Unexpected behaviour with gridplot and layout

Community Support
1

Hello,

I am trying to understand gridplots and layouts. My starting point is code from the webpage:

‘’’
Code from Creating layouts — Bokeh 2.4.2 Documentation
‘’’

from bokeh.io import output_server, show
from bokeh.layouts import gridplot, layout
from bokeh.palettes import Viridis3
from bokeh.plotting import figure
import subprocess
import time

if name == ‘main’:
args = [‘python’, ‘-m’, ‘bokeh’, ‘serve’]
p = subprocess.Popen(args)
time.sleep(1)
output_server(session_id=‘test’, autopush = True)

x = list(range(11))
y0 = x
y1 = [10 - i for i in x]
y2 = [abs(i - 5) for i in x]

# create three plots
p1 = figure(width=250, plot_height=250, title=None)
p1.circle(x, y0, size=10, color=Viridis3[0])
p2 = figure(width=250, height=250, title=None)
p2.triangle(x, y1, size=10, color=Viridis3[1])
p3 = figure(width=250, height=250, title=None)
p3.square(x, y2, size=10, color=Viridis3[2])

# make a grid
grid = gridplot([[p1, p2], [None, p3]])

testlayout = layout([[p1, p2, p3]], sizing_mode=‘stretch_both’)

# show the results
show(grid)

show(testlayout)

``

This works as shown on the webpage:

When I add an (unused) layout,

# make a grid
grid = gridplot([[p1, p2], [None, p3]])
testlayout = layout([[p1, p2, p3]], sizing_mode='stretch_both')

# show the results
show(grid)

show(testlayout)

``

``

``

the output changes to (screenshot):

When I use the save-button I get three images, although only one is visible in the webpage.

Console output:

2016-10-26 21:02:09,054 Starting Bokeh server version 0.12.2

2016-10-26 21:02:09,061 Starting Bokeh server on port 5006 with applications at paths [’/’]

2016-10-26 21:02:09,061 Starting Bokeh server with process id: 736

2016-10-26 21:02:09,109 WebSocket connection opened

2016-10-26 21:02:09,110 ServerConnection created

2016-10-26 21:02:09,260 WebSocket connection closed: code=1000, reason=‘closed’

INFO:bokeh.client._connection:Connection closed by server

2016-10-26 21:02:09,520 200 GET /?bokeh-session-id=test (::1) 13.01ms

2016-10-26 21:02:10,058 WebSocket connection opened

2016-10-26 21:02:10,059 ServerConnection created

2016-10-26 21:02:10,648 error handling message Message ‘PATCH-DOC’ (revision 1): DeserializationError('Int exp

ected Integral, got 1.4210854715202004e-14’,)

Last version, only layout

make a grid

grid = gridplot([[p1, p2], [None, p3]])

testlayout = layout([[p1, p2, p3]], sizing_mode=‘stretch_both’)

show the results

show(grid)

show(testlayout

``

is working as expected:

Is there a specific reason, why gridplot can not be mixed with an layout? Even if the layout variable is not used at all?

Thanks

Daniel

2

Hi,

···

On Wed, Oct 26, 2016 at 9:21 PM, ‘Daniel Krause’ via Bokeh Discussion - Public [email protected] wrote:

Hello,

I am trying to understand gridplots and layouts. My starting point is code from the webpage:

‘’’
Code from http://bokeh.pydata.org/en/latest/docs/user_guide/layout.html
‘’’

from bokeh.io import output_server, show
from bokeh.layouts import gridplot, layout
from bokeh.palettes import Viridis3
from bokeh.plotting import figure
import subprocess
import time

if name == ‘main’:
args = [‘python’, ‘-m’, ‘bokeh’, ‘serve’]
p = subprocess.Popen(args)
time.sleep(1)
output_server(session_id=‘test’, autopush = True)

x = list(range(11))
y0 = x
y1 = [10 - i for i in x]
y2 = [abs(i - 5) for i in x]

# create three plots
p1 = figure(width=250, plot_height=250, title=None)
p1.circle(x, y0, size=10, color=Viridis3[0])
p2 = figure(width=250, height=250, title=None)
p2.triangle(x, y1, size=10, color=Viridis3[1])
p3 = figure(width=250, height=250, title=None)
p3.square(x, y2, size=10, color=Viridis3[2])

