[Bioc-devel] No build report today

[Shepherd, Lori]

The experiment data build report from yesterday  and the  software build report 
for today will not be posted.  There was an ERROR generating the builds 
stemming from the SSL certificate issues we were having.  The issues have been 
remedied and we expect the build report to post as normal tomorrow.

Thanks for your understanding


Lori Shepherd

Bioconductor Core Team

Roswell Park Comprehensive Cancer Center

Department of Biostatistics & Bioinformatics

Elm & Carlton Streets

Buffalo, New York 14263


This email message may contain legally privileged and/or confidential 
information.  If you are not the intended recipient(s), or the employee or 
agent responsible for the delivery of this message to the intended 
recipient(s), you are hereby notified that any disclosure, copying, 
distribution, or use of this email message is prohibited.  If you have received 
this message in error, please notify the sender immediately by e-mail and 
delete this email message from your computer. Thank you.
[[alternative HTML version deleted]]

___
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel

[Rd] Error in [.terms

[Therneau, Terry M., Ph.D. via R-devel]

Martin,

   There are a couple of issues with [.terms that have bitten my survival code. 
 At the 
useR conference I promised you a detailed (readable) explanation, and have been 
lax in 
getting it to you. The error was first pointed out in a bugzilla note from 
2016, by the 
way.  The current survival code works around these.

Consider the following formula:

<>=
library(survival)  # only to get access to the lung data set
test <- Surv(time, status) ~  age + offset(ph.ecog) + strata(inst)
tform <- terms(test, specials="strata")
mf <- model.frame(tform, data=lung)
mterm <- terms(mf)
@

The strata term is handled in a special way by coxph, and then needs to be 
removed from 
the model formula before calling model.matrix.
To do this the code uses essentially the following, which fails for the formula 
above.

<>=
strata <- attr(mterm, "specials")$strata - attr(mterm, "response")
X <- model.matrix(mterm[-strata], mf)
@

The root problem is the need for multiple subscripts.
\begin{itemize}
   \item The formula itself has length 5, with `~' as the first element
   \item The variables and predvars attributes are call objects, each a list() 
with 4 
elments: the response and all 3 predictors
   \item The term.labels attribute omits the resonse and the offset, so has  
length 2
   \item The factors attribute has 4 rows and 2 columns
   \item The dataClasses attribute is a character vector of length 4
\end{itemize}

So the ideal result of  mterm[remove the specials] would use subscript of
\begin{itemize}
   \item [-5] on the formula itself, variables and predvars attributes
   \item [-2] for term.labels
   \item [-4 , -2, drop=FALSE] for factor attribute
   \item [-2] for order attribute
   \item [-4] for the dataClasses attribute
\end{itemize}

That will recreate the formula that ``would have been'' had there been no 
strata term.  
Now look at the first portion of the code in models.R
<<>>=
`[.terms` <- function (termobj, i)
{
     resp <- if (attr(termobj, "response")) termobj[[2L]]
     newformula <- attr(termobj, "term.labels")[i]
     if (length(newformula) == 0L) newformula <- "1"
     newformula <- reformulate(newformula, resp, attr(termobj, "intercept"), 
environment(termobj))
     result <- terms(newformula, specials = names(attr(termobj, "specials")))

     # Edit the optional attributes
}
@

The use of reformulate() is a nice trick.  However, the index reported in the 
specials 
attribute is generated with reference to the variables
attribute, or equivalently the row names of the factors attribute, not with 
respect to the 
term.labels attribute. For consistency the second line should instead be
<<>>=
newformula <- row.names(attr(termobj, "factors"))[i]
@

Of course, this will break code for anyone else who has used [.terms and, like 
me, has 
been adjusting for the ``response is counted in specials but
not in term.labels'' feature.  R core will have to discuss/decide what is the 
right thing 
to do, and I'll adapt.

The reformulate trick breaks in another way, one that only appeared on my radar 
this week 
via a formula like the following.

<>=
Surv(time, status) ~ age + (sex=='male') + strata(inst)
@

In both the term.labels attribute and the row/col names of the factors 
attribute the 
parentheses disappear, and the result of the reformulate call is not a proper 
formula.  
The + binds tighter than == leading to an error message that will confuse most 
users. We 
can argue, and I probably would, that the user should have used I(sex=='male'). 
 But they 
didn't, and without the I() it is a legal formula, or at least one that 
currently works.  
Fixing this issue is a lot harder.

