Writing Reproducible Research Reports

LaTex is a type setting program. It is widely used in the mathematical and computing communities for its ability to produce precisely and correctly typeset