Figures drawn with plot() break code chunks into pieces #17
Labels
bug
Something isn't working
Comments
|
It's not intended and sounds like a bug. I'll investigate. Thanks for the report! |
|
I can't reproduce the problem on macOS or Windows. Are you using Linux? |
|
I just tested it on Ubuntu and couldn't reproduce the problem, either. Please provide |
|
I may have found the cause of the problem. It seems to happen only when litedown::reactor(dev = svglite::svglite)
Rmd = '
```{r}
#### Data ####
set.seed(2023)
subj_idx = 1:30
gender = rep(1:2, each=15)
heights = c(160, 175)[gender] + rnorm(30, sd=15)
#### Plot ####
ylim = c( min(heights)-5, max(heights)+5 )
plot( 1, type="n", xlim=c(.5,30.5), ylim=ylim,
xlab="Subject Index", ylab="Height (cm)" )
points( 1:15, heights[ 1:15], col=2, pch=19 )
points( 16:30, heights[16:30], col=4, pch=19 )
abline( v=15.5, col="grey", lty="dashed" )
mtext( c("Girls","Boys"), at=c(7,23), col=c(2,4), padj=-.5 )
```
'
litedown::fuse(Rmd, output="index.md")Output: ``` {.r}
#### Data ####
set.seed(2023)
subj_idx = 1:30
gender = rep(1:2, each=15)
heights = c(160, 175)[gender] + rnorm(30, sd=15)
#### Plot ####
ylim = c( min(heights)-5, max(heights)+5 )
plot( 1, type="n", xlim=c(.5,30.5), ylim=ylim,
xlab="Subject Index", ylab="Height (cm)" )
```
``` {.r}
points( 1:15, heights[ 1:15], col=2, pch=19 )
points( 16:30, heights[16:30], col=4, pch=19 )
abline( v=15.5, col="grey", lty="dashed" )
mtext( c("Girls","Boys"), at=c(7,23), col=c(2,4), padj=-.5 )
```Changing litedown::reactor(dev = svg)
Rmd = '
```{r}
#### Data ####
set.seed(2023)
subj_idx = 1:30
gender = rep(1:2, each=15)
heights = c(160, 175)[gender] + rnorm(30, sd=15)
#### Plot ####
ylim = c( min(heights)-5, max(heights)+5 )
plot( 1, type="n", xlim=c(.5,30.5), ylim=ylim,
xlab="Subject Index", ylab="Height (cm)" )
points( 1:15, heights[ 1:15], col=2, pch=19 )
points( 16:30, heights[16:30], col=4, pch=19 )
abline( v=15.5, col="grey", lty="dashed" )
mtext( c("Girls","Boys"), at=c(7,23), col=c(2,4), padj=-.5 )
```
'
litedown::fuse(Rmd, output="index.md")Output: ``` {.r}
#### Data ####
set.seed(2023)
subj_idx = 1:30
gender = rep(1:2, each=15)
heights = c(160, 175)[gender] + rnorm(30, sd=15)
#### Plot ####
ylim = c( min(heights)-5, max(heights)+5 )
plot( 1, type="n", xlim=c(.5,30.5), ylim=ylim,
xlab="Subject Index", ylab="Height (cm)" )
points( 1:15, heights[ 1:15], col=2, pch=19 )
points( 16:30, heights[16:30], col=4, pch=19 )
abline( v=15.5, col="grey", lty="dashed" )
mtext( c("Girls","Boys"), at=c(7,23), col=c(2,4), padj=-.5 )
```
xfun::session_info('litedown')
#> R version 4.2.1 (2022-06-23)
#> Platform: aarch64-apple-darwin20 (64-bit)
#> Running under: macOS 14.3.1, RStudio c(2023, 6, 2, 561)
#>
#> Locale: en_US.UTF-8 / en_US.UTF-8 / en_US.UTF-8 / C / en_US.UTF-8 / en_US.UTF-8
#>
#> Package version:
#> commonmark_1.9.1 graphics_4.2.1 grDevices_4.2.1 litedown_0.0.30 stats_4.2.1
#> tools_4.2.1 utils_4.2.1 xfun_0.46 |
|
It's a bug of record('plot(1)\npoints(1)') # default png; okay
record('plot(1)\npoints(1)', dev = svglite::svglite) # plot in wrong placeI understand what's going on now and will fix it. |
|
Fixed in xfun: https://github.com/yihui/xfun litedown::fuse('```{r, dev=svglite::svglite}
plot(1)
points(1)
```', output = 'markdown')``` {.r}
plot(1)
```
``` {.r}
points(1)
``` |
clrpackages
pushed a commit
to clearlinux-pkgs/R-xfun
that referenced
this issue
Aug 20, 2024
Yihui Xie (25):
add functions lazy_save() and lazy_load()
provide the command directly in the message
amend 02973d25c413f27dda43534dde0f279342ac3d23: coerce the version into character, otherwise c() will concatenate version numbers as a list
fix yihui/litedown#17: svglite will create a plot file upon the call to a high-level plot function, and modify this file later if low-level plot functions are called later
correct issue number
no need to use on.exit(after = FALSE)
support lower versions of R
typo
add endsWidth()
add trimws()
add file.size()
typo
move to zzz.R
add anyNA()
remove functions that are available in R 3.2.0
it seems 9000 is still too big
make the batch size an argument of cloud_check()
separate status and report
rename get_fun to call_fun and call the object directly if it's a function
`...` is not actually used
get rid of the assign() hack and use the job name directly
suppress all messages by default, and only report status and report
show submission progress
no need to show the call
CRAAN release v0.47
netbsd-srcmastr
pushed a commit
to NetBSD/pkgsrc
that referenced
this issue
Oct 26, 2024
