I have a layout that returns a variable number of plot depending on the parameter selected in Select Widget. I have observed this intermittent error message in bokeh server and I do not know what it means. I have observed that the error occurred mainly when there are many plots (>10). What can I do to fix this problem? My code is here if necessary.
It looks like you are creating new plots and adding them? There are some know issues with regards to adding *new* objects to documents. The primary initial use-case we sought, and the one that is very well demonstrated is rather to *update* existing objects in the document. That is, change the data in a data source to update a plot, or change the values properties on an existing objects to update titles or colors or transparency, etc. We certainly intend to address these issues, as the limitations of human resources permit. In the mean time the best general guidance I can give is to try an structure your app as much as possible around creating one set of models once, up front, then *updating* them. If you need to hide/show things there are .visible properties in some cases, or alpha values that can be set to zero, etc.
I have a layout that returns a variable number of plot depending on the parameter selected in Select Widget. I have observed this intermittent error message in bokeh server and I do not know what it means. I have observed that the error occurred mainly when there are many plots (>10). What can I do to fix this problem? My code is here if necessary.
.
Thanks for any help.
Rafael Pais
plataform: win-64
conda version 4.2.9
bokeh 0.12.2
An improvement in that direction would be interesting. I need to work with lots of plots(>200) in different categories. In an excel file I have a spreadsheet with the plot configuration (axis label, plot title…) that I use to generate the plots in this way.
def figure_ticker(n):
plts=[]
ps = [figure(plot_width=550, plot_height=380) for i in range (n)]
for i in range (n):
ps[i].circle()
plts.append(ps[i])
p = gridplot(plts, ncols=2, responsive=False)
return p
``
and change the layout by calling .children
def update_ticker(attrname, old, new):
n = ticker.value
layout.children[1] = figure_ticker(int(n))
``
with the excel spreadsheet i make correspondence with the database (for ColumnDataSource). That’s interesting because to change the plot configuration I just need to change the excel spreadsheet but I have find the error that you explained very well. So I’ll implement your suggestion and use hide/show.
Thanks,
Rafael Pais
···
Em sexta-feira, 16 de dezembro de 2016 15:25:31 UTC-2, Bryan Van de ven escreveu:
Hi,
It looks like you are creating new plots and adding them? There are some know issues with regards to adding new objects to documents. The primary initial use-case we sought, and the one that is very well demonstrated is rather to update existing objects in the document. That is, change the data in a data source to update a plot, or change the values properties on an existing objects to update titles or colors or transparency, etc. We certainly intend to address these issues, as the limitations of human resources permit. In the mean time the best general guidance I can give is to try an structure your app as much as possible around creating one set of models once, up front, then updating them. If you need to hide/show things there are .visible properties in some cases, or alpha values that can be set to zero, etc.
I have a layout that returns a variable number of plot depending on the parameter selected in Select Widget. I have observed this intermittent error message in bokeh server and I do not know what it means. I have observed that the error occurred mainly when there are many plots (>10). What can I do to fix this problem? My code is here if necessary.
.
Thanks for any help.
Rafael Pais
plataform: win-64
conda version 4.2.9
bokeh 0.12.2
–
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].