I’ve had good success modifying the columnDataSource for a plot and then calling push_notebook() to cause the new information to appear in the graph (running in a Jupyter python 3 notebook). However when I attempt the same approach with a Bokeh table, calling push_notebook() doesn’t cause the table to update. However, a subsequent call “show” will cause the table to display the updated information.
The example below illustrates the situation where I print the data that the source for the table showing that it now only contains two rows of data. However, even after calling push_notebook() the table still shows the old values consisting of many rows. If I then run the following cell to “show” the table again it will correctly reflect only the two rows described by its source.
Should push_notebook cause a table to refresh to reflect an updated data source? Or is there some other way I can trigger it to update other than calling “show” which I can’t do since the table needs to be displayed from a separate notebook cell (strange things happen if I try to call Bokeh show in the callback of an ipywidget button since both the table and button would display together creating a problem in the notebook cell publishing machinery)
It occurred to me after making this post that it could be a case of push_notebook only updating the last shown object. I do have a number of different Bokeh objects visible in this notebook and the source of the data that is being displayed in this table is derived from the selected points in a bokeh scatter plot, so I suppose that might be the “last shown” object. Is there any known workaround to this? I’m ok with even creating a button to force the refresh/update if there is something I could do in the CustomJS of it’s callback.
I can't think of anything offhand. The comms problem was introduced at the last minute by an important hotfix. I intend to look into it in the next few days, and will have a fix in a dev build as soon as possible.
It occurred to me after making this post that it could be a case of push_notebook only updating the *last shown object*. I do have a number of different Bokeh objects visible in this notebook and the source of the data that is being displayed in this table is derived from the selected points in a bokeh scatter plot, so I suppose that might be the "last shown" object. Is there any known workaround to this? I'm ok with even creating a button to force the refresh/update if there is something I could do in the CustomJS of it's callback.
I’ve looked through the open issues on github but didn’t see anything specific to this particular comms bug. Is there some place I should be watching for progress on it’s resolution (Is there a particular github issue that covers it that I missed and where can I find any notes for what is in the dev builds?). T
he problem is getting to be more of a big deal on our end as time passes and at some point in the near future will basically amount to a showstopper issue that would force us to move away from bokeh.
···
On Friday, January 15, 2016 at 10:00:12 PM UTC-7, Bryan Van de ven wrote:
I can’t think of anything offhand. The comms problem was introduced at the last minute by an important hotfix. I intend to look into it in the next few days, and will have a fix in a dev build as soon as possible.
It occurred to me after making this post that it could be a case of push_notebook only updating the last shown object. I do have a number of different Bokeh objects visible in this notebook and the source of the data that is being displayed in this table is derived from the selected points in a bokeh scatter plot, so I suppose that might be the “last shown” object. Is there any known workaround to this? I’m ok with even creating a button to force the refresh/update if there is something I could do in the CustomJS of it’s callback.
–
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].
I've looked through the open issues on github but didn't see anything specific to this particular comms bug. Is there some place I should be watching for progress on it's resolution (Is there a particular github issue that covers it that I missed and where can I find any notes for what is in the dev builds?). T
he problem is getting to be more of a big deal on our end as time passes and at some point in the near future will basically amount to a showstopper issue that would force us to move away from bokeh.
On Friday, January 15, 2016 at 10:00:12 PM UTC-7, Bryan Van de ven wrote:
I can't think of anything offhand. The comms problem was introduced at the last minute by an important hotfix. I intend to look into it in the next few days, and will have a fix in a dev build as soon as possible.
Bryan
> On Jan 15, 2016, at 10:44 PM, Jim Sharpe <[email protected]> wrote:
>
> It occurred to me after making this post that it could be a case of push_notebook only updating the *last shown object*. I do have a number of different Bokeh objects visible in this notebook and the source of the data that is being displayed in this table is derived from the selected points in a bokeh scatter plot, so I suppose that might be the "last shown" object. Is there any known workaround to this? I'm ok with even creating a button to force the refresh/update if there is something I could do in the CustomJS of it's callback.
>
> --
> 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 bokeh+un...@continuum.io.
> To post to this group, send email to bo...@continuum.io.
> To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/1fc8d28b-3a54-490f-a9dc-432a724171b5%40continuum.io\.
> For more options, visit https://groups.google.com/a/continuum.io/d/optout\.
I’ve looked through the open issues on github but didn’t see anything specific to this particular comms bug. Is there some place I should be watching for progress on it’s resolution (Is there a particular github issue that covers it that I missed and where can I find any notes for what is in the dev builds?). T
he problem is getting to be more of a big deal on our end as time passes and at some point in the near future will basically amount to a showstopper issue that would force us to move away from bokeh.
On Friday, January 15, 2016 at 10:00:12 PM UTC-7, Bryan Van de ven wrote:
I can’t think of anything offhand. The comms problem was introduced at the last minute by an important hotfix. I intend to look into it in the next few days, and will have a fix in a dev build as soon as possible.
