In TIBCOSoftware/RinR: Evaluate R Expressions in a Different R Interpreter
The RinR package provides functions for running code in one R interpreter
from within another R interpreter.
The R interpreters can be different versions of open-source R
or the TERR (TIBCO Enterprise Runtime for R) interpreter.
The RCompare function in the package can run an expression in two different
R interpreters and compare the results.
The RGraph function can be used to generate graphics in TERR
by calling graphics functions in an open-source R interpreter.
The RinR package is included with TERR 2.0 (and later versions).
The package can be downloaded from the TERR community site for
use with R.
The package should also soon be available on CRAN.
Along with the RinR package, this vignette uses the MASS package that is
included with open-source R.
The minimax and devtools packages from CRAN are also used.
library(RinR)
Evaluators
The main functions of the RinR package are the evaluators.
These functions execute a given expression in the specified R interpreter.
The specified R interpreter is started with these command
line arguments: "--quiet --vanilla --no-echo".
Note that this means that neither the system Rprofile.site nor
the user's .Rprofile (for R) or .TERRprofile for TERR will be run.
The return value from an evaluator is a list of length 1
containing the value of the evaluated expression.
The name of the list component is the version$verstion.string
from the R interpreter used in the evaluation.
The REvaluate function returns just the value of the
evaluated expression in the specified R interpreter.
You can have many evaluator functions, one for each R interpreter
that you want to evaluate an expression in.
The makeEvaluator function is used to create an evaluator
for a particular R interpreter.
vs_4.0.5 <-
list(`R version 4.0.5 (2021-03-31)` = "R version 4.0.5 (2021-03-31)")
vs_4.1.2 <-
list(`R version 4.1.2 (2021-11-01)` = "R version 4.1.2 (2021-11-01)")
R4.1.2 <- makeREvaluator("R", RHome="/home/R/R-4.1.2")
R4.1.2(version$version.string)
## $`R version 4.1.2 (2021-11-01)`
## [1] "R version 4.1.2 (2021-11-01)"
R4.0.5 <- makeREvaluator("R", RHome="/home/R/R-4.0.5")
R4.0.5(version$version.string)
## $`R version 4.0.5 (2021-03-31)`
## [1] "R version 4.0.5 (2021-03-31)"
The RinR package comes with several default evaluators.
LocalEvaluator - the current interpreter (R or TERR)REvaluator - an R interpreterTERREvaluator - a TERR interpreter
These evaluators are created with the configureREvaluator function
when the RinR package is loaded.
Finding the R Interpreters
The RinR package must be able to find the various R interpreters that you
want the package to evaluate code in.
RinR tries to find the location of the R interpreter based on the values of
options("RinR_R_FULL_PATH"),options("RinR_TERR_FULL_PATH")- values in the Windows registry,
- commonly-used locations in the file system,
- directories in
Sys.getenv("PATH").
If you regularly use the RinR package it is recommended what you set the
options RinR_R_FULL_PATH and/or RinR_TERR_FULL_PATH in your
startup file.
The value of the option should be the full path to the R binary e.g.:
options(RinR_R_FULL_PATH="/usr/local/bin/R-4.0.5")
options(RinR_TERR_FULL_PATH="/usr/local/bin/TERR-6.0")
This will ensure that you always get the specified R interpreter when using
the default REvaluator and TERREvaluator functions.
If a default evaluators could not be configured when the
package is loaded,
the evaluator will look in the system path (Sys.getenv("PATH"))
for the appropriate interpreter when the evaluator function is called.
The pushPath function can add the path to an R interpreter to the system path e.g.
pushPATH("/home/R/R-4.0.5/bin")
The popPATH function can be used to remove the first value from the
system path (to undo a pushPATH call).
The pushPATH and popPATH function only modify the system path
for the current R / TERR session.
Using REvaluate
REvaluate directly returns the value from the expression.
REvaluate(version$version.string, LocalEvaluator)
REvaluate(version$language, REvaluator)
REvaluate(version$language, TERREvaluator)
## [1] "TERR"
The data argument to REvaluate is used to pass data to the R engine that
will be called.
The argument can either be a character vector or a list with named components.
p <- 60:67
REvaluate(log2(p), REvaluator, data="p")
To load a package for use by the REvaluate expression you can
either explicilty call library in the expression
or use the packages argument to REvaluate.
rlmFit <- REvaluate({library(MASS); rlm(body_mass_g ~ bill_depth_mm + sex,
peng)}, data = list(peng = palmerpenguins::penguins))
multiREvaluator and RCompare
The multiREvaluator function evaluates an expression in one
or more R interpreters.
The specific R interpreters used are specified as a list of REvaluator
objects in the REvaluators argument.
The default is DefaultREvaluators() which contains an evaluator
for R and for TERR.
The RCompare function uses the multiREvaluator to evaluate
an expression in multiple R interpreters and return the results in a list.
