mirror of https://git.sr.ht/~hrbrmstr/splashr
16 changed files with 451 additions and 10 deletions
@ -0,0 +1,200 @@ |
|||
make_splash_call <- function(splash_obj) { |
|||
|
|||
sprintf(' |
|||
function main(splash) |
|||
%s |
|||
end |
|||
', paste0(sprintf(" %s", splash_obj$calls), collapse="\n")) -> out |
|||
|
|||
out |
|||
|
|||
} |
|||
|
|||
#' Enable or disable response content tracking. |
|||
#' |
|||
#' By default Splash doesn’t keep bodies of each response in memory, for efficiency reasons. |
|||
#' |
|||
#' @param splash_obj splashr object |
|||
#' @param enable logical |
|||
#' @export |
|||
#' @examples \dontrun{ |
|||
#' splash_local %>% |
|||
#' splash_response_body(TRUE) %>% |
|||
#' splash_go("https://rud.is/b") %>% |
|||
#' splash_wait(2) %>% |
|||
#' splash_har() -> rud_har |
|||
#' } |
|||
splash_response_body <- function(splash_obj, enable=FALSE) { |
|||
splash_obj$calls <- c(splash_obj$calls, sprintf('splash.response_body_enabled = %s', |
|||
if (enable) "true" else "false")) |
|||
splash_obj |
|||
} |
|||
|
|||
#' Enable or disable browser plugins (e.g. Flash). |
|||
#' |
|||
#' Plugins are disabled by default. |
|||
#' |
|||
#' @param splash_obj splashr object |
|||
#' @param enable logical |
|||
#' @export |
|||
#' @examples \dontrun{ |
|||
#' splash_local %>% |
|||
#' splash_plugins(TRUE) %>% |
|||
#' splash_go("https://rud.is/b") %>% |
|||
#' splash_wait(2) %>% |
|||
#' splash_har() -> rud_har |
|||
#' } |
|||
splash_plugins <- function(splash_obj, enable=FALSE) { |
|||
splash_obj$calls <- c(splash_obj$calls, sprintf('splash.plugins_enabled = %s', |
|||
if (enable) "true" else "false")) |
|||
splash_obj |
|||
} |
|||
|
|||
#' Enable/disable images |
|||
#' |
|||
#' By default, images are enabled. Disabling of the images can save a lot of network |
|||
#' traffic (usually around ~50%) and make rendering faster. Note that this option can |
|||
#' affect the JavaScript code inside page: disabling of the images may change sizes and |
|||
#' positions of DOM elements, and scripts may read and use them. |
|||
#' |
|||
#' @param splash_obj splashr object |
|||
#' @param enable logical |
|||
#' @export |
|||
#' @examples \dontrun{ |
|||
#' splash_local %>% |
|||
#' splash_images(TRUE) %>% |
|||
#' splash_go("https://rud.is/b") %>% |
|||
#' splash_wait(2) %>% |
|||
#' splash_har() -> rud_har |
|||
#' } |
|||
splash_images <- function(splash_obj, enable=TRUE) { |
|||
splash_obj$calls <- c(splash_obj$calls, sprintf('splash.images_enabled = %s', |
|||
if (enable) "true" else "false")) |
|||
splash_obj |
|||
} |
|||
|
|||
#' Go to an URL. |
|||
#' |
|||
#' This is similar to entering an URL in a browser address bar, pressing Enter and waiting |
|||
#' until page loads. |
|||
#' |
|||
#' @param splash_obj splashr object |
|||
#' @param url - URL to load; |
|||
#' @export |
|||
#' @examples \dontrun{ |
|||
#' splash_local %>% |
|||
#' splash_response_body(TRUE) %>% |
|||
#' splash_go("https://rud.is/b") %>% |
|||
#' splash_wait(2) %>% |
|||
#' splash_har() -> rud_har |
|||
#' } |
|||
splash_go <- function(splash_obj, url) { |
|||
splash_obj$calls <- c(splash_obj$calls, |
|||
sprintf('url = "%s"', url), |
|||
"splash:go(url)") |