It occurred to me after making this post that it could be a case of push_notebook only updating the last shown object. I do have a number of different Bokeh objects visible in this notebook and the source of the data that is being displayed in this table is derived from the selected points in a bokeh scatter plot, so I suppose that might be the “last shown” object. Is there any known workaround to this? I’m ok with even creating a button to force the refresh/update if there is something I could do in the CustomJS of it’s callback.
Looks like the changes aren’t working for tables. I just tried a test with 11.1.dev2 where I update the data in the source for a table and nothing happens when I then call push_notebook. However, if I then make another explicit call to show() the table the updated values appear in the new copy. Note that this is using bokeh offline in a python 3 jupyter notebook.
···
On Thursday, January 21, 2016 at 3:02:42 PM UTC-7, Jim Sharpe wrote:
That’s fantastic news! I’ll go grab it right now.
On Thursday, January 21, 2016 at 2:52:02 PM UTC-7, Bryan Van de ven wrote:
The fix for comms handles is in 0.11.1dev2 which was published today. This notebook shows the expected behavior with plots:
I’ve looked through the open issues on github but didn’t see anything specific to this particular comms bug. Is there some place I should be watching for progress on it’s resolution (Is there a particular github issue that covers it that I missed and where can I find any notes for what is in the dev builds?). T
he problem is getting to be more of a big deal on our end as time passes and at some point in the near future will basically amount to a showstopper issue that would force us to move away from bokeh.
On Friday, January 15, 2016 at 10:00:12 PM UTC-7, Bryan Van de ven wrote:
I can’t think of anything offhand. The comms problem was introduced at the last minute by an important hotfix. I intend to look into it in the next few days, and will have a fix in a dev build as soon as possible.
It occurred to me after making this post that it could be a case of push_notebook only updating the last shown object. I do have a number of different Bokeh objects visible in this notebook and the source of the data that is being displayed in this table is derived from the selected points in a bokeh scatter plot, so I suppose that might be the “last shown” object. Is there any known workaround to this? I’m ok with even creating a button to force the refresh/update if there is something I could do in the CustomJS of it’s callback.
Will require some dedicated investigation then, can you make an issue on GH? If you have a notebook laying around that you can share that will help tremendously with getting started.
Looks like the changes aren't working for tables. I just tried a test with 11.1.dev2 where I update the data in the source for a table and nothing happens when I then call push_notebook. However, if I then make another explicit call to show() the table the updated values appear in the new copy. Note that this is using bokeh offline in a python 3 jupyter notebook.
I apologize in advance if I’ve misunderstood how updating data sources should work and would welcome any pointers as to what I’m doing wrong. I’ve attached a small example notebook that I believe reproduces the behavior of my other, much larger, application. Although it only uses a couple of simple dummy data structures it appears to demonstrate the same behavior. (I.e. the table doesn’t update after modifying the data source and then calling push_notebook but the updated information is reflected if I show it again.) I think it should be fairly self explanatory but please let me know if you have any questions.
On Tuesday, January 26, 2016 at 1:52:58 PM UTC-7, Jim Sharpe wrote:
Looks like the changes aren’t working for tables. I just tried a test with 11.1.dev2 where I update the data in the source for a table and nothing happens when I then call push_notebook. However, if I then make another explicit call to show() the table the updated values appear in the new copy. Note that this is using bokeh offline in a python 3 jupyter notebook.
On Thursday, January 21, 2016 at 3:02:42 PM UTC-7, Jim Sharpe wrote:
That’s fantastic news! I’ll go grab it right now.
On Thursday, January 21, 2016 at 2:52:02 PM UTC-7, Bryan Van de ven wrote:
The fix for comms handles is in 0.11.1dev2 which was published today. This notebook shows the expected behavior with plots:
I’ve looked through the open issues on github but didn’t see anything specific to this particular comms bug. Is there some place I should be watching for progress on it’s resolution (Is there a particular github issue that covers it that I missed and where can I find any notes for what is in the dev builds?). T
he problem is getting to be more of a big deal on our end as time passes and at some point in the near future will basically amount to a showstopper issue that would force us to move away from bokeh.
On Friday, January 15, 2016 at 10:00:12 PM UTC-7, Bryan Van de ven wrote:
I can’t think of anything offhand. The comms problem was introduced at the last minute by an important hotfix. I intend to look into it in the next few days, and will have a fix in a dev build as soon as possible.
It occurred to me after making this post that it could be a case of push_notebook only updating the last shown object. I do have a number of different Bokeh objects visible in this notebook and the source of the data that is being displayed in this table is derived from the selected points in a bokeh scatter plot, so I suppose that might be the “last shown” object. Is there any known workaround to this? I’m ok with even creating a button to force the refresh/update if there is something I could do in the CustomJS of it’s callback.
I was hoping that might be the fix for you, but alas apparently not. Thanks for the notebook it will make getting started investigating much easier to do. I will try to take a poke at it next week sometime.