# make a grid
grid = gridplot([[p1, p2], [None, p3]])

testlayout = layout([[p1, p2, p3]], sizing_mode=‘stretch_both’)

# show the results
show(grid)

show(testlayout)

``

This works as shown on the webpage:

When I add an (unused) layout,

# make a grid
grid = gridplot([[p1, p2], [None, p3]])
testlayout = layout([[p1, p2, p3]], sizing_mode='stretch_both')

# show the results
show(grid)

show(testlayout)

``

``

``

the output changes to (screenshot):

When I use the save-button I get three images, although only one is visible in the webpage.

Console output:

2016-10-26 21:02:09,054 Starting Bokeh server version 0.12.2

2016-10-26 21:02:09,061 Starting Bokeh server on port 5006 with applications at paths [‘/’]

2016-10-26 21:02:09,061 Starting Bokeh server with process id: 736

2016-10-26 21:02:09,109 WebSocket connection opened

2016-10-26 21:02:09,110 ServerConnection created

2016-10-26 21:02:09,260 WebSocket connection closed: code=1000, reason=‘closed’

INFO:bokeh.client._connection:Connection closed by server

2016-10-26 21:02:09,520 200 GET /?bokeh-session-id=test (::1) 13.01ms

2016-10-26 21:02:10,058 WebSocket connection opened

2016-10-26 21:02:10,059 ServerConnection created

2016-10-26 21:02:10,648 error handling message Message ‘PATCH-DOC’ (revision 1): DeserializationError('Int exp

ected Integral, got 1.4210854715202004e-14’,)

Given the above error message, I think that the behavior you experience isn’t a problem with layout, but some internal plumbing, most likely due to using output_server(), which is broken and has been deprecated. If you just want to play with layouts, I suggest using the most recent version of this example, which uses output_file() instead.

Mateusz

Last version, only layout

make a grid

grid = gridplot([[p1, p2], [None, p3]])

testlayout = layout([[p1, p2, p3]], sizing_mode=‘stretch_both’)

show the results

show(grid)

show(testlayout

``

is working as expected:

Is there a specific reason, why gridplot can not be mixed with an layout? Even if the layout variable is not used at all?

Thanks

Daniel

–

You received this message because you are subscribed to the Google Groups “Bokeh Discussion - Public” group.

To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].

To post to this group, send email to [email protected].

To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/15778262-7821-4d8d-8594-a649ddd18a7b%40continuum.io.

For more options, visit https://groups.google.com/a/continuum.io/d/optout.

3

Yes, I noticed the first error message as well.
In the meantime I changed the imports to

from bokeh.client import push_session
from bokeh.layouts import gridplot, layout
from bokeh.palettes import Viridis3
from bokeh.plotting import figure, curdoc
import subprocess
import time

``

``

and the code with grid and layout as follows

# make a grid
grid = gridplot([[p1, p2], [None, p3]])
testlayout = layout([[p1, p2, p3]], sizing_mode='stretch_both')

# show the results
session = push_session(curdoc())
session.show(grid) # open the document in a browser

show(testlayout)

``

The result did not change, the behaviour stays the same.

But thanks for idea,

Daniel

···

Am Mittwoch, 26. Oktober 2016 22:11:20 UTC+2 schrieb mateusz.paprocki:

Hi,

On Wed, Oct 26, 2016 at 9:21 PM, ‘Daniel Krause’ via Bokeh Discussion - Public [email protected] wrote:

Hello,

I am trying to understand gridplots and layouts. My starting point is code from the webpage:

‘’’
Code from http://bokeh.pydata.org/en/latest/docs/user_guide/layout.html
‘’’

from bokeh.io import output_server, show
from bokeh.layouts import gridplot, layout
from bokeh.palettes import Viridis3
from bokeh.plotting import figure
import subprocess
import time

if name == ‘main’:
args = [‘python’, ‘-m’, ‘bokeh’, ‘serve’]
p = subprocess.Popen(args)
time.sleep(1)
output_server(session_id=‘test’, autopush = True)

x = list(range(11))
y0 = x
y1 = [10 - i for i in x]
y2 = [abs(i - 5) for i in x]

