[Bioc-devel] annotation() to BiocGenerics?

Fri Sep 21 19:47:53 CEST 2012

Hi Martin,

On Sep 21, 2012, at 6:07 PM, Martin Morgan wrote:

> On 09/21/2012 08:37 AM, Nicolas Delhomme wrote:
>> Hi all,
>> 
>> I have the same issue for the easyRNASeq package:
>> 
>> unable to find an inherited method for function "annotation", for
>> signature "Genome_intervals_stranded"
> 
> this doesn't show up in the build report? How can it be reproduced?

Not that I've seen. I suppose there are several reason to this, one being that for packages to be hosted on Bioc, they should run R CMD check in less than 5 minutes, which is understandable given the amount of package to be built daily. But this impacts the number of tests one can do. The other (shame on me) being that I would need to spend the time to update these tests and add proper unit tests. 

> 
>> easyRNASeq uses genomeIntervals (another Bioc package) and due to the
>> mentioned change (moving of the annotation generic), both packages
>> are now partially non-functional.
> 
> we'll clean up our mess, but thanks for the report.

I've just looked up the affected code and it should be straightforward to fix, especially given your additional comments.

> 
>> While I completely support such consolidation, this time, the
>> timeline is very tight. The change was done on the 18th, it took 2
>> days to propagate, luckily the bug was reported immediately, but it's
>> already friday evening here (21st, CET)  and the API freeze is on the
>> 24th, so if we would play by the rule, there'd just be one day left
>> to fix it (the 24th, can't expect to have someone fixing this at the
>> WE, really).
>> 
>> This raises a more general concern as there's now really a lot of
>> contributed package in Bioc that depends on the excellent core Bioc
>> API. Would not it make sense to have an earlier freeze of the core
>> Bioc API than just a week before the release? Just to make sure that
>> any effect in downstream contributed packages is timely detected? I'm
>> saying this, because the exact same happened for the 2.10 release 6
>> months ago. I would not mind sparing that extra bit of stress :-)
> 
> Sorry for the inconvenience; we do try to avoid things that will definitely be disruptive.

I know and I'm extremely grateful for all the excellent work you are all doing. Sincerely it's amazing. And without your infrastructure I could not do much of the research I'm doing, so there's no inconvenience. What I wrote was not a critic (sorry if my frenglish sounded so), it's just that I realized that Bioc is growing steadily and that I could imagine this situation to occur more frequently or affect more packages in the future. Which is why I was thinking that the core packages could have a different release schedule than the contributed packages, since many of these depend on the former. For example the core packages could have deadlines a week earlier than contributed packages. Does that sound sensible?

> 
>> I've contacted the maintainer of the genomeIntervals package to see
>> if they can use the BiocGeneric generic instead.
> 
> genomeIntervals should previously have importFrom(Biobase, annotation, "annotation<-") in its NAMESPACE.

Well it did not and that is part of the reason why it breaks/broke I believe.