|||
splash_obj |
|||
} |
|||
|
|||
#' Wait for a period time |
|||
#' |
|||
#' When script is waiting WebKit continues processing the webpage |
|||
#' |
|||
#' @md |
|||
#' @param splash_obj splashr object |
|||
#' @param time number of seconds to wait |
|||
#' @export |
|||
#' @examples \dontrun{ |
|||
#' splash_local %>% |
|||
#' splash_response_body(TRUE) %>% |
|||
#' splash_go("https://rud.is/b") %>% |
|||
#' splash_wait(2) %>% |
|||
#' splash_har() -> rud_har |
|||
#' } |
|||
splash_wait <- function(splash_obj, time=2) { |
|||
splash_obj$calls <- c(splash_obj$calls, sprintf('splash:wait(%s)', time)) |
|||
splash_obj |
|||
} |
|||
|
|||
#' Return information about Splash interaction with a website in HAR format. |
|||
#' |
|||
#' Similar to [render_har] but used in a script context. Should be the LAST element in |
|||
#' a DSL script chain as this will execute the script and return the HAR content |
|||
#' |
|||
#' @md |
|||
#' @param splash_obj splashr object |
|||
#' @export |
|||
#' @examples \dontrun{ |
|||
#' splash_local %>% |
|||
#' splash_response_body(TRUE) %>% |
|||
#' splash_go("https://rud.is/b") %>% |
|||
#' splash_wait(2) %>% |
|||
#' splash_har() -> rud_har |
|||
#' } |
|||
splash_har <- function(splash_obj) { |
|||
|
|||
splash_obj$calls <- c(splash_obj$calls, 'return(splash:har())') |
|||
|
|||
call_function <- make_splash_call(splash_obj) |
|||
|
|||
res <- execute_lua(splash_obj, call_function) |
|||
as_har(res) |
|||
|
|||
} |
|||
|
|||
#' Return a HTML snapshot of a current page. |
|||
#' |
|||
#' Similar to [render_html] but used in a script context. Should be the LAST element in |
|||
#' a DSL script chain as this will execute the script and return the HTML content |
|||
#' |
|||
#' @md |
|||
#' @param splash_obj splashr object |
|||
#' @param raw_html if `TRUE` then return a character vector vs an XML document. |
|||
#' @export |
|||
#' @examples \dontrun{ |
|||
#' splash_local %>% |
|||
#' splash_response_body(TRUE) %>% |
|||
#' splash_go("https://rud.is/b") %>% |
|||
#' splash_wait(2) %>% |
|||
#' splash_html() -> rud_pg |
|||
#' } |
|||
splash_html <- function(splash_obj, raw_html=FALSE) { |
|||
|
|||
splash_obj$calls <- c(splash_obj$calls, 'return(splash:html())') |
|||
|
|||
call_function <- make_splash_call(splash_obj) |
|||
|
|||
out <- execute_lua(splash_obj, call_function) |
|||
|
|||
if (!raw_html) out <- xml2::read_html(out) |
|||
|
|||
out |
|||
|
|||
} |
|||
|
|||
#' Return a screenshot of a current page in PNG format. |
|||
#' |
|||
#' Similar to [render_png] but used in a script context. Should be the LAST element in |
|||
#' a DSL script chain as this will execute the script and return the PNG content |
|||
#' |
|||
#' @md |
|||
#' @param splash_obj splashr object |
|||
#' @return a [magick] image object |
|||
#' @export |
|||
#' @examples \dontrun{ |
|||
#' splash_local %>% |
|||
#' splash_go("https://rud.is/b") %>% |
|||
#' splash_wait(2) %>% |
|||
#' splash_png() |
|||
#' } |
|||
splash_png <- function(splash_obj) { |
|||
|
|||
splash_obj$calls <- c(splash_obj$calls, 'return splash:png{render_all=true}') |
|||
|
|||
call_function <- make_splash_call(splash_obj) |
|||
|
|||