# create three plots
p1 = figure(width=250, plot_height=250, title=None)
p1.circle(x, y0, size=10, color=Viridis3[0])
p2 = figure(width=250, height=250, title=None)
p2.triangle(x, y1, size=10, color=Viridis3[1])
p3 = figure(width=250, height=250, title=None)
p3.square(x, y2, size=10, color=Viridis3[2])

# make a grid
grid = gridplot([[p1, p2], [None, p3]])

testlayout = layout([[p1, p2, p3]], sizing_mode=‘stretch_both’)

# show the results
show(grid)

show(testlayout)

``

This works as shown on the webpage:

When I add an (unused) layout,

# make a grid
grid = gridplot([[p1, p2], [None, p3]])
testlayout = layout([[p1, p2, p3]], sizing_mode='stretch_both')

# show the results
show(grid)

show(testlayout)

``

``

``

the output changes to (screenshot):

When I use the save-button I get three images, although only one is visible in the webpage.

Console output:

2016-10-26 21:02:09,054 Starting Bokeh server version 0.12.2

2016-10-26 21:02:09,061 Starting Bokeh server on port 5006 with applications at paths [‘/’]

2016-10-26 21:02:09,061 Starting Bokeh server with process id: 736

2016-10-26 21:02:09,109 WebSocket connection opened

2016-10-26 21:02:09,110 ServerConnection created

2016-10-26 21:02:09,260 WebSocket connection closed: code=1000, reason=‘closed’

INFO:bokeh.client._connection:Connection closed by server

2016-10-26 21:02:09,520 200 GET /?bokeh-session-id=test (::1) 13.01ms

2016-10-26 21:02:10,058 WebSocket connection opened

2016-10-26 21:02:10,059 ServerConnection created

2016-10-26 21:02:10,648 error handling message Message ‘PATCH-DOC’ (revision 1): DeserializationError('Int exp

ected Integral, got 1.4210854715202004e-14’,)

Given the above error message, I think that the behavior you experience isn’t a problem with layout, but some internal plumbing, most likely due to using output_server(), which is broken and has been deprecated. If you just want to play with layouts, I suggest using the most recent version of this example, which uses output_file() instead.

Mateusz

Last version, only layout

make a grid

grid = gridplot([[p1, p2], [None, p3]])

testlayout = layout([[p1, p2, p3]], sizing_mode=‘stretch_both’)

show the results

show(grid)

show(testlayout

``

is working as expected:

Is there a specific reason, why gridplot can not be mixed with an layout? Even if the layout variable is not used at all?

Thanks

Daniel

–

You received this message because you are subscribed to the Google Groups “Bokeh Discussion - Public” group.

To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].

To post to this group, send email to [email protected].

To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/15778262-7821-4d8d-8594-a649ddd18a7b%40continuum.io.

For more options, visit https://groups.google.com/a/continuum.io/d/optout.

4

Sorry, my last message was not complete.
I updated to bokeh 0.12.3. In the console output appeared than an additional message stating the used output_server is deprecated. That was the reason why I changed the code.

As for the

error handling message Message ‘PATCH-DOC’ (revision 1): DeserializationError('Int exp

ected Integral, got 1.4210854715202004e-14’,)

``

I have no idea what that could mean.

Daniel

···

Am Mittwoch, 26. Oktober 2016 22:20:15 UTC+2 schrieb Daniel Krause:

Yes, I noticed the first error message as well.
In the meantime I changed the imports to

from bokeh.client import push_session
from bokeh.layouts import gridplot, layout
from bokeh.palettes import Viridis3
from bokeh.plotting import figure, curdoc
import subprocess
import time

``

``

and the code with grid and layout as follows

# make a grid
grid = gridplot([[p1, p2], [None, p3]])
testlayout = layout([[p1, p2, p3]], sizing_mode='stretch_both')

# show the results
session = push_session(curdoc())
session.show(grid) # open the document in a browser

show(testlayout)

``

The result did not change, the behaviour stays the same.