# CHANGES IN xfun VERSION 0.48 - Added utilities for HTML tags: `html_tag()`, `html_escape()`, `html_escape()`, and `html_view()`. Removed the soft dependency on the **htmltools** package accordingly. - `base_pkgs()` is faster now: it calls `tools::standard_package_names()` if the function exists (R >= 4.4.0), otherwise it just returns a constant vector of base package names (thanks, @arnaudgallou, #91). - Added a function `mime_type()` to obtain the MIME types of files via `mime::guess_type()` if **mime** is installed, otherwise it will call `tools:::mime_type()`, and fall back to using a system command (e.g., `file --mime-type`) to obtain the types. - Added a function `file_rename()` to deal with `file.rename()` failures by calling `file.copy()` (thanks, @Giqles @katrinabrock, rstudio/bookdown#804). - `new_app()` will use `utils::browseURL()` to open the app if `options('viewer')` is not configured (thanks, @AlbertLei, yihui/litedown#29). - Added a method `record_print.record_asis()` to return the object as is. # CHANGES IN xfun VERSION 0.47 - Added functions `lazy_save()` and `lazy_load()` to save objects to files and lazy-load them. - Fixed a bug in `record(dev = svglite::svglite)` that misplaced plots when low-level plot functions are used (thanks, @liao961120, yihui/litedown#17). - Specified the lowest R version required (v3.2.0) for this package. # CHANGES IN xfun VERSION 0.46 - `md_table()` should add a vertical ellipsis to row names when rows are truncated by the `limit` argument. - `session_info()` recognizes Positron now (thanks, @chuxinyuan, #89). # CHANGES IN xfun VERSION 0.45 - For `record()` with `verbose = 1` or `2`, invisible `NULL` is no longer printed. - `Rscript_call()` will show the actual error message (if an error occurred) during calling the function in a new R session. # CHANGES IN xfun VERSION 0.44 - Added a function `cache_exec()` to cache the execution of an expression either in memory or on disk. It is much more general and flexible than `cache_rds()`. For example, it supports custom reading/writing methods for cache files, and can load locally created variables in the expression while loading cache. - Added an argument `cache` to `record()` to make it possible to enable caching. - Added arguments `message` and `warning` to `record()` to decide whether messages and warnings should be recorded. - Changed the default value of the argument `error` of `record()` from `FALSE` to `NA`. Now `FALSE` means to suppress error messages, and `NA` means to throw errors normally. This is for consistency with the `message` and `warning` arguments. - Added an S3 generic function `record_print()`, which is similar to `knitr::knit_print()` but for the purpose of printing visible values in `record()`. - The `record()` function gained new arguments `print` and `print.args` to support custom printing functions and arguments. - Added a function `md_table()`, which is a minimal Markdown table generator. - Exported the internal function `md5()` to calculate the MD5 checksums of R objects. The function is essentially a workaround for `tools::md5sum()` (see HenrikBengtsson/Wishlist-for-R#21). - For `fenced_block()`, a space is added between the backticks and the language name, e.g., ```` ```r ```` has become ```` ``` r ```` now. This will affect snapshot tests based on Markdown ([an example](yihui/knitr-examples@931e0a2)). - Added a shorthand `fenced_div()` for `fenced_block(char = ':')`. - `write_utf8()` returns the `con` argument (typically a file path) now. Previously, it returns `NULL`. - Added an experimental function `new_app()` to create a local web application. - The returned value of `yaml_body()` contains a new element `lines` in the list indicating the line numbers of YAML metadata if exists. - Removed the `skip` argument from `split_source()`. - For `split_source(line_number = TRUE)`, the attribute name for line numbers in the returned value was changed from `line_start` (a single starting line number) to `lines` (both the starting and ending numbers). - Fixed an edge case in `prose_index()`, in which inline code was incorrectly recognized as a code block fence. # CHANGES IN xfun VERSION 0.43 - Added a function `upload_imgur()`, which was adapted from `knitr::imgur_upload()`. The latter will call the former in the future. `xfun::upload_imgur()` allows users to choose whether to use the system command `curl` or the R package **curl** to upload the image. It also has a new argument `include_xml` to specify whether the XML response needs to be included in the returned value. - Added a function `fenced_block()` to create a fenced block in Markdown (thanks, @cderv, yihui/knitr#2331). The block can be either a code block or a fenced Div. - Fixed a bug in `xfun::record()` when the argument `verbose = 1` or `2`.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I find this counter-intuitive. Not sure if it's intended?
Here's an example:
Code in
main.Rmd:Output in Markdown
Output in HTML
Huge thanks for this awesome work!
The text was updated successfully, but these errors were encountered: