I am a package developer and want to state the minimum R version required to use my package in the DESCRIPTION
file.
我是一个软件包开发人员,想要说明在DESCRIPTION文件中使用我的软件包所需的最低R版本。
available.packages
parses DESCRIPTIONS of packages, but the result is not (easily) machine readable to lookup recursive dependencies, since the Imports and Depends fields are comma separated text, and sometimes contain packages have version requirements.
available.packages解析包的描述,但结果不是(容易)机器可读的查找递归依赖,因为Imports和Depends字段是逗号分隔文本,有时包含包有版本要求。
The solution described in: Listing R Package Dependencies Without Installing Packages is not a recursive solution. If a nested dependency needed R > 3.3, I want to know about it.
解决方案中描述的解决方案:列表R包依赖性而不安装包不是递归解决方案。如果嵌套依赖关系需要R> 3.3,我想知道它。
At a minimum, I would like to see the minimum version of R and imported, linked, and depends packages for a given CRAN package. Better still would be to list the package or packages which set the minimum R or package version.
至少,我希望看到给定CRAN包的R和导入,链接和依赖包的最低版本。更好的是列出设置最小R或包版本的包或包。
By eliminating dependencies which have higher version requirements, I can serve more people with institutionally old R versions they can't fix: some are still on R 2.x.
通过消除具有更高版本要求的依赖关系,我可以为更多人提供他们无法解决的系统旧R版本:一些仍然在R 2.x.
2 个解决方案
#1
4
min_r_version <- function(package="ggplot2", exclude_main_pkg=TRUE) {
purrr::walk(c("tools", "purrr", "devtools", "stringi", "tidyr", "dplyr"),
require, character.only=TRUE)
deps <- package_dependencies(package, recursive=TRUE)
if (exclude_main_pkg) {
pkgs <- deps[[1]]
} else {
pkgs <- c(package, deps[[1]])
}
available.packages() %>%
as_data_frame() %>%
filter(Package %in% pkgs) %>%
select(Depends) %>%
unlist() -> pkg_list
# if main pkg only relied on core R packages (i.e. pkgs that aren't in CRAN) and we
# excluded the pkg itself from the min version calculation, this is an edge case we need
# to handle.
if (length(pkg_list) == 0) return("Unspecified")
stri_split_regex(pkg_list, "[,]") %>%
unlist() %>%
trimws() %>%
stri_match_all_regex(c("^R$|^R \\(.*\\)$")) %>%
unlist() %>%
discard(is.na(.)) %>%
unique() %>%
stri_replace_all_regex("[R >=\\(\\)]", "") %>%
data_frame(vs=.) %>%
separate(vs, c("a", "b", "c"), fill="right") %>%
mutate(c=ifelse(is.na(c), 0, c)) %>%
arrange(a, b, c) %>%
tail(1) %>%
unite(min, a:c, sep=".") -> vs
return(vs$min)
}
# did we handle the edge cases well enought?
base <- c("base", "compiler", "datasets", "grDevices", "graphics", "grid", "methods", "parallel", "profile", "splines", "stats", "stats4", "tcltk", "tools", "translations")
(base_reqs <- purrr::map_chr(base, min_r_version))