But thanks for idea,

Daniel

Am Mittwoch, 26. Oktober 2016 22:11:20 UTC+2 schrieb mateusz.paprocki:

Hi,

On Wed, Oct 26, 2016 at 9:21 PM, ‘Daniel Krause’ via Bokeh Discussion - Public [email protected] wrote:

Hello,

I am trying to understand gridplots and layouts. My starting point is code from the webpage:

‘’’
Code from http://bokeh.pydata.org/en/latest/docs/user_guide/layout.html
‘’’

from bokeh.io import output_server, show
from bokeh.layouts import gridplot, layout
from bokeh.palettes import Viridis3
from bokeh.plotting import figure
import subprocess
import time

if name == ‘main’:
args = [‘python’, ‘-m’, ‘bokeh’, ‘serve’]
p = subprocess.Popen(args)
time.sleep(1)
output_server(session_id=‘test’, autopush = True)

x = list(range(11))
y0 = x
y1 = [10 - i for i in x]
y2 = [abs(i - 5) for i in x]

# create three plots
p1 = figure(width=250, plot_height=250, title=None)
p1.circle(x, y0, size=10, color=Viridis3[0])
p2 = figure(width=250, height=250, title=None)
p2.triangle(x, y1, size=10, color=Viridis3[1])
p3 = figure(width=250, height=250, title=None)
p3.square(x, y2, size=10, color=Viridis3[2])

# make a grid
grid = gridplot([[p1, p2], [None, p3]])

testlayout = layout([[p1, p2, p3]], sizing_mode=‘stretch_both’)

# show the results
show(grid)

show(testlayout)

``

This works as shown on the webpage:

When I add an (unused) layout,

# make a grid
grid = gridplot([[p1, p2], [None, p3]])
testlayout = layout([[p1, p2, p3]], sizing_mode='stretch_both')

# show the results
show(grid)

show(testlayout)

``

``

``

the output changes to (screenshot):

When I use the save-button I get three images, although only one is visible in the webpage.

Console output:

2016-10-26 21:02:09,054 Starting Bokeh server version 0.12.2

2016-10-26 21:02:09,061 Starting Bokeh server on port 5006 with applications at paths [‘/’]

2016-10-26 21:02:09,061 Starting Bokeh server with process id: 736

2016-10-26 21:02:09,109 WebSocket connection opened

2016-10-26 21:02:09,110 ServerConnection created

2016-10-26 21:02:09,260 WebSocket connection closed: code=1000, reason=‘closed’

INFO:bokeh.client._connection:Connection closed by server

2016-10-26 21:02:09,520 200 GET /?bokeh-session-id=test (::1) 13.01ms

2016-10-26 21:02:10,058 WebSocket connection opened

2016-10-26 21:02:10,059 ServerConnection created

2016-10-26 21:02:10,648 error handling message Message ‘PATCH-DOC’ (revision 1): DeserializationError('Int exp

ected Integral, got 1.4210854715202004e-14’,)

Given the above error message, I think that the behavior you experience isn’t a problem with layout, but some internal plumbing, most likely due to using output_server(), which is broken and has been deprecated. If you just want to play with layouts, I suggest using the most recent version of this example, which uses output_file() instead.

Mateusz

Last version, only layout

make a grid

grid = gridplot([[p1, p2], [None, p3]])

testlayout = layout([[p1, p2, p3]], sizing_mode=‘stretch_both’)

show the results

show(grid)

show(testlayout

``

is working as expected:

Is there a specific reason, why gridplot can not be mixed with an layout? Even if the layout variable is not used at all?

Thanks

Daniel

–

You received this message because you are subscribed to the Google Groups “Bokeh Discussion - Public” group.

To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].

To post to this group, send email to [email protected].

To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/15778262-7821-4d8d-8594-a649ddd18a7b%40continuum.io.

For more options, visit https://groups.google.com/a/continuum.io/d/optout.

