding_in_qt4_wtoolbar.html
and http://matplotlib.org/examples/user_interfaces/embedding_in_qt4.html
 are a good place to start.

Embedding a mpl figure in a QWidget is how the QtAgg backend family works
and how the interactive windows work.

Tom

On Wed Jan 28 2015 at 2:39:16 PM <mo...@po...> wrote:

> On 2015-01-28 16:25 Thomas Caswell <tca...@gm...> wrote:
> > It should be exactly the same as for Qt4, just importing from the Qt5
> > version
>
> Sorry, but my question was to unspecific.
>
> I don't know how it worked with Qt4.
>
> I need to know I can "draw" a plot into a QWidget or something else. I
> couldn't find an example or documentation about it.
>
> ------------------------------------------------------------
> ------------------
> Dive into the World of Parallel Programming. The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

On 2015-01-28 16:25 Thomas Caswell <tca...@gm...> wrote:
> It should be exactly the same as for Qt4, just importing from the Qt5
> version

Sorry, but my question was to unspecific.

I don't know how it worked with Qt4.

I need to know I can "draw" a plot into a QWidget or something else. I
couldn't find an example or documentation about it.

It should be exactly the same as for Qt4, just importing from the Qt5
version

Tom

On Wed Jan 28 2015 at 11:24:55 AM <mo...@po...> wrote:

> I read the documentation a little bit. But I couldn't find a hint how
> to integrate a plot with matplotlib into a Qt5-Gui.
>
> ------------------------------------------------------------
> ------------------
> Dive into the World of Parallel Programming. The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

I read the documentation a little bit. But I couldn't find a hint how
to integrate a plot with matplotlib into a Qt5-Gui.

Hi,

creating a pickle of the Basmap instance was the correct way. Now it works
great! Thank's all for that quick and helpful support!

Solution as follows:

m = Basemap (...)
pickle.dump(m,open('map.pickle','wb'),-1)  # pickle it


for ... :    # loop to generate x plots
    pickle.load(open('map.pickle','rb'))   # load here the above pickle    
    # do some other stuff

Regards
Sappy85



--
View this message in context: http://matplotlib.1069221.n5.nabble.com/Plotting-maps-with-matplotlib-basemap-very-slow-tp44755p44761.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

On 2015/01/24 6:11 AM, Sappy85 wrote:
> I would like to draw very, very simple maps of only europe in matplotlib /
> basemap, which takes very much time (around 10 seconds!). This is just
> unreal!? Setting of resolution is only "l" (low).
> I need to plot hundreds of those maps every few hours. This would be
> impossible!!? :(
>
> Here is the very simple code:
>
> import matplotlib
> matplotlib.use('Agg')
> import matplotlib.pyplot as plt
> from mpl_toolkits.basemap import Basemap
>
> m = Basemap(projection='stere',lon_0=5,lat_0=90.0,rsphere=6371200.,\
>
> llcrnrlon=-25.0,urcrnrlon=72.0,llcrnrlat=26.0,urcrnrlat=65.0,resolution='l')
>

You can save a lot of time by pickling the Basemap instance at this 
point, saving it to a cache.  Then the next time, just pickle.load it 
from the cache.

Eric

> m.drawcoastlines(linewidth=0.2)
> m.drawcountries(linewidth=0.2)
> plt.savefig('/var/www/map.png')
>
>
> Hope you can help! :(
>
> Regards
>
>
>
>
>
> --
> View this message in context: http://matplotlib.1069221.n5.nabble.com/Plotting-maps-with-matplotlib-basemap-very-slow-tp44755.html
> Sent from the matplotlib - users mailing list archive at Nabble.com.
>
> ------------------------------------------------------------------------------
> New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
> GigeNET is offering a free month of service with a new server in Ashburn.
> Choose from 2 high performing configs, both with 100TB of bandwidth.
> Higher redundancy.Lower latency.Increased capacity.Completely compliant.
> http://p.sf.net/sfu/gigenet
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Hi Ben, 

i run the process as user "root" on a virtual linux server. Is this a
problem?





--
View this message in context: http://matplotlib.1069221.n5.nabble.com/Plotting-maps-with-matplotlib-basemap-very-slow-tp44755p44759.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

10 seconds? That doesn't seem right. Are you running the process as an
apache user or something without a home directory? I suspect that the font
cache is not being created, and so it is having to build the font listing
every time it starts up.