res <- execute_lua(splash_obj, call_function) |
|||
|
|||
magick::image_read(res) |
|||
|
|||
} |
|||
|
|||
@ -0,0 +1,26 @@ |
|||
% Generated by roxygen2: do not edit by hand |
|||
% Please edit documentation in R/dsl.r |
|||
\name{splash_go} |
|||
\alias{splash_go} |
|||
\title{Go to an URL.} |
|||
\usage{ |
|||
splash_go(splash_obj, url) |
|||
} |
|||
\arguments{ |
|||
\item{splash_obj}{splashr object} |
|||
|
|||
\item{url}{- URL to load;} |
|||
} |
|||
\description{ |
|||
This is similar to entering an URL in a browser address bar, pressing Enter and waiting |
|||
until page loads. |
|||
} |
|||
\examples{ |
|||
\dontrun{ |
|||
splash_local \%>\% |
|||
splash_response_body(TRUE) \%>\% |
|||
splash_go("https://rud.is/b") \%>\% |
|||
splash_wait(2) \%>\% |
|||
splash_har() -> rud_har |
|||
} |
|||
} |
|||
@ -0,0 +1,24 @@ |
|||
% Generated by roxygen2: do not edit by hand |
|||
% Please edit documentation in R/dsl.r |
|||
\name{splash_har} |
|||
\alias{splash_har} |
|||
\title{Return information about Splash interaction with a website in HAR format.} |
|||
\usage{ |
|||
splash_har(splash_obj) |
|||
} |
|||
\arguments{ |
|||
\item{splash_obj}{splashr object} |
|||
} |
|||
\description{ |
|||
Similar to \link{render_har} but used in a script context. Should be the LAST element in |
|||
a DSL script chain as this will execute the script and return the HAR content |
|||
} |
|||
\examples{ |
|||
\dontrun{ |
|||
splash_local \%>\% |
|||
splash_response_body(TRUE) \%>\% |
|||
splash_go("https://rud.is/b") \%>\% |
|||
splash_wait(2) \%>\% |
|||
splash_har() -> rud_har |
|||
} |
|||
} |
|||
@ -0,0 +1,26 @@ |
|||
% Generated by roxygen2: do not edit by hand |
|||
% Please edit documentation in R/dsl.r |
|||
\name{splash_html} |
|||
\alias{splash_html} |
|||
\title{Return a HTML snapshot of a current page.} |
|||
\usage{ |
|||
splash_html(splash_obj, raw_html = FALSE) |
|||
} |
|||
\arguments{ |
|||
\item{splash_obj}{splashr object} |
|||
|
|||
\item{raw_html}{if \code{TRUE} then return a character vector vs an XML document.} |
|||
} |
|||
\description{ |
|||
Similar to \link{render_html} but used in a script context. Should be the LAST element in |
|||
a DSL script chain as this will execute the script and return the HTML content |
|||
} |
|||
\examples{ |
|||
\dontrun{ |
|||
splash_local \%>\% |
|||
splash_response_body(TRUE) \%>\% |
|||
splash_go("https://rud.is/b") \%>\% |
|||
splash_wait(2) \%>\% |
|||
splash_html() -> rud_pg |
|||
} |
|||
} |
|||
@ -0,0 +1,28 @@ |
|||
% Generated by roxygen2: do not edit by hand |
|||
% Please edit documentation in R/dsl.r |
|||
\name{splash_images} |
|||
\alias{splash_images} |
|||
\title{Enable/disable images} |
|||
\usage{ |
|||
splash_images(splash_obj, enable = TRUE) |
|||
} |
|||
\arguments{ |
|||
\item{splash_obj}{splashr object} |
|||
|
|||
\item{enable}{logical} |
|||
} |
|||
\description{ |
|||
By default, images are enabled. Disabling of the images can save a lot of network |
|||
traffic (usually around ~50%) and make rendering faster. Note that this option can |
|||
affect the JavaScript code inside page: disabling of the images may change sizes and |
|||
positions of DOM elements, and scripts may read and use them. |
