Region within contour --> 2D array

_Matthew_Czesarski · August 28, 2014, 1:02pm

Hi Matplotlib Users!

I have some 2-d arrays, which i am displaying with implot, and deriving contours for with contour. Easy - I’m just pulling them out of collections[0].get_paths() .

However what’s not easy is that I would like to recover a 1-0 or True-False array of the array values (pixels) that fall within the contours. Some line crossing algorithm/floodfill could do it, but I guess that matplotlib’s fill() or contourf() must do this under the hood anyway. I’ve looked into the output both functions, but I don’t see anything obvious…

Does anybody know if there’s an a way to pull out a such an array from matplotlib? Any pointers are appreciated!

Cheers,
Matt

Shantha_Kumara · August 28, 2014, 1:24pm

Hi All,

Thank you so much for your help, It really worked for me.

I need one more favor,

I have ploted the graph with 2 Y-axes

Here is the code

   lns1 = ax1.plot(x1, y1, 'r-o',label=LY1,markersize=4)
        ax1 = self.set_ylim(ax1,y1,label=LY1)
        lns2 = ax2.plot(x2, y2, 'b-o',label=LY2,markersize=4)

I want to set the grid to the second y-axes in the same graph.

Please help on the same.

See here the difference

(Embedded image moved to file: pic19302.gif)

SHANTHA KUMARA REVANASIDDAPPA
Python Developper
Operations & Production Control, A1 Telekom
IBM
M +43-6642196132
@ Shantha.Kumara.Kadur.Revanasiddappa@...4563...

Benjamin_Root · August 28, 2014, 1:31pm

That stuff is done in the deep underbelly of matplotlib and isn’t exposed to the user. It is done as part of the rendering process in AGG or whichever other backend is performing the render. I have done something very similar to what you are asking for my job, and while I can’t share the code, I can point out that GDAL has a very fast polygon rasterizer. I can also point you to this link: http://stackoverflow.com/questions/2220749/rasterizing-a-gdal-layer

I will also say that there are some subtle errors in that code, but it should get you to where you need to go.

Cheers!
Ben Root

···

On Thu, Aug 28, 2014 at 9:02 AM, Matthew Czesarski <matthew.czesarski@…287…> wrote:

Hi Matplotlib Users!

I have some 2-d arrays, which i am displaying with implot, and deriving contours for with contour. Easy - I’m just pulling them out of collections[0].get_paths() .

However what’s not easy is that I would like to recover a 1-0 or True-False array of the array values (pixels) that fall within the contours. Some line crossing algorithm/floodfill could do it, but I guess that matplotlib’s fill() or contourf() must do this under the hood anyway. I’ve looked into the output both functions, but I don’t see anything obvious…

Does anybody know if there’s an a way to pull out a such an array from matplotlib? Any pointers are appreciated!

Cheers,
Matt

Slashdot TV.

Video for Nerds. Stuff that matters.

http://tv.slashdot.org/

Matplotlib-users mailing list

Matplotlib-users@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Benjamin_Root · August 28, 2014, 1:33pm

You are asking for the twinx() feature: http://matplotlib.org/examples/api/two_scales.html

···

On Thu, Aug 28, 2014 at 9:24 AM, Shantha Kumara <shkumara@…4562…> wrote:

Hi All,

Thank you so much for your help, It really worked for me.

I need one more favor,

I have ploted the graph with 2 Y-axes

Here is the code
     lns1 = ax1.plot(x1, y1, 'r-o',label=LY1,markersize=4)

    ax1 = self.set_ylim(ax1,y1,label=LY1)

    lns2 = ax2.plot(x2, y2, 'b-o',label=LY2,markersize=4)
I want to set the grid to the second y-axes in the same graph.

Please help on the same.

See here the difference

(Embedded image moved to file: pic19302.gif)

SHANTHA KUMARA REVANASIDDAPPA

Python Developper

Operations & Production Control, A1 Telekom

IBM

M +43-6642196132

@ Shantha.Kumara.Kadur.Revanasiddappa@…4563…

Slashdot TV.

Video for Nerds. Stuff that matters.

http://tv.slashdot.org/

Matplotlib-users mailing list

Matplotlib-users@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Eric_Firing1 · August 29, 2014, 1:23am

Hi Matplotlib Users!

I have some 2-d arrays, which i am displaying with implot, and deriving
contours for with contour. Easy - I'm just pulling them out of
collections[0].get_paths() .

However what's not easy is that I would like to recover a 1-0 or
True-False array of the array values (pixels) that fall within the
contours. Some line crossing algorithm/floodfill could do it, but I
guess that matplotlib's fill() or contourf() must do this under the hood
anyway. I've looked into the output both functions, but I don't see
anything obvious..

Does anybody know if there's an a way to pull out a such an array from
matplotlib? Any pointers are appreciated!

Make an array of (x, y) pairs from the X and Y you use in your call to contour, and then feed that array to the contains_points() method of your contour Path. This will give you the desired Boolean array for any given Path; depending on what you want, you might need to combine arrays for more than one Path.

To get closed paths, I think you will want to use contourf, not contour.

Eric

···

On 2014/08/28, 3:02 AM, Matthew Czesarski wrote:

Cheers,
Matt

------------------------------------------------------------------------------
Slashdot TV.
Video for Nerds. Stuff that matters.
http://tv.slashdot.org/

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
matplotlib-users List Signup and Options

Joe_Kington3 · August 29, 2014, 2:18am

Why not just use boolean indexing?