## [1] "Unspecified" "Unspecified" "Unspecified" "Unspecified" "Unspecified"
## [6] "Unspecified" "Unspecified" "Unspecified" "Unspecified" "Unspecified"
## [11] "Unspecified" "Unspecified" "Unspecified" "Unspecified" "Unspecified"
# a few of the "core" contributed pkgs rely on a pkg or two outside of base
# but many only rely on base packages, to this is another gd edge case to
# text for.
contrib <- c("KernSmooth", "MASS", "Matrix", "boot", "class", "cluster", "codetools", "foreign", "lattice", "mgcv", "nlme", "nnet", "rpart", "spatial", "survival")
contrib_reqs <- purrr::map_chr(contrib, min_r_version)
## [1] "Unspecified" "Unspecified" "3.0.0" "Unspecified" "3.1.0"
## [6] "Unspecified" "Unspecified" "Unspecified" "Unspecified" "3.0.2"
## [11] "3.0.0" "Unspecified" "Unspecified" "Unspecified" "3.0.1"
# See what the min version of R shld be for some of my pkgs
min_r_version("ggalt") # I claim R (>= 3.0.0) in DESCRIPTION
## [1] "3.1.2"
min_r_version("curlconverter") # I claim R (>= 3.0.0) in DESCRIPTION
## [1] "3.1.2"
min_r_version("iptools") # I claim R (>= 3.0.0) in DESCRIPTION
## [1] "3.0.0"
#2
0
Based on ideas from @hrbrmstr and written with base functions, I'm now using the following function:
基于@hrbrmstr的想法并用基本函数编写,我现在使用以下函数:
min_r_version <- function(pkg) {
requireNamespace("tools")
requireNamespace("utils")
avail <- utils::available.packages(utils::contrib.url(repo))
deps <- tools::package_dependencies(pkg, db = avail, recursive = TRUE)
if (is.null(deps))
stop("package not found")
pkgs <- deps[[1]]
repo = getOption("repo")
if (is.null(repo))
repo <- "https://cloud.r-project.org"
matches <- avail[ , "Package"] %in% pkgs
pkg_list <- avail[matches, "Depends"]
vers <- grep("^R$|^R \\(.*\\)$", pkg_list, value = TRUE)
vers <- gsub("[^0-9.]", "", vers)
if (length(vers) == 0)
return("Not specified")
max_ver = vers[1]
if (length(vers) == 1)
return(max_ver)
for (v in 2:length(vers))
if (utils::compareVersion(vers[v], max_ver) > 0)
max_ver <- vers[v]
max_ver
}
#1
4
min_r_version <- function(package="ggplot2", exclude_main_pkg=TRUE) {
purrr::walk(c("tools", "purrr", "devtools", "stringi", "tidyr", "dplyr"),
require, character.only=TRUE)
deps <- package_dependencies(package, recursive=TRUE)
if (exclude_main_pkg) {
pkgs <- deps[[1]]
} else {
pkgs <- c(package, deps[[1]])
}
available.packages() %>%
as_data_frame() %>%
filter(Package %in% pkgs) %>%
select(Depends) %>%
unlist() -> pkg_list
# if main pkg only relied on core R packages (i.e. pkgs that aren't in CRAN) and we
# excluded the pkg itself from the min version calculation, this is an edge case we need
# to handle.
if (length(pkg_list) == 0) return("Unspecified")
stri_split_regex(pkg_list, "[,]") %>%
unlist() %>%
trimws() %>%
stri_match_all_regex(c("^R$|^R \\(.*\\)$")) %>%
unlist() %>%
discard(is.na(.)) %>%
unique() %>%
stri_replace_all_regex("[R >=\\(\\)]", "") %>%
data_frame(vs=.) %>%
separate(vs, c("a", "b", "c"), fill="right") %>%
mutate(c=ifelse(is.na(c), 0, c)) %>%
arrange(a, b, c) %>%
tail(1) %>%
unite(min, a:c, sep=".") -> vs
return(vs$min)
}
# did we handle the edge cases well enought?
base <- c("base", "compiler", "datasets", "grDevices", "graphics", "grid", "methods", "parallel", "profile", "splines", "stats", "stats4", "tcltk", "tools", "translations")
(base_reqs <- purrr::map_chr(base, min_r_version))
## [1] "Unspecified" "Unspecified" "Unspecified" "Unspecified" "Unspecified"
## [6] "Unspecified" "Unspecified" "Unspecified" "Unspecified" "Unspecified"
## [11] "Unspecified" "Unspecified" "Unspecified" "Unspecified" "Unspecified"
# a few of the "core" contributed pkgs rely on a pkg or two outside of base
# but many only rely on base packages, to this is another gd edge case to
# text for.
contrib <- c("KernSmooth", "MASS", "Matrix", "boot", "class", "cluster", "codetools", "foreign", "lattice", "mgcv", "nlme", "nnet", "rpart", "spatial", "survival")
contrib_reqs <- purrr::map_chr(contrib, min_r_version)
## [1] "Unspecified" "Unspecified" "3.0.0" "Unspecified" "3.1.0"
## [6] "Unspecified" "Unspecified" "Unspecified" "Unspecified" "3.0.2"
## [11] "3.0.0" "Unspecified" "Unspecified" "Unspecified" "3.0.1"
# See what the min version of R shld be for some of my pkgs
min_r_version("ggalt") # I claim R (>= 3.0.0) in DESCRIPTION
## [1] "3.1.2"
min_r_version("curlconverter") # I claim R (>= 3.0.0) in DESCRIPTION
## [1] "3.1.2"
min_r_version("iptools") # I claim R (>= 3.0.0) in DESCRIPTION
## [1] "3.0.0"
#2
0
Based on ideas from @hrbrmstr and written with base functions, I'm now using the following function:
基于@hrbrmstr的想法并用基本函数编写,我现在使用以下函数:
min_r_version <- function(pkg) {
requireNamespace("tools")
requireNamespace("utils")
avail <- utils::available.packages(utils::contrib.url(repo))
deps <- tools::package_dependencies(pkg, db = avail, recursive = TRUE)
if (is.null(deps))
stop("package not found")
pkgs <- deps[[1]]
repo = getOption("repo")
if (is.null(repo))
repo <- "https://cloud.r-project.org"
matches <- avail[ , "Package"] %in% pkgs
pkg_list <- avail[matches, "Depends"]
vers <- grep("^R$|^R \\(.*\\)$", pkg_list, value = TRUE)
vers <- gsub("[^0-9.]", "", vers)
if (length(vers) == 0)
return("Not specified")
max_ver = vers[1]
if (length(vers) == 1)
return(max_ver)
for (v in 2:length(vers))
if (utils::compareVersion(vers[v], max_ver) > 0)
max_ver <- vers[v]
max_ver
}