The names of the list are the names of the evaluators used.
The return value also includes the results of running
all.equal comparing the first value in the list with each of the
other values.
The return object is of class "sideBySide" and there is a print
method for the class to display the results.
## RCompare(log2(xTest), data = list(xTest=4:8))
## R version 4.1.2 (2021-11-01) | TERR version 6.0.0 (2021-11-30)
## [1] 2.000000 2.321928 2.584963 | [1] 2.000000 2.321928 2.584963
## [4] 2.807355 3.000000 | [4] 2.807355 3.000000
##
## Attributes:
## $all.equal
## $all.equal$R version 4.1.2 (2021-11-01) vs. TERR version 6.0.0 (2021-11-30)
## [1] TRUE
Here is an example where TERR and open-source R differ.
The seq.Date function in TERR uses the last day of each month if the from argument is
a last day, even if the subsequent months have less days.
Open-source R moves into the next month if the number of days is less than the from argument.
RCompare(seq(as.Date("2000-01-31"), by = "months", length = 12))
## R version 4.1.2 (2021-11-01) | TERR version 6.0.0 (2021-11-30)
## [1] "2000-01-31" "2000-03-02" | [1] "2000-01-31" "2000-02-29"
## [3] "2000-03-31" "2000-05-01" | [3] "2000-03-31" "2000-04-30"
## [5] "2000-05-31" "2000-07-01" | [5] "2000-05-31" "2000-06-30"
## [7] "2000-07-31" "2000-08-31" | [7] "2000-07-31" "2000-08-31"
## [9] "2000-10-01" "2000-10-31" | [9] "2000-09-30" "2000-10-31"
## [11] "2000-12-01" "2000-12-31" | [11] "2000-11-30" "2000-12-31"
##
## Attributes:
## $all.equal
## $all.equal$R version 4.1.2 (2021-11-01) vs. TERR version 6.0.0 (2021-11-3 0)
## [1] "Mean relative difference: 0.0001076079"
There is a sideBySide function that converts any list into
an object of class "sideBySide".
This is a useful way to compare multiple results, even if
they were not computed in different R interpreters.
The .PRINTFUN argument can be used to specify a particular
print method for the display (see the example below).
sideBySide(list(NaiveFactorial=sapply(1:10, function(i) prod(1:i)),
ViaGamma=gamma((1:10)+1)))
# may need options(width=124)
sideBySide(list(Default=terms(y ~ 1), NoIntercept=terms(y ~ -1)),
PRINTFUN="str")
TIBCOSoftware/RinR documentation built on March 19, 2022, 5:16 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
The RinR package provides functions for running code in one R interpreter
from within another R interpreter.
The R interpreters can be different versions of open-source R
or the TERR (TIBCO Enterprise Runtime for R) interpreter.
The RCompare function in the package can run an expression in two different
R interpreters and compare the results.
The RGraph function can be used to generate graphics in TERR
by calling graphics functions in an open-source R interpreter.
The RinR package is included with TERR 2.0 (and later versions). The package can be downloaded from the TERR community site for use with R. The package should also soon be available on CRAN.
Along with the RinR package, this vignette uses the MASS package that is
included with open-source R.
The minimax and devtools packages from CRAN are also used.
library(RinR)
Evaluators
The main functions of the RinR package are the evaluators.
These functions execute a given expression in the specified R interpreter.
The specified R interpreter is started with these command
line arguments: "--quiet --vanilla --no-echo".
Note that this means that neither the system Rprofile.site nor
the user's .Rprofile (for R) or .TERRprofile for TERR will be run.
The return value from an evaluator is a list of length 1
containing the value of the evaluated expression.
The name of the list component is the version$verstion.string
from the R interpreter used in the evaluation.
The REvaluate function returns just the value of the
evaluated expression in the specified R interpreter.
You can have many evaluator functions, one for each R interpreter that you want to evaluate an expression in.
The makeEvaluator function is used to create an evaluator
for a particular R interpreter.
vs_4.0.5 <- list(`R version 4.0.5 (2021-03-31)` = "R version 4.0.5 (2021-03-31)") vs_4.1.2 <- list(`R version 4.1.2 (2021-11-01)` = "R version 4.1.2 (2021-11-01)")
R4.1.2 <- makeREvaluator("R", RHome="/home/R/R-4.1.2") R4.1.2(version$version.string) ## $`R version 4.1.2 (2021-11-01)` ## [1] "R version 4.1.2 (2021-11-01)" R4.0.5 <- makeREvaluator("R", RHome="/home/R/R-4.0.5") R4.0.5(version$version.string) ## $`R version 4.0.5 (2021-03-31)` ## [1] "R version 4.0.5 (2021-03-31)"
The RinR package comes with several default evaluators.
LocalEvaluator- the current interpreter (R or TERR)REvaluator- an R interpreterTERREvaluator- a TERR interpreter
These evaluators are created with the configureREvaluator function
when the RinR package is loaded.