An offset term causes issues in the 'Edit the optional attributes' part of the 
routine as 
well.  If you and/or R core will tell me what you think
the code should do, I'll create a patch.  My vote would be to use rownames per 
the above 
and ignore the () edge case.

The same basic code appears in drop.terms, by the way.

Terry T.


[[alternative HTML version deleted]]

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel

Re: [Bioc-devel] pwrEWAS - checkResults error

[Shepherd, Lori]

The experiment data builds were suppose to post yesterday but it appears there 
was a delay on our end.  Once those post the data experiment package will be 
available for your software package.  Please wait just a little bit longer.  
Sorry for the delay.


Lori Shepherd

Bioconductor Core Team

Roswell Park Comprehensive Cancer Center

Department of Biostatistics & Bioinformatics

Elm & Carlton Streets

Buffalo, New York 14263


From: Graw, Stefan H 
Sent: Friday, October 4, 2019 9:10 AM
To: Graw, Stefan H ; Shepherd, Lori 
; bioc-devel@r-project.org 

Subject: RE: pwrEWAS - checkResults error

Dear Bioc-Team,

CheckResults is still resulting in the error where pwrEWAS cannot find the 
dependency pwrEWAS.data, which is a related ExperimentHub package:

ERROR: dependency 'pwrEWAS.data' is not available for package 'pwrEWAS'

How do I address this issue?

Thank you,
Stefan


-Original Message-
From: Bioc-devel  On Behalf Of Graw, Stefan H
Sent: Monday, September 30, 2019 9:46 AM
To: Shepherd, Lori ; bioc-devel@r-project.org
Subject: Re: [Bioc-devel] pwrEWAS - checkResults error

Hey Lori,

Yes, pwrEWAS.data is a separate ExperimentHub package to cache the data.
The package was originally submitted as pwrEWAS_data and subsequently changed 
to pwrEWAS.data. Maybe that cause the error.

Thank you for looking into this.
Stefan

From: Shepherd, Lori 
Sent: Monday, September 30, 2019 9:40 AM
To: Graw, Stefan H ; bioc-devel@r-project.org
Subject: Re: pwrEWAS - checkResults error

There is a separate Experiment data package pwrEWAS.data correct?

I think this is our mistake and it wasn't added to the experiment data package 
manifest.  We will correct on our end.  The next Data Experiment package builds 
will happen Tuesday so the software package will not clear up until later in 
the week.



Lori Shepherd

Bioconductor Core Team

Roswell Park Comprehensive Cancer Center

Department of Biostatistics & Bioinformatics

Elm & Carlton Streets

Buffalo, New York 14263


From: Bioc-devel 
mailto:bioc-devel-boun...@r-project.org>> on 
behalf of Graw, Stefan H mailto:shg...@uams.edu>>
Sent: Monday, September 30, 2019 10:26 AM
To: bioc-devel@r-project.org 
mailto:bioc-devel@r-project.org>>
Subject: [Bioc-devel] pwrEWAS - checkResults error

Dear Bioc-Team,

A packages (pwrEWAS) I submitted caused an error in checkResults:

https://urldefense.proofpoint.com/v2/url?u=https-3A__bioconductor.org_checkResults_3.10_bioc-2DLATEST_pwrEWAS_malbec1-2Dinstall.html=DwICAg=27AKQ-AFTMvLXtgZ7shZqsfSXu-Fwzpqk4BoASshREk=i78VfP1bq28ww6AUFrJ6Ww=qi9ynP2AS8G5-Z6i5W3fuLDGTRJN6wZO1PSc0dlCyOo=lszvydHABltQmpE9440StzNTtxbaYHtQJ8cCAgsAu00=

ERROR: dependency 'pwrEWAS.data' is not available for package 'pwrEWAS'

The error was caused by failing to load the related ExperimentHub package 
(pwrEWAS.data).
How do I address this issue?

Thank you,
Stefan




--
Confidentiality Notice: This e-mail message, including a...{{dropped:28}}

___
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel

Re: [Bioc-devel] pwrEWAS - checkResults error

[Graw, Stefan H]

Dear Bioc-Team,

CheckResults is still resulting in the error where pwrEWAS cannot find the 
dependency pwrEWAS.data, which is a related ExperimentHub package:

ERROR: dependency 'pwrEWAS.data' is not available for package 'pwrEWAS'

How do I address this issue?