Cheers,

Nico

> 
> Martin
> 
>> 
>> Robert, it would be good if you were to tell the GSVA developer. And
>> btw, thanks for the tip Benilton!
>> 
>> Cheers,
>> 
>> Nico
>> 
>> ---------------------------------------------------------------
>> Nicolas Delhomme
>> 
>> Genome Biology Computational Support
>> 
>> European Molecular Biology Laboratory
>> 
>> Tel: +49 6221 387 8310 Email: nicolas.delhomme at embl.de
>> Meyerhofstrasse 1 - Postfach 10.2209 69102 Heidelberg, Germany
>> ---------------------------------------------------------------
>> 
>> 
>> 
>> 
>> 
>> On Sep 21, 2012, at 12:49 PM, Robert Castelo wrote:
>> 
>>> Dear Valerie and other developers,
>>> 
>>> executing the vignette from package GSVA fails with an error
>>> related to the change described in your email (because before it
>>> did not appear):
>>> 
>>> Error: processing vignette ‘GSVA.Rnw’ failed with diagnostics:
>>> chunk 19 Error in function (classes, fdef, mtable)  : unable to
>>> find an inherited method for function "annotation<-", for signature
>>> "ExpressionSet", "missing" Execution halted
>>> 
>>> the main function of the package, gsva(), calls at some point
>>> 'annotation(eset)' for some ExpressionSet object 'eset'.
>>> 
>>> i've tried the following solutions:
>>> 
>>> 1. put in NAMESPACE
>>> 
>>> importFrom(Biobase, annotation)
>>> 
>>> and replace every call to 'annotation()' by
>>> 'Biobase::annotation()'
>>> 
>>> 2. put in NAMESPACE
>>> 
>>> importFrom(BiocGenerics, annotation)
>>> 
>>> and replace every call to 'annotation()' by
>>> 'BiocGenerics::annotation()'
>>> 
>>> 
>>> but none of them work, both give still the same error.
>>> 
>>> i've executed the vignette interactively and this is the
>>> traceback() and sessionInfo() at the line of code producing the
>>> error:
>>> 
>>>> gbm_es <- gsva(gbm_eset, brainTxDbSets, mx.diff=FALSE,
>>>> verbose=FALSE)$es.obs
>>> Error in function (classes, fdef, mtable)  : unable to find an
>>> inherited method for function "annotation<-", for signature
>>> "ExpressionSet", "missing"
>>>> traceback()
>>> 6: stop("unable to find an inherited method for function \"",
>>> fdef at generic, "\", for signature ", cnames) 5: function (classes,
>>> fdef, mtable) { methods <- .findInheritedMethods(classes, fdef,
>>> mtable) if (length(methods) == 1L) return(methods[[1L]]) else if
>>> (length(methods) == 0L) { cnames <- paste0("\"", sapply(classes,
>>> as.character), "\"", collapse = ", ") stop("unable to find an
>>> inherited method for function \"", fdef at generic, "\", for signature
>>> ", cnames) } else stop("Internal error in finding inherited
>>> methods; didn't return a unique method") }(list("ExpressionSet",
>>> "missing"), function (object, ..., value)
>>> standardGeneric("annotation<-"), <environment>) 4:
>>> Biobase::`annotation<-`(eScoEset, "") 3: .local(expr,
>>> gset.idx.list, annotation, ...) 2: gsva(gbm_eset, brainTxDbSets,
>>> mx.diff = FALSE, verbose = FALSE) 1: gsva(gbm_eset, brainTxDbSets,
>>> mx.diff = FALSE, verbose = FALSE)
>>> 
>>> whose interpretation goes beyond my expertise, so i hope somebody
>>> more experienced than me has a hint about the problem :)
>>> 
>>> cheers, robert.
>>> 
>>> On 09/18/2012 08:48 PM, Valerie Obenchain wrote:
>>>> 'annotation' and 'annotation<-' have been moved from Biobase
>>>> (2.17.8) to BiocGenerics (0.3.2). This should propagate through
>>>> tomorrows build.
>>>> 
>>>> Valerie
>>>> 
>>>> 
>>>> On 08/30/2012 11:35 AM, Valerie Obenchain wrote:
>>>>> Hi Vince,
>>>>> 
>>>>> Yes we can add this to BiocGenerics. I'll post back when it's
>>>>> done.
>>>>> 
>>>>> Valerie
>>>>> 
>>>>> 
>>>>> On 08/29/2012 07:25 AM, Vincent Carey wrote:
>>>>>> currently in Biobase
>>>>>> 
>>>>>> [[alternative HTML version deleted]]
>>>>>> 
>>>>>> _______________________________________________
>>>>>> Bioc-devel at r-project.org mailing list
>>>>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>>>> 
>>>>> _______________________________________________
>>>>> Bioc-devel at r-project.org mailing list
>>>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>>> 
>>>> _______________________________________________
>>>> Bioc-devel at r-project.org mailing list
>>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>>>> 
>>> 
>>> -- Robert Castelo, PhD Associate Professor Dept. of Experimental
>>> and Health Sciences Universitat Pompeu Fabra (UPF) Barcelona
>>> Biomedical Research Park (PRBB) Dr Aiguader 88 E-08003 Barcelona,
>>> Spain telf: +34.933.160.514 fax: +34.933.160.550
>>> 
>>> _______________________________________________
>>> Bioc-devel at r-project.org mailing list
>>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>> 
>> _______________________________________________
>> Bioc-devel at r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/bioc-devel
>> 
> 
> 
> -- 
> Computational Biology / Fred Hutchinson Cancer Research Center
> 1100 Fairview Ave. N.
> PO Box 19024 Seattle, WA 98109
> 
> Location: Arnold Building M1 B861
> Phone: (206) 667-2793