Ben Root
On Jan 24, 2015 1:02 PM, "Sappy85" <rob...@gm...> wrote:

> Hi, yes that helps a lot, reading in the boundaries from a shapefile.
>
> But does anybody know, how avoid the "double inner borders"? Coastlines
> seem
> to have a thinner border. Borders between two countries be drawn twice.
>
> <http://matplotlib.1069221.n5.nabble.com/file/n44757/map.png>
>
> Regards
>
>
>
> --
> View this message in context:
> http://matplotlib.1069221.n5.nabble.com/Plotting-maps-with-matplotlib-basemap-very-slow-tp44755p44757.html
> Sent from the matplotlib - users mailing list archive at Nabble.com.
>
>
> ------------------------------------------------------------------------------
> New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
> GigeNET is offering a free month of service with a new server in Ashburn.
> Choose from 2 high performing configs, both with 100TB of bandwidth.
> Higher redundancy.Lower latency.Increased capacity.Completely compliant.
> http://p.sf.net/sfu/gigenet
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Hi, yes that helps a lot, reading in the boundaries from a shapefile.

But does anybody know, how avoid the "double inner borders"? Coastlines seem
to have a thinner border. Borders between two countries be drawn twice.

<http://matplotlib.1069221.n5.nabble.com/file/n44757/map.png> 

Regards



--
View this message in context: http://matplotlib.1069221.n5.nabble.com/Plotting-maps-with-matplotlib-basemap-very-slow-tp44755p44757.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Hi,

Have you considered reading from a suitably low-resolution shapefile
instead? I suppose overlays or colors change per generation but not
the geographical area.

Cheers,

Christian


On Sat, Jan 24, 2015 at 4:11 PM, Sappy85 <rob...@gm...> wrote:
> I would like to draw very, very simple maps of only europe in matplotlib /
> basemap, which takes very much time (around 10 seconds!). This is just
> unreal!? Setting of resolution is only "l" (low).
> I need to plot hundreds of those maps every few hours. This would be
> impossible!!? :(
>
> Here is the very simple code:
>
> import matplotlib
> matplotlib.use('Agg')
> import matplotlib.pyplot as plt
> from mpl_toolkits.basemap import Basemap
>
> m = Basemap(projection='stere',lon_0=5,lat_0=90.0,rsphere=6371200.,\
>
> llcrnrlon=-25.0,urcrnrlon=72.0,llcrnrlat=26.0,urcrnrlat=65.0,resolution='l')
>
> m.drawcoastlines(linewidth=0.2)
> m.drawcountries(linewidth=0.2)
> plt.savefig('/var/www/map.png')
>
>
> Hope you can help! :(
>
> Regards
>
>
>
>
>
> --
> View this message in context: http://matplotlib.1069221.n5.nabble.com/Plotting-maps-with-matplotlib-basemap-very-slow-tp44755.html
> Sent from the matplotlib - users mailing list archive at Nabble.com.
>
> ------------------------------------------------------------------------------
> New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
> GigeNET is offering a free month of service with a new server in Ashburn.
> Choose from 2 high performing configs, both with 100TB of bandwidth.
> Higher redundancy.Lower latency.Increased capacity.Completely compliant.
> http://p.sf.net/sfu/gigenet
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users

I would like to draw very, very simple maps of only europe in matplotlib /
basemap, which takes very much time (around 10 seconds!). This is just
unreal!? Setting of resolution is only "l" (low).
I need to plot hundreds of those maps every few hours. This would be
impossible!!? :(

Here is the very simple code:

import matplotlib
matplotlib.use('Agg')
import matplotlib.pyplot as plt
from mpl_toolkits.basemap import Basemap

m = Basemap(projection='stere',lon_0=5,lat_0=90.0,rsphere=6371200.,\
       
llcrnrlon=-25.0,urcrnrlon=72.0,llcrnrlat=26.0,urcrnrlat=65.0,resolution='l')

m.drawcoastlines(linewidth=0.2)
m.drawcountries(linewidth=0.2)
plt.savefig('/var/www/map.png')


Hope you can help! :(

Regards





--
View this message in context: http://matplotlib.1069221.n5.nabble.com/Plotting-maps-with-matplotlib-basemap-very-slow-tp44755.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

On Fri, Jan 23, 2015 at 1:42 PM, Thomas Caswell <tca...@gm...> wrote:

> Have you looked at using the mpl tables?
> http://matplotlib.org/examples/pylab_examples/table_demo.html
>

Just pointing out:  the numbers in those tables and the words other than
"Quake" are slightly cut-off at the top in the demo.

Hi Tom,

Thanks for the tip, I'll take a look.
Concerning the TeX behavior, do you know what is going on?