Thank you,
Stefan


-Original Message-
From: Bioc-devel  On Behalf Of Graw, Stefan H
Sent: Monday, September 30, 2019 9:46 AM
To: Shepherd, Lori ; bioc-devel@r-project.org
Subject: Re: [Bioc-devel] pwrEWAS - checkResults error

Hey Lori,

Yes, pwrEWAS.data is a separate ExperimentHub package to cache the data.
The package was originally submitted as pwrEWAS_data and subsequently changed 
to pwrEWAS.data. Maybe that cause the error.

Thank you for looking into this.
Stefan

From: Shepherd, Lori 
Sent: Monday, September 30, 2019 9:40 AM
To: Graw, Stefan H ; bioc-devel@r-project.org
Subject: Re: pwrEWAS - checkResults error

There is a separate Experiment data package pwrEWAS.data correct?

I think this is our mistake and it wasn't added to the experiment data package 
manifest.  We will correct on our end.  The next Data Experiment package builds 
will happen Tuesday so the software package will not clear up until later in 
the week.



Lori Shepherd

Bioconductor Core Team

Roswell Park Comprehensive Cancer Center

Department of Biostatistics & Bioinformatics

Elm & Carlton Streets

Buffalo, New York 14263


From: Bioc-devel 
mailto:bioc-devel-boun...@r-project.org>> on 
behalf of Graw, Stefan H mailto:shg...@uams.edu>>
Sent: Monday, September 30, 2019 10:26 AM
To: bioc-devel@r-project.org 
mailto:bioc-devel@r-project.org>>
Subject: [Bioc-devel] pwrEWAS - checkResults error

Dear Bioc-Team,

A packages (pwrEWAS) I submitted caused an error in checkResults:

https://urldefense.proofpoint.com/v2/url?u=https-3A__bioconductor.org_checkResults_3.10_bioc-2DLATEST_pwrEWAS_malbec1-2Dinstall.html=DwICAg=27AKQ-AFTMvLXtgZ7shZqsfSXu-Fwzpqk4BoASshREk=i78VfP1bq28ww6AUFrJ6Ww=qi9ynP2AS8G5-Z6i5W3fuLDGTRJN6wZO1PSc0dlCyOo=lszvydHABltQmpE9440StzNTtxbaYHtQJ8cCAgsAu00=

ERROR: dependency 'pwrEWAS.data' is not available for package 'pwrEWAS'

The error was caused by failing to load the related ExperimentHub package 
(pwrEWAS.data).
How do I address this issue?

Thank you,
Stefan




--
Confidentiality Notice: This e-mail message, including a...{{dropped:17}}

___
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel

[Bioc-devel] Package Submission Deadline is TODAY!!

[Shepherd, Lori]

The final day to submit new packages to the Bioconductor contributions tracker 
to have a shot at being included in the upcoming 3.10 release is TODAY 5 PM EST!

Please note: submission by this date does not guarantee it will be included - 
the package must undergo an official review and be accepted by Wednesday 
October 23th.

Submit new packages here:
https://github.com/Bioconductor/Contributions
[https://avatars2.githubusercontent.com/u/2286807?s=400=4]
Bioconductor/Contributions: Contribute Packages to Bioconductor - The world's 
leading software development platform � 
GitHub
Join GitHub today. GitHub is home to over 40 million developers working 
together to host and review code, manage projects, and build software together.
github.com


Package submitted after this date or that are submitted but do not pass the 
review process in time can still be accepted to Bioconductor but will be only 
available in the development version until the Spring release.


Lori Shepherd

Bioconductor Core Team

Roswell Park Comprehensive Cancer Center

Department of Biostatistics & Bioinformatics

Elm & Carlton Streets

Buffalo, New York 14263


This email message may contain legally privileged and/or confidential 
information.  If you are not the intended recipient(s), or the employee or 
agent responsible for the delivery of this message to the intended 
recipient(s), you are hereby notified that any disclosure, copying, 
distribution, or use of this email message is prohibited.  If you have received 
this message in error, please notify the sender immediately by e-mail and 
delete this email message from your computer. Thank you.
[[alternative HTML version deleted]]

___
Bioc-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/bioc-devel