5

Hi,

I lost the thread of this, but here's a complete code sample that works in a couple of configurations with some comments made inline:

···

-----

from bokeh.client import push_session
from bokeh.layouts import gridplot, layout
from bokeh.palettes import Viridis3
from bokeh.plotting import figure, curdoc
import subprocess
import time

if __name__ == '__main__':
    args = ['python', '-m', 'bokeh', 'serve']
    p = subprocess.Popen(args)

    # had to increase, bokeh serve takes more time to start
    # probably better to coordinate externally, or find a way to wait
    # until the server start can be confirmed as completed
    time.sleep(3)

    x = list(range(11))
    y0 = x
    y1 = [10 - i for i in x]
    y2 = [abs(i - 5) for i in x]

    # create three plots
    p1 = figure(width=250, plot_height=250, title=None)
    p1.circle(x, y0, size=10, color=Viridis3[0])
    p2 = figure(width=250, height=250, title=None)
    p2.triangle(x, y1, size=10, color=Viridis3[1])
    p3 = figure(width=250, height=250, title=None)
    p3.square(x, y2, size=10, color=Viridis3[2])

     # either one or the other of these works -- but not BOTH at once
     # cannot re-use same plot twice in a Bokeh app
    #layout = gridplot([[p1, p2], [None, p3]])
    layout = layout([[p1, p2, p3]], sizing_mode='stretch_both')

    # show the results
    session = push_session(curdoc())
    session.show(layout) # open the document in a browser
    session.loop_until_closed()

6

Thanks again Bryan!

···

Am Mittwoch, 26. Oktober 2016 21:21:04 UTC+2 schrieb Daniel Krause:

Hello,

I am trying to understand gridplots and layouts. My starting point is code from the webpage:

‘’’
Code from http://bokeh.pydata.org/en/latest/docs/user_guide/layout.html
‘’’

from bokeh.io import output_server, show
from bokeh.layouts import gridplot, layout
from bokeh.palettes import Viridis3
from bokeh.plotting import figure
import subprocess
import time

if name == ‘main’:
args = [‘python’, ‘-m’, ‘bokeh’, ‘serve’]
p = subprocess.Popen(args)
time.sleep(1)
output_server(session_id=‘test’, autopush = True)

x = list(range(11))
y0 = x
y1 = [10 - i for i in x]
y2 = [abs(i - 5) for i in x]

# create three plots
p1 = figure(width=250, plot_height=250, title=None)
p1.circle(x, y0, size=10, color=Viridis3[0])
p2 = figure(width=250, height=250, title=None)
p2.triangle(x, y1, size=10, color=Viridis3[1])
p3 = figure(width=250, height=250, title=None)
p3.square(x, y2, size=10, color=Viridis3[2])

# make a grid
grid = gridplot([[p1, p2], [None, p3]])

testlayout = layout([[p1, p2, p3]], sizing_mode=‘stretch_both’)

# show the results
show(grid)

show(testlayout)

``

This works as shown on the webpage:

When I add an (unused) layout,

# make a grid
grid = gridplot([[p1, p2], [None, p3]])
testlayout = layout([[p1, p2, p3]], sizing_mode='stretch_both')

# show the results
show(grid)

show(testlayout)

``

``

``

the output changes to (screenshot):

When I use the save-button I get three images, although only one is visible in the webpage.

Console output:

2016-10-26 21:02:09,054 Starting Bokeh server version 0.12.2

2016-10-26 21:02:09,061 Starting Bokeh server on port 5006 with applications at paths [‘/’]

2016-10-26 21:02:09,061 Starting Bokeh server with process id: 736

2016-10-26 21:02:09,109 WebSocket connection opened

2016-10-26 21:02:09,110 ServerConnection created

2016-10-26 21:02:09,260 WebSocket connection closed: code=1000, reason=‘closed’

INFO:bokeh.client._connection:Connection closed by server

2016-10-26 21:02:09,520 200 GET /?bokeh-session-id=test (::1) 13.01ms

2016-10-26 21:02:10,058 WebSocket connection opened

2016-10-26 21:02:10,059 ServerConnection created

2016-10-26 21:02:10,648 error handling message Message ‘PATCH-DOC’ (revision 1): DeserializationError('Int exp

ected Integral, got 1.4210854715202004e-14’,)

Last version, only layout

make a grid

grid = gridplot([[p1, p2], [None, p3]])

testlayout = layout([[p1, p2, p3]], sizing_mode=‘stretch_both’)

show the results

show(grid)

show(testlayout

``

is working as expected:

Is there a specific reason, why gridplot can not be mixed with an layout? Even if the layout variable is not used at all?

Thanks

Daniel