Finding the R Interpreters
The RinR package must be able to find the various R interpreters that you want the package to evaluate code in.
RinR tries to find the location of the R interpreter based on the values of
options("RinR_R_FULL_PATH"),options("RinR_TERR_FULL_PATH")- values in the Windows registry,
- commonly-used locations in the file system,
- directories in
Sys.getenv("PATH").
If you regularly use the RinR package it is recommended what you set the
options RinR_R_FULL_PATH and/or RinR_TERR_FULL_PATH in your
startup file.
The value of the option should be the full path to the R binary e.g.:
options(RinR_R_FULL_PATH="/usr/local/bin/R-4.0.5") options(RinR_TERR_FULL_PATH="/usr/local/bin/TERR-6.0")
This will ensure that you always get the specified R interpreter when using
the default REvaluator and TERREvaluator functions.
If a default evaluators could not be configured when the
package is loaded,
the evaluator will look in the system path (Sys.getenv("PATH"))
for the appropriate interpreter when the evaluator function is called.
The pushPath function can add the path to an R interpreter to the system path e.g.
pushPATH("/home/R/R-4.0.5/bin")
The popPATH function can be used to remove the first value from the
system path (to undo a pushPATH call).
The pushPATH and popPATH function only modify the system path
for the current R / TERR session.
Using REvaluate
REvaluate directly returns the value from the expression.
REvaluate(version$version.string, LocalEvaluator)
REvaluate(version$language, REvaluator)
REvaluate(version$language, TERREvaluator) ## [1] "TERR"
The data argument to REvaluate is used to pass data to the R engine that
will be called.
The argument can either be a character vector or a list with named components.
p <- 60:67 REvaluate(log2(p), REvaluator, data="p")
To load a package for use by the REvaluate expression you can
either explicilty call library in the expression
or use the packages argument to REvaluate.
rlmFit <- REvaluate({library(MASS); rlm(body_mass_g ~ bill_depth_mm + sex, peng)}, data = list(peng = palmerpenguins::penguins))
multiREvaluator and RCompare
The multiREvaluator function evaluates an expression in one
or more R interpreters.
The specific R interpreters used are specified as a list of REvaluator
objects in the REvaluators argument.
The default is DefaultREvaluators() which contains an evaluator
for R and for TERR.
The RCompare function uses the multiREvaluator to evaluate
an expression in multiple R interpreters and return the results in a list.
The names of the list are the names of the evaluators used.
The return value also includes the results of running
all.equal comparing the first value in the list with each of the
other values.
The return object is of class "sideBySide" and there is a print
method for the class to display the results.
## RCompare(log2(xTest), data = list(xTest=4:8)) ## R version 4.1.2 (2021-11-01) | TERR version 6.0.0 (2021-11-30) ## [1] 2.000000 2.321928 2.584963 | [1] 2.000000 2.321928 2.584963 ## [4] 2.807355 3.000000 | [4] 2.807355 3.000000 ## ## Attributes: ## $all.equal ## $all.equal$R version 4.1.2 (2021-11-01) vs. TERR version 6.0.0 (2021-11-30) ## [1] TRUE
Here is an example where TERR and open-source R differ.
The seq.Date function in TERR uses the last day of each month if the from argument is
a last day, even if the subsequent months have less days.
Open-source R moves into the next month if the number of days is less than the from argument.
RCompare(seq(as.Date("2000-01-31"), by = "months", length = 12)) ## R version 4.1.2 (2021-11-01) | TERR version 6.0.0 (2021-11-30) ## [1] "2000-01-31" "2000-03-02" | [1] "2000-01-31" "2000-02-29" ## [3] "2000-03-31" "2000-05-01" | [3] "2000-03-31" "2000-04-30" ## [5] "2000-05-31" "2000-07-01" | [5] "2000-05-31" "2000-06-30" ## [7] "2000-07-31" "2000-08-31" | [7] "2000-07-31" "2000-08-31" ## [9] "2000-10-01" "2000-10-31" | [9] "2000-09-30" "2000-10-31" ## [11] "2000-12-01" "2000-12-31" | [11] "2000-11-30" "2000-12-31" ## ## Attributes: ## $all.equal ## $all.equal$R version 4.1.2 (2021-11-01) vs. TERR version 6.0.0 (2021-11-3 0) ## [1] "Mean relative difference: 0.0001076079"
There is a sideBySide function that converts any list into
an object of class "sideBySide".
This is a useful way to compare multiple results, even if
they were not computed in different R interpreters.
The .PRINTFUN argument can be used to specify a particular
print method for the display (see the example below).
sideBySide(list(NaiveFactorial=sapply(1:10, function(i) prod(1:i)), ViaGamma=gamma((1:10)+1))) # may need options(width=124) sideBySide(list(Default=terms(y ~ 1), NoIntercept=terms(y ~ -1)), PRINTFUN="str")
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.