Cheers,
Arnaldo.

Have you looked at using the mpl tables?
http://matplotlib.org/examples/pylab_examples/table_demo.html

Tom
On Fri Jan 23 2015 at 1:15:09 PM Arnaldo Russo <arn...@gm...>
wrote:

> Hi,
>
> I'm dealing with this issue but no clues I have found!
> When I use simple matplotlib engine, my plot does not render the correct
> font (which I have set to Arial) but it renders the \hline and specif
> spaces in the table inserted inside the figure area [Fig 1]
> <http://imgur.com/J1A01h5>.
>
> In the second try, using Xelatex engine (made possible by mpl.use('pgf')
> ), I get correct Arial font but my table inside becomes a mess [Fig 2]
> <http://imgur.com/4TckUyb>.
>
> Does anyone knows what is the difference between the two codes [Code]
> <https://gist.github.com/arnaldorusso/e542b04245560300d4f7>, and how
> should I get a figure with Arial font and a table inside the figure area?
>
> It's strange when I have to deal with latex codes and matplotlib... In a
> simple plot (whitout LaTeX elements) I'd have set simply: font.family:
> ["sans-serif"] and font.sans-serif: ["Arial"].
>
> Is there a way to include tables inside a figure with another approach
> (e.g. whithout using LaTeX)?
>
> At least but not less important: The figures in their different part of
> the [Code] <https://gist.github.com/arnaldorusso/e542b04245560300d4f7>,
> must be executed in different session. If I run everything in a simple
> console, I really dont know which parameter is maintained, but Arial is not
> used (in the second plot).
>
>
> [Fig 1] http://imgur.com/J1A01h5
> [Fig 2] http://imgur.com/4TckUyb
> [Code] https://gist.github.com/arnaldorusso/e542b04245560300d4f7
>
> Thank you all!
>
> Cheers,
> Arnaldo.
>
>  ------------------------------------------------------------
> ------------------
> New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
> GigeNET is offering a free month of service with a new server in Ashburn.
> Choose from 2 high performing configs, both with 100TB of bandwidth.
> Higher redundancy.Lower latency.Increased capacity.Completely compliant.
> http://p.sf.net/sfu/gigenet_______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>

Hi,

I'm dealing with this issue but no clues I have found!
When I use simple matplotlib engine, my plot does not render the correct
font (which I have set to Arial) but it renders the \hline and specif
spaces in the table inserted inside the figure area [Fig 1]
<http://imgur.com/J1A01h5>.

In the second try, using Xelatex engine (made possible by mpl.use('pgf') ),
I get correct Arial font but my table inside becomes a mess [Fig 2]
<http://imgur.com/4TckUyb>.

Does anyone knows what is the difference between the two codes [Code]
<https://gist.github.com/arnaldorusso/e542b04245560300d4f7>, and how should
I get a figure with Arial font and a table inside the figure area?

It's strange when I have to deal with latex codes and matplotlib... In a
simple plot (whitout LaTeX elements) I'd have set simply: font.family:
["sans-serif"] and font.sans-serif: ["Arial"].

Is there a way to include tables inside a figure with another approach
(e.g. whithout using LaTeX)?

At least but not less important: The figures in their different part of the
[Code] <https://gist.github.com/arnaldorusso/e542b04245560300d4f7>, must be
executed in different session. If I run everything in a simple console, I
really dont know which parameter is maintained, but Arial is not used (in
the second plot).


[Fig 1] http://imgur.com/J1A01h5
[Fig 2] http://imgur.com/4TckUyb
[Code] https://gist.github.com/arnaldorusso/e542b04245560300d4f7

Thank you all!

Cheers,
Arnaldo.

+1 - that's really great news!

