[chimera-dev] Chimera extension threads

[Tanja Štular]

Hey,

Thanks a lot. I will create files like you suggested.

Regards,
Tanja

2016-05-18 19:48 GMT+02:00 Eric Pettersen <pett at cgl.ucsf.edu>:

> Ah, I see, these aren’t actually molecules — they’re grids.  Therefore
> Chimera wastes a lot of time looking for bonds that aren’t actually
> relevant.  I assume you have control over how these files are created.
> What you should do is put the grid points in HETATM records instead of ATOM
> records, and each in its own residue.  This will prevent Chimera from
> looking for any bonds.  The way you have it set up now, Chimera is looking
> at each atom pair to see if it’s within bonding distance.  With the setup I
> propose, there are no other atoms within the residue to look at for
> possible bonds, and the use of HETATM records prevents the creation of
> bonds between consecutive residues.
>
> I’ve attached a version of your first file with these changes.  You can
> see it opens quickly.  You can’t put a surface on it, but you can change
> the atoms to “sphere” mode (Actions->Atoms/Bonds->sphere) to get a
> space-filling depiction.
>
> —Eric
>
>
>
> On May 18, 2016, at 7:19 AM, Tanja Štular <ts6544 at student.uni-lj.si>
> wrote:
>
> Hi all,
>
> Here is some information regarding the questions:
>
> Yes, my plugin is written in Python and it communicates with several
> external web services. I tried loading the object with initial smart
> display set to false, but there is no noted difference.
>
> I'm not sure that the loading can be speeded up since the data is large
> (I'm attaching two examples). I would just like to send the loading process
> to the separate thread, then the plugin, so that plugin stays interactive.
> Is there an option for that?
>
> Here the command that I use to load and change the structure:
>
> chimera.openModels.open(data/grid-"+name+".pdb", identifyAs="bsite-"+str(id))
>
> chimera_id = [objekt.id for objekt in chimera.openModels.list() if objekt.name == 'bsite-'+str(id)][0]
>
> chimera.runCommand("~show #"+str(chimera_id))
>
> chimera.runCommand("vdw #"+str(chimera_id))
>
> chimera.runCommand("color "+colors[id]+" #"+str(chimera_id))
>
> Basically what I'm doing is load the object and then change its surface.
>
> I tried running the commands through SubprocessMonitor.Popen with flag
> daemon = True. Shouldn't this work in the background?
>
> Thanks,
>
> Tanja
>
>
> 2016-05-17 19:44 GMT+02:00 Eric Pettersen <pett at cgl.ucsf.edu>:
>
>> Hi Tanja,
>> Well, depending on you data it may still be possible to speed things up
>> somewhat.  It’s possible that computing the initial “smart display” is slow
>> with your structure.  You can test this out interactively by going to
>> Favorites->Preferences, switch to the New Molecules category of
>> Preferences, and change “smart initial display” from “true” to “false”,
>> then open your structure.  If it’s significantly faster to open with smart
>> display off, then let me know what calls you are making to open the
>> structure and I’ll tell you what to change to prevent smart display from
>> happening.
>>
>> —Eric
>>
>> Eric Pettersen
>> UCSF Computer Graphics Lab
>>
>> On May 16, 2016, at 3:06 PM, Tanja Štular <ts6544 at student.uni-lj.si>
>> wrote:
>>
>> Dear Conrad,
>>
>> Thank you for fast reply. Unfortunately, I'm having the problem due to
>> the large data. As I understand correctly for now there is no way to solve
>> this problem? I would be happy even with just making a plugin responsive
>> while Chimera is loading.
>>
>> Regards,
>> Tanja
>>
>> 2016-05-16 23:53 GMT+02:00 Conrad Huang <conrad at cgl.ucsf.edu>:
>>
>>> Hi, Tanja.
>>>
>>> I think the answer depends on where the bottleneck lies.  If the problem
>>> is that your plugin data is large and Chimera is just slow to process the
>>> data set, then there's not a whole lot we can do to speed that up. (That's
>>> why we're working on ChimeraX.)  If, on the other hand, the issue is that
>>> the plugin takes some time to send data back, I may be able to help.  Do
>>> you have some idea of which type of problem you're encountering?  Thanks.
>>>
>>> Conrad
>>>
>>>
>>> On 5/16/2016 9:49 AM, Tanja Štular wrote:
>>>
>>>> Dear Sir or Madam,
>>>>
>>>> I have been developing an extension for UCSF Chimera. I'm wondering if
>>>> there is any option for background threading for loading object to
>>>> Chimera view? Our plugin allows loading large molecules to Chimera which
>>>> sometimes causes freezing of Chimera and plugin for a little time. I
>>>> know about Subprocess class
>>>> (
>>>> https://www.cgl.ucsf.edu/chimera/docs/ProgrammersGuide/Examples/Main_RunSubprocess.html
>>>> )
>>>> but it doesn't seem to work (same goes for python threading).
>>>>
>>>> Thank you in advance.
>>>>
>>>> Best regards,
>>>> Tanja
>>>>
>>>>
>>>> _______________________________________________
>>>> Chimera-dev mailing list
>>>> Chimera-dev at cgl.ucsf.edu
>>>> http://www.rbvi.ucsf.edu/mailman/listinfo/chimera-dev
>>>>
>>>>
>>>
>> _______________________________________________
>> Chimera-dev mailing list
>> Chimera-dev at cgl.ucsf.edu
>> http://www.rbvi.ucsf.edu/mailman/listinfo/chimera-dev
>>
>>
>>
> <grid-2y03.35.Y01.401.A.A_3.pdb><grid-1all.43.CYC.175.A.A_0.pdb>
> _______________________________________________
> Chimera-dev mailing list
> Chimera-dev at cgl.ucsf.edu
> http://www.rbvi.ucsf.edu/mailman/listinfo/chimera-dev
>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://plato.cgl.ucsf.edu/pipermail/chimera-dev/attachments/20160518/03b8e4a2/attachment.html>