This example is about using the MATLAB bode() function.
The approach I've worked up here is to write an m-script in an R code chunk, have MATLAB run the code and save the outputs, and read the outputs into the R Markdown environment to create the course document.
I use relative file paths with respect to the RStudio Project working directory. My examples require a file structure that includes:
project\
|-- derived\
|-- results\
|-- scripts\
`-- project.Rproj
I open an Rmd script and save it in the scripts\ directory.
I set the knitr working directory to match the project working directory.
library(knitr)
opts_knit$set(root.dir = "../")Load the packages.
library(readr)
library(stringr)
library(reach)
library(R.matlab)The string variable m_script (below) contains the MATLAB commands to add the project folders to the MATLAB search path.
m_script <- "pathstr = cd;
addpath(genpath(pathstr), '-end');
savepath;"
reach::runMatlabCommand(m_script)The header for this code chunk includes the knitr option eval = FALSE. If MATLAB indicates a path problem, e.g., Warning: Name is nonexistent or not directory..., then I manually run the code chunk.
However, this code chunk adds all directories to the search path, including hidden directories such as .git/ and .Rproj.user/. In later scripts, I show how to set the MATLAB path while excluding the hidden folders (see functions.R).
This R code chunk contains the MATLAB script, surrounded by a single set of quotes and assigned to object m_script.
m_script <- "% assign parameters
K = 1;
wn = 10;
z = 0.05;
% create the transfer function
n = K;
d = [1/wn^2 2*z/wn 1];
sys = tf(n, d);
% compute and plot the frequency response
bode(sys)
grid
"# end m-fileTo run the script in MATLAB:
reach::runMatlabCommand(m_script)
Sys.sleep(0)Running this script, a MATLAB command window opens, the graph opens and closes, and the command window closes. The Sys.sleep(0) function has no effect here (set to 0 seconds) but will be used later.
Next we add some housekeeping code to be able to save the MATLAB output.
Add new lines to the m_script string following the bode(sys) line, starting with comment % break.
- Lines above the break comment will be printed in the output document (code I want students to see and use); lines below are not.
- the
% write sys to txtchunk saves the transfer functionsysto'results/sys01.txt' - the
% write gcf to pngchunk saves the graph to'results/m01_bode.png'
# MATLAB commands for a bode plot
m_script <- "% assign parameters
K = 1;
wn = 10;
z = 0.05;
% create the transfer function
n = K;
d = [1/wn^2 2*z/wn 1];
sys = tf(n, d);
% compute and plot the frequency response
bode(sys)
grid
% print_stop
% write sys to txt
fid = fopen('results/sys01.txt', 'wt');
sys_string = evalc('sys');
fprintf(fid, sys_string);
fclose(fid);
% write gcf to png
fig = gcf;
fig.PaperUnits = 'inches';
fig.PaperPosition = [0 0 4 4];
saveas(fig, 'results/m01_bode.png');
"# end m-fileAfter knitting the document, the results directory has two files.
results\
|-- sys01.txt
`-- m01_bode.png
I use str_split() from the stringr package to separate the MATLAB commands at the % print_stop comment and print the upper half of the code to the output document for student use.
if (str_detect(m_script, "% print_stop")) {
code_for_students <- str_split(m_script, "% print_stop", n = 2)[[1]][1]
} else {
code_for_students <- m_script
}
cat(code_for_students)The output printed to the document is:
% assign parameters
K = 1;
wn = 10;
z = 0.05;
% create the transfer function
n = K;
d = [1/wn^2 2*z/wn 1];
sys = tf(n, d);
% compute and plot the frequency response
bode(sys)
grid
Before reading the files from results\ to place them in the Rmd output document, I edit Sys.sleep() to pause R execution for 12 seconds (your mileage may vary) to give MATLAB time to write outputs to file before R reads those files.