-Shawn

On Wed, Jan 21, 2015 at 2:36 PM, Nicolas P. Rougier
<Nic...@in...> wrote:
>
> +1. Great news.
>
> Nicolas.
>
>> On 21 Jan 2015, at 20:22, Chris Barker <chr...@no...> wrote:
>>
>> +1 -- sounds great!
>>
>>
>>
>> On Tue, Jan 20, 2015 at 7:48 AM, Michael Droettboom <md...@st...> wrote:
>> Matplotlib
>>           is a widely used, well regarded, and powerful visualization
>>           library that has dominated the Python visualization stack for
>>           over a decade. However, to maintain that position, matplotlib
>>           must continue to evolve. Complementary or alternative
>>           libraries are appearing at an increasing rate, including
>>           browser-based plotting and GPU acceleration. To maintain its
>>           leadership position for the next decade, Matplotlib must
>>           interface with these alternatives while simultaneously
>>           expanding its capabilities and becoming easier to use and
>>           learn.
>>
>>
>> Matplotlib’s
>>           large existing user base (greater than 50,000) means that new
>>           developments need to be carefully balanced with maintaining
>>           existing interfaces.  With the large user and code base comes
>>           a significant maintenance and user-support burden.  These
>>           responsibilities currently account for a majority of the
>>           core-developer time spent on matplotlib and has resulted in
>>           both the code base and community being in a healthier state
>>           than ever before. Even 6 years ago there was no automated
>>           testing to speak of and the number of contributors continues
>>           to soar on github. However, this effort is, for the most part,
>>           done on a volunteer basis in the nights and weekends of the
>>           core developers.  To go beyond this maintenance level—to make
>>           step-change improvements for the benefit of matplotlib’s
>>           users—will require funding for full-time developers. Inspired
>>           and encouraged by the example of IPython, we would like to
>>           begin the process of fundraising.
>>
>>
>> Managing
>>           funding on the needed scale is a complex and time-consuming
>>           process.  Thankfully, NumFOCUS, a 501(c)3 charity organisation
>>           co-founded by John Hunter, offers a fiscal sponsorship
>>           agreement to minimize the administrative and legal burden on
>>           open source projects. We would like to enlist NumFOCUS as our
>>           agents in all legal and financial matters, including banking,
>>           accepting donations as a non-profit, payroll, and access to
>>           legal counsel.  As part of the agreement, NumFOCUS would
>>           charge a percentage of all funds raised to cover their costs.
>>            The full text of the agreement is attached.
>>
>>
>> To
>>           comply with the legal and accounting requirements of a
>>           non-profit, matplotlib needs to form an administrative body to
>>           interact with NumFOCUS and direct the disbursement of any
>>           funds.  The proposed initial members of the body, are myself
>>           (Mike Droettboom), Eric Firing, Phil Elson, and Thomas
>>           Caswell, with Thomas acting as the point of contact with
>>           NumFOCUS.
>>
>>
>> In
>>           practice, signing an FSA will have very little impact on the
>>           matplotlib project itself - it will still be BSD-licensed and
>>           community-driven as it has always been, and the only
>>           motivation for doing this is to give us an opportunity to
>>           apply for funding to do more work on matplotlib. We'd like to
>>           canvas the community's opinion on the matter, but to put a
>>           concrete timeline on the discussion, we would like to propose
>>           signing an FSA with NumFOCUS in 3 weeks (Feb 10th 2015) unless
>>           there is a major community discomfort with us doing so.
>>
>>
>>
>>
>>
>>
>>           Cheers,
>>
>>
>>           Michael Droettboom
>>
>>
>>
>>
>>
>> --
>> Michael Droettboom
>> Science Software Branch
>> Space Telescope Science Institute
>>
>>
>> http://www.droettboom.com
>>
>> ------------------------------------------------------------------------------
>> New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
>> GigeNET is offering a free month of service with a new server in Ashburn.
>> Choose from 2 high performing configs, both with 100TB of bandwidth.
>> Higher redundancy.Lower latency.Increased capacity.Completely compliant.
>> http://p.sf.net/sfu/gigenet
>> _______________________________________________
>> Matplotlib-devel mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>>
>>
>>
>>
>> --
>>
>> Christopher Barker, Ph.D.
>> Oceanographer
>>
>> Emergency Response Division
>> NOAA/NOS/OR&R            (206) 526-6959   voice
>> 7600 Sand Point Way NE   (206) 526-6329   fax
>> Seattle, WA  98115       (206) 526-6317   main reception
>>
>> Chr...@no...
>> ------------------------------------------------------------------------------
>> New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
>> GigeNET is offering a free month of service with a new server in Ashburn.
>> Choose from 2 high performing configs, both with 100TB of bandwidth.
>> Higher redundancy.Lower latency.Increased capacity.Completely compliant.
>> http://p.sf.net/sfu/gigenet_______________________________________________
>> Matplotlib-devel mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>
>
> ------------------------------------------------------------------------------
> New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
> GigeNET is offering a free month of service with a new server in Ashburn.
> Choose from 2 high performing configs, both with 100TB of bandwidth.
> Higher redundancy.Lower latency.Increased capacity.Completely compliant.
> http://p.sf.net/sfu/gigenet
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users



-- 
Yuxiang "Shawn" Wang
Gerling Research Lab
University of Virginia
yw...@vi...
+1 (434) 284-0836
https://sites.google.com/a/virginia.edu/yw5aj/

+1. Great news.