|||
} |
|||
\examples{ |
|||
\dontrun{ |
|||
splash_local \%>\% |
|||
splash_images(TRUE) \%>\% |
|||
splash_go("https://rud.is/b") \%>\% |
|||
splash_wait(2) \%>\% |
|||
splash_har() -> rud_har |
|||
} |
|||
} |
|||
@ -0,0 +1,25 @@ |
|||
% Generated by roxygen2: do not edit by hand |
|||
% Please edit documentation in R/dsl.r |
|||
\name{splash_plugins} |
|||
\alias{splash_plugins} |
|||
\title{Enable or disable browser plugins (e.g. Flash).} |
|||
\usage{ |
|||
splash_plugins(splash_obj, enable = FALSE) |
|||
} |
|||
\arguments{ |
|||
\item{splash_obj}{splashr object} |
|||
|
|||
\item{enable}{logical} |
|||
} |
|||
\description{ |
|||
Plugins are disabled by default. |
|||
} |
|||
\examples{ |
|||
\dontrun{ |
|||
splash_local \%>\% |
|||
splash_plugins(TRUE) \%>\% |
|||
splash_go("https://rud.is/b") \%>\% |
|||
splash_wait(2) \%>\% |
|||
splash_har() -> rud_har |
|||
} |
|||
} |
|||
@ -0,0 +1,26 @@ |
|||
% Generated by roxygen2: do not edit by hand |
|||
% Please edit documentation in R/dsl.r |
|||
\name{splash_png} |
|||
\alias{splash_png} |
|||
\title{Return a screenshot of a current page in PNG format.} |
|||
\usage{ |
|||
splash_png(splash_obj) |
|||
} |
|||
\arguments{ |
|||
\item{splash_obj}{splashr object} |
|||
} |
|||
\value{ |
|||
a \link{magick} image object |
|||
} |
|||
\description{ |
|||
Similar to \link{render_png} but used in a script context. Should be the LAST element in |
|||
a DSL script chain as this will execute the script and return the PNG content |
|||
} |
|||
\examples{ |
|||
\dontrun{ |
|||
splash_local \%>\% |
|||
splash_go("https://rud.is/b") \%>\% |
|||
splash_wait(2) \%>\% |
|||
splash_png() |
|||
} |
|||
} |
|||
@ -0,0 +1,25 @@ |
|||
% Generated by roxygen2: do not edit by hand |
|||
% Please edit documentation in R/dsl.r |
|||
\name{splash_response_body} |
|||
\alias{splash_response_body} |
|||
\title{Enable or disable response content tracking.} |
|||
\usage{ |
|||
splash_response_body(splash_obj, enable = FALSE) |
|||
} |
|||
\arguments{ |
|||
\item{splash_obj}{splashr object} |
|||
|
|||
\item{enable}{logical} |
|||
} |
|||
\description{ |
|||
By default Splash doesn’t keep bodies of each response in memory, for efficiency reasons. |
|||
} |
|||
\examples{ |
|||
\dontrun{ |
|||
splash_local \%>\% |
|||
splash_response_body(TRUE) \%>\% |
|||
splash_go("https://rud.is/b") \%>\% |
|||
splash_wait(2) \%>\% |
|||
splash_har() -> rud_har |
|||
} |
|||
} |
|||
@ -0,0 +1,25 @@ |
|||
% Generated by roxygen2: do not edit by hand |
|||
% Please edit documentation in R/dsl.r |
|||
\name{splash_wait} |
|||
\alias{splash_wait} |
|||
\title{Wait for a period time} |
|||
\usage{ |
|||
splash_wait(splash_obj, time = 2) |
|||
} |
|||
\arguments{ |
|||
\item{splash_obj}{splashr object} |
|||
|
|||
\item{time}{number of seconds to wait} |
|||
} |
|||
\description{ |
|||
When script is waiting WebKit continues processing the webpage |
|||
} |
|||
\examples{ |
|||
\dontrun{ |
|||
splash_local \%>\% |
|||
splash_response_body(TRUE) \%>\% |
|||
splash_go("https://rud.is/b") \%>\% |
|||
splash_wait(2) \%>\% |
|||
splash_har() -> rud_har |
|||
} |
|||
} |
|||
Loading…
Reference in new issue