The revised code chunk for running MATLAB:
reach::runMatlabCommand(m_script)
Sys.sleep(12)Read the MATLAB transfer function expression with read_lines() from the readr package. Select specific lines and print with cat().
sys <- read_lines('results/sys01.txt', skip = 3, n_max = 3)
cat(sys, sep = "\n") 1
---------------------
0.01 s^2 + 0.01 s + 1
Or I can get fancy and pretty-print the transfer function,
# trim spaces and isolate the numerator and denominator
Gs <- str_trim(sys)
num <- Gs[1]
den <- Gs[3]using inline code chunks in a math expression,
$$
G(s) = \frac{`r num`}{`r den`}
$$Sadly, this syntax does not render properly when the R Markdown output format is set to github_document. However, it does render correctly to HTML, PDF, and Word output formats.
The frequency response graph is obtained by reading the PNG file. Note that because include_graphics() does not honor the root.dir = "../" option set earlier, the path to the file name has to go up one level before finding the results directory.
knitr::include_graphics('../results/m01_bode.png')
Executing the MATLAB script is necessary only if the m-script changes. I can save run time by checking if the m-script has changed since the last time it was run.
This code chunk runs the MATLAB script for the first pass or if the MATLAB code has been changed. This chunk replaces the earlier lines runMatlabCommand(m_script) and Sys.sleep(12). I still need the pause, but only when the m-file has been changed.
The Sys.sleep(0) function allows R to temporarily be given very low priority and hence not to interfere with more important foreground tasks. A typical use is to allow a process launched from R to set itself up and read its input files before R execution is resumed.
# save current version of m-script
saveRDS(m_script, "derived/m01-new.rds")
# create an empty RDS on the first pass
if (!file.exists("derived/m01-old.rds")) {
saveRDS(" ", "derived/m01-old.rds")
}
# run MATLAB if old and new differ
newrds <- readRDS("derived/m01-new.rds")
oldrds <- readRDS("derived/m01-old.rds")
if (!identical(oldrds, newrds)) {
saveRDS(m_script, "derived/m01-old.rds")
runMatlabCommand(m_script)
Sys.sleep(12)
}- Several of the housekeeping tasks in this script have since been made into functions and saved in
functions.R. - These functions are used in the revised Bode plot tutorial.
setting value
version R version 3.3.1 (2016-06-21)
system x86_64, mingw32
ui RStudio (0.99.902)
language (EN)
collate English_United States.1252
tz America/New_York
date 2016-11-06
package * version date source
assertthat 0.1 2013-12-06 CRAN (R 3.2.1)
devtools * 1.12.0 2016-06-24 CRAN (R 3.3.1)
digest 0.6.10 2016-08-02 CRAN (R 3.3.1)
evaluate 0.10 2016-10-11 CRAN (R 3.3.1)
htmltools 0.3.5 2016-03-21 CRAN (R 3.3.0)
knitr * 1.14.15 2016-11-03 Github (yihui/knitr@56faff4)
magrittr 1.5 2014-11-22 CRAN (R 3.2.1)
memoise 1.0.0 2016-01-29 CRAN (R 3.2.3)
png 0.1-7 2013-12-03 CRAN (R 3.3.0)
R.matlab * 3.6.1 2016-10-20 CRAN (R 3.3.1)
R.methodsS3 1.7.1 2016-02-16 CRAN (R 3.2.3)
R.oo 1.20.0 2016-02-17 CRAN (R 3.2.3)
R.utils 2.4.0 2016-09-14 CRAN (R 3.3.1)
Rcpp 0.12.7 2016-09-05 CRAN (R 3.3.1)
reach * 0.3.0 2015-10-17 Github (schmidtchristoph/reach@f503d44)
readr * 1.0.0 2016-08-03 CRAN (R 3.3.1)
rmarkdown * 1.1 2016-10-16 CRAN (R 3.3.1)
rsconnect 0.4.3 2016-05-02 CRAN (R 3.3.0)
rstudioapi 0.6 2016-06-27 CRAN (R 3.3.1)
stringi 1.1.2 2016-10-01 CRAN (R 3.3.1)
stringr * 1.1.0 2016-08-19 CRAN (R 3.3.1)
tibble 1.2 2016-08-26 CRAN (R 3.3.1)
withr 1.0.2 2016-06-20 CRAN (R 3.3.1)
yaml 2.1.13 2014-06-12 CRAN (R 3.2.1)