E.g. to find the region that falls between 5 and 10, do “(z >=5) & (z <= 10)”:

In [1]: import numpy as np

In [2]: x, y = np.mgrid[-10:10, -10:10]

In [3]: z = np.hypot(x, y)

In [4]: result = (z >= 5) & (z <= 10)

In [5]: result.astype(int)
Out[5]:
array([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0],

   [0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0],
   [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0],
   [0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0],
   [0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0],

   [0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0],
   [0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1],
   [0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1],
   [0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1],

   [0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1],
   [1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1],
   [0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1],
   [0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1],

   [0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1],
   [0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1],
   [0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0],
   [0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0],

   [0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0],
   [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0],
   [0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0]])

Cheers,

-Joe

···

On Thu, Aug 28, 2014 at 8:23 PM, Eric Firing <efiring@…202…> wrote:

On 2014/08/28, 3:02 AM, Matthew Czesarski wrote:

Hi Matplotlib Users!

I have some 2-d arrays, which i am displaying with implot, and deriving

contours for with contour. Easy - I’m just pulling them out of

collections[0].get_paths() .

However what’s not easy is that I would like to recover a 1-0 or

True-False array of the array values (pixels) that fall within the

contours. Some line crossing algorithm/floodfill could do it, but I

guess that matplotlib’s fill() or contourf() must do this under the hood

anyway. I’ve looked into the output both functions, but I don’t see

anything obvious…

Does anybody know if there’s an a way to pull out a such an array from

matplotlib? Any pointers are appreciated!

Make an array of (x, y) pairs from the X and Y you use in your call to

contour, and then feed that array to the contains_points() method of

your contour Path. This will give you the desired Boolean array for any

given Path; depending on what you want, you might need to combine arrays

for more than one Path.

To get closed paths, I think you will want to use contourf, not contour.

Eric

Cheers,

Matt

Slashdot TV.

Video for Nerds. Stuff that matters.

http://tv.slashdot.org/

Matplotlib-users mailing list

Matplotlib-users@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-users

Slashdot TV.

Video for Nerds. Stuff that matters.

http://tv.slashdot.org/

Matplotlib-users mailing list

Matplotlib-users@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-users

_Sterling_P_Smith · August 29, 2014, 5:18am

Joe and list,

This is off topic, but can you point me to good documentation on the use of '&' as opposed to numpy.logical_and ?

Thanks,
Sterling

···

On Aug 28, 2014, at 7:18PM, Joe Kington wrote:

Why not just use boolean indexing?

E.g. to find the region that falls between 5 and 10, do "(z >=5) & (z <= 10)":

In [1]: import numpy as np

In [2]: x, y = np.mgrid[-10:10, -10:10]

In [3]: z = np.hypot(x, y)

In [4]: result = (z >= 5) & (z <= 10)

In [5]: result.astype(int)
Out[5]:
array([[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0],
       [0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0],
       [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0],
       [0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0],
       [0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0],
       [0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0],
       [0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1],
       [0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1],
       [0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1],
       [0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1],
       [1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1],
       [0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1],
       [0, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1],
       [0, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1],
       [0, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1],
       [0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0],
       [0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0],
       [0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0],
       [0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0],
       [0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0]])

Cheers,
-Joe

On Thu, Aug 28, 2014 at 8:23 PM, Eric Firing <efiring@...202...> wrote:
On 2014/08/28, 3:02 AM, Matthew Czesarski wrote:
> Hi Matplotlib Users!
>
>
>
> I have some 2-d arrays, which i am displaying with implot, and deriving
> contours for with contour. Easy - I'm just pulling them out of
> collections[0].get_paths() .
>
> However what's not easy is that I would like to recover a 1-0 or
> True-False array of the array values (pixels) that fall within the
> contours. Some line crossing algorithm/floodfill could do it, but I
> guess that matplotlib's fill() or contourf() must do this under the hood
> anyway. I've looked into the output both functions, but I don't see
> anything obvious..
>
> Does anybody know if there's an a way to pull out a such an array from
> matplotlib? Any pointers are appreciated!

Make an array of (x, y) pairs from the X and Y you use in your call to
contour, and then feed that array to the contains_points() method of
your contour Path. This will give you the desired Boolean array for any
given Path; depending on what you want, you might need to combine arrays
for more than one Path.

To get closed paths, I think you will want to use contourf, not contour.

Eric

>
> Cheers,
> Matt
>
>
> ------------------------------------------------------------------------------
> Slashdot TV.
> Video for Nerds. Stuff that matters.
> http://tv.slashdot.org/
>
>
>
> _______________________________________________
> Matplotlib-users mailing list
> Matplotlib-users@lists.sourceforge.net
> matplotlib-users List Signup and Options
>

------------------------------------------------------------------------------
Slashdot TV.
Video for Nerds. Stuff that matters.
http://tv.slashdot.org/
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
matplotlib-users List Signup and Options

------------------------------------------------------------------------------
Slashdot TV.
Video for Nerds. Stuff that matters.
Slashdot: News for nerds, stuff that matters
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
matplotlib-users List Signup and Options

Benjamin_Root · August 29, 2014, 1:28pm

slaps forehead…

Joe, you just won the “duh!” moment of the month award!

Cheers!

Ben Root

···

On Thu, Aug 28, 2014 at 10:18 PM, Joe Kington <joferkington@…287…> wrote: