[Bioc-devel] Replacement for CytoscapeWindowClass from devel RCy3

Robert M. Flight rflight79 at gmail.com
Sat Mar 31 02:18:18 CEST 2018


Thanks again Alex!

Copying bioc-devel list in case anyone else has similar issues.

On Fri, Mar 30, 2018, 8:15 PM Alex Pico <alex.pico at gladstone.ucsf.edu>
wrote:

> I think that’s correct.  A few tips based on looking at your particular
> usage…
>
> 1) cw <- CytoscapeWindow(graphName, graph=ccGraph, …)
>
> ...becomes
>
> net.suid <- createNetworkFromGraph(graph=ccGraph,title=graphName, ...)
>
> 2) You don’t need `displayGraph(cw)` anymore
>
> 3) You can simply delete `cw` from all subsequent functions. Optionally,
> you can replace it with `network=net.suid`
>
> 4) You don’t need `redraw(cw)` anymore
>
> 5) Your “Rule” functions become...
>
> setNodeTooltipMapping(nodeAtts[toolTipLoc[1]])
>
> setNodeColorMapping(‘fillcolor’, tmpCols, tmpCols, ‘d’, default.color='
> #FF0000’)
> # or, if you have a column of hex colors already, try this:
> setNodeColorMapping(‘fillcolor’, tmpCols, mapping.type= ‘p’,
> default.color='#FF0000’)
>
> setNodeShapeMapping(‘shape’, nodeShapes, nodeShapes, ‘d’, default.shape='
> ellipse' )
> # again, if you have a column of hex colors already, try this:
> setNodeShapeMapping(‘shape’, nodeShapes, mapping.type= ‘p’, default.shape=
> 'ellipse')
>
>
> Ok. Hope that helps!
>  - Alex
>
> On Mar 30, 2018, at 4:54 PM, Robert M. Flight <rflight79 at gmail.com> wrote:
>
> OK, I'll have to try out some of the examples and see if it works. In most
> of my methods it would just mean removing that part of the signature, I
> think.
>
> Robert
>
> On Fri, Mar 30, 2018, 7:44 PM Alex Pico <alex.pico at gladstone.ucsf.edu>
> wrote:
>
>> Correct. This simplification comes at the cost of Object types. The
>> advantages, however, include avoiding a false sense of state when CyREST is
>> stateless. There was an whole class of errors and user issues based on
>> breaks between the R Object and the Cytoscape network. Now, you can even
>> restart R and pick right up in the middle of script (as far as RCy3 is
>> concerned).
>>
>> Also, since in most cases you are working with one network at a time, you
>> can actually ignore the SUID and let the default behavior (of working with
>> the “current” network) take over and greatly simplify your scripts.
>>
>>  - Alex
>>
>> On Mar 30, 2018, at 4:05 PM, Robert M. Flight <rflight79 at gmail.com>
>> wrote:
>>
>> Alex,
>>
>> That makes sense. Appreciate the info on SUID. Again this is only a
>> simple numeric though, right?? No class attribrute or formal class with it??
>>
>> I like the fact that currently my methods know they are operating on a
>> cytoscape network object, this breaks that paradigm.
>>
>> Will have to think about the best way to handle this.
>>
>> Robert
>>
>> On Fri, Mar 30, 2018, 6:50 PM Alex Pico <alex.pico at gladstone.ucsf.edu>
>> wrote:
>>
>>> Hi Robert,
>>>
>>> Indeed, lots of changes being made in this major version bump to RCy3.
>>> The goals are to align better with the rapid development of Cytoscape’s
>>> REST API. A description of how to “upgrade” existing scripts is provided
>>> here:
>>> https://github.com/cytoscape/RCy3/wiki/Upgrading-Existing-Scripts
>>>
>>>
>>> In the case you pointed to the upgrade is not too bad (see attached):
>>>  * Replace all instances/references to CyWindow Objects with a simple
>>> numeric, the network SUID.
>>> *Note: that even this reference is now optional for all functions, i.e.,
>>> you can leave it out completely if you know that you are operating on the
>>> “current” network in Cytoscape. If you’re not sure, then this is how you
>>> specify networks now.*
>>>  * Replaced getAllEdgeAttributes with getTableColumns(‘edge’)
>>>
>>> That’s it!  If you have any more questions, please ask.
>>>
>>>  - Alex
>>>
>>> On Mar 30, 2018, at 9:13 AM, Robert M. Flight <rflight79 at gmail.com>
>>> wrote:
>>>
>>> My package, categoryCompare, is failing in Bioc devel.
>>>
>>> The reason appears to be exports in previous versions of RCy3 no longer
>>> being exported, namely the "CytoscapeWindowClass".
>>>
>>> I had some methods where behavior was defined based on being passed an
>>> object of "CytoscapeWindowClass" (see
>>> https://github.com/rmflight/categoryCompare/blob/7305185a65f80b108d50971b350c165a4e0cbe47/R/outCytoscape.r#L37 for
>>> an example).
>>>
>>> It appears this class no longer exists in the devel version of RCy3, and
>>> what is noted as a replacement for the deprecated function that created it
>>> (example being createNetworkFromGraph), has no specific class outside of
>>> returning a "function".
>>>
>>> I could use some advice on how I should modify my method signature to
>>> handle this for the next version of RCy3, or if the object returned will
>>> indeed have a "class" attached to them that can be checked via method
>>> signatures.
>>>
>>> Thanks,
>>>
>>> -Robert
>>>
>>> Robert M Flight, PhD
>>> Bioinformatics Research Associate
>>> Puller of Rabbits from Hats
>>> Research Parasite
>>> Resource Center for Stable Isotope Resolved Metabolomics
>>> Manager, Systems Biology and Omics Integration Journal Club
>>> Markey Cancer Center
>>> CC434 Roach Building
>>> University of Kentucky
>>> Lexington, KY
>>>
>>> Twitter: @rmflight
>>> Web: rmflight.github.io
>>> ORCID: http://orcid.org/0000-0001-8141-7788
>>> EM rflight79 at gmail.com
>>> PH 502-509-1827 <(502)%20509-1827>
>>>
>>> To call in the statistician after the experiment is done may be no more
>>> than asking him to perform a post-mortem examination: he may be able to say
>>> what the experiment died of. - Ronald Fisher
>>>
>>>
>>>
>>
>

	[[alternative HTML version deleted]]



More information about the Bioc-devel mailing list