Nicolas.

> On 21 Jan 2015, at 20:22, Chris Barker <chr...@no...> wrote:
> 
> +1 -- sounds great!
> 
> 
> 
> On Tue, Jan 20, 2015 at 7:48 AM, Michael Droettboom <md...@st...> wrote:
> Matplotlib
>           is a widely used, well regarded, and powerful visualization
>           library that has dominated the Python visualization stack for
>           over a decade. However, to maintain that position, matplotlib
>           must continue to evolve. Complementary or alternative
>           libraries are appearing at an increasing rate, including
>           browser-based plotting and GPU acceleration. To maintain its
>           leadership position for the next decade, Matplotlib must
>           interface with these alternatives while simultaneously
>           expanding its capabilities and becoming easier to use and
>           learn.
> 
> 
> Matplotlib’s
>           large existing user base (greater than 50,000) means that new
>           developments need to be carefully balanced with maintaining
>           existing interfaces.  With the large user and code base comes
>           a significant maintenance and user-support burden.  These
>           responsibilities currently account for a majority of the
>           core-developer time spent on matplotlib and has resulted in
>           both the code base and community being in a healthier state
>           than ever before. Even 6 years ago there was no automated
>           testing to speak of and the number of contributors continues
>           to soar on github. However, this effort is, for the most part,
>           done on a volunteer basis in the nights and weekends of the
>           core developers.  To go beyond this maintenance level—to make
>           step-change improvements for the benefit of matplotlib’s
>           users—will require funding for full-time developers. Inspired
>           and encouraged by the example of IPython, we would like to
>           begin the process of fundraising.
> 
> 
> Managing
>           funding on the needed scale is a complex and time-consuming
>           process.  Thankfully, NumFOCUS, a 501(c)3 charity organisation
>           co-founded by John Hunter, offers a fiscal sponsorship
>           agreement to minimize the administrative and legal burden on
>           open source projects. We would like to enlist NumFOCUS as our
>           agents in all legal and financial matters, including banking,
>           accepting donations as a non-profit, payroll, and access to
>           legal counsel.  As part of the agreement, NumFOCUS would
>           charge a percentage of all funds raised to cover their costs.
>            The full text of the agreement is attached.
> 
> 
> To
>           comply with the legal and accounting requirements of a
>           non-profit, matplotlib needs to form an administrative body to
>           interact with NumFOCUS and direct the disbursement of any
>           funds.  The proposed initial members of the body, are myself
>           (Mike Droettboom), Eric Firing, Phil Elson, and Thomas
>           Caswell, with Thomas acting as the point of contact with
>           NumFOCUS.
> 
> 
> In
>           practice, signing an FSA will have very little impact on the
>           matplotlib project itself - it will still be BSD-licensed and
>           community-driven as it has always been, and the only
>           motivation for doing this is to give us an opportunity to
>           apply for funding to do more work on matplotlib. We'd like to
>           canvas the community's opinion on the matter, but to put a
>           concrete timeline on the discussion, we would like to propose
>           signing an FSA with NumFOCUS in 3 weeks (Feb 10th 2015) unless
>           there is a major community discomfort with us doing so. 
> 
> 
>         
> 
> 
> 
>           Cheers,
> 
> 
>           Michael Droettboom
> 
> 
>         
> 
> 
> -- 
> Michael Droettboom
> Science Software Branch
> Space Telescope Science Institute
> 
> 
> http://www.droettboom.com
> 
> ------------------------------------------------------------------------------
> New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
> GigeNET is offering a free month of service with a new server in Ashburn.
> Choose from 2 high performing configs, both with 100TB of bandwidth.
> Higher redundancy.Lower latency.Increased capacity.Completely compliant.
> http://p.sf.net/sfu/gigenet
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
> 
> 
> 
> 
> -- 
> 
> Christopher Barker, Ph.D.
> Oceanographer
> 
> Emergency Response Division
> NOAA/NOS/OR&R            (206) 526-6959   voice
> 7600 Sand Point Way NE   (206) 526-6329   fax
> Seattle, WA  98115       (206) 526-6317   main reception
> 
> Chr...@no...
> ------------------------------------------------------------------------------
> New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
> GigeNET is offering a free month of service with a new server in Ashburn.
> Choose from 2 high performing configs, both with 100TB of bandwidth.
> Higher redundancy.Lower latency.Increased capacity.Completely compliant.
> http://p.sf.net/sfu/gigenet_______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel