Please read the appropriate sections below, which give specific
instructions for installing and testing the pomp
software
and other packages we will be using for data analysis using POMP models.
First follow the instructions for “all users”, then those for your
specific operating system (OS).
Important note: If you run into problems, email ionides@umich.edu with a detailed description of the problem you’ve encountered. In this message, be certain to include all of the following information:
In particular, it is often easiest to send a screenshot or transcript
showing the commands you’ve entered and the error messages you’ve
received. In R, you can run Sys.info()
to
get a printout of the operating system and software version numbers.
You’ll need at least version 4.3.1 of R. Source code and binaries are available on CRAN (https://cran.r-project.org/). Install the latest version of RStudio from rstudio.com.
Open a session in RStudio and run the following:
> update.packages()
> source("https://ionides.github.io/531w24/pomp_prep/packages.R")
[The >
is the command prompt; it is not part of
the command. Also, depending on your email client propgram, you may need
to replace the quotation marks with plain keyboard double
quotes.]
The first command updates your installed packages. You may be prompted to specify a CRAN mirror: choose one geographically near you. In RStudio, you can also push the “Update” button on the “Packages” tab to accomplish this.
The second command runs a script on my website. It will install some needed packages if these are not already installed on your system.
If your machine runs Windows, you must install Rtools. This will give you the ability to compile C code and dynamically link it into an R session.
Download Rtools from CRAN and install it. A video tutorial demonstrating how to install Rtools may be helpful. Also, detailed installation instructions are available on CRAN. Note that, after installation, there is one more step to be completed.
So that you can compile C code and dynamically link it into an R session, you will need to have the Xcode app installed. This is gratis and can be installed via the App Store or downloaded from developer.apple.com. Video tutorials demonstrating how to check if you need to install Xcode, how to install Xcode, and how to install pomp once you have installed Xcode are available.
Note that you must go beyond merely installing the Xcode app. After you’ve installed the app, open a unix terminal (listed as the Terminal app under “Utilities” in the Finder) and run the following line
xcode-select --install
This will install the “Command Line Tools” that are needed to compile native C codes. Running this command is tantamount to the part of the Xcode installation tutorial video where we complete the installation using the graphical user interface.
Open a session in RStudio and run the following:
> source("https://kingaa.github.io/scripts/pompTest.R")
This will check whether you can work with pomp.
If it fails, try the following:
> source("https://kingaa.github.io/scripts/hello.R",echo=TRUE)
If this fails to give the “Hello!” message, you will need to follow
the instructions below that correspond to your OS before re-trying the
pompTest.R
script.
If you have trouble with any of the scripts above, make sure you have the GNU compiler collection (GCC), including gfortran, installed on your computer. Linux distributions typically include this by default but it is not impossible that you have somehow avoided this.
If the pompTest.R
script fails because you cannot load
pomp, try installing it from source. The easiest way to
do this is to use the devtools package. Do
install.packages("devtools")
library(devtools)
install_github("kingaa/pomp")
If, while trying to install from source, you receive the error,
make: gfortran-8.2: No such file or directory
or one that otherwise refers to gfortran
, then it is
likely that you do not have the necessary version of
gfortran installed. Have a look at these instructions and
contact me at the address above if these don’t work for you.
Some users have reported receiving an error complaining that
'stdlib.h' file not found
This indicates that the command-line tools from Xcode have not been properly installed. In a unix terminal, run
xcode-select --install
to correct this problem.
You have probably failed to install the Rtools correctly. Revisit the instructions above.
Back to course homepage
Video
instructions for Windows users
Video
instructions for MacOS users
Acknowledgment:
these instructions are derived from
https://kingaa.github.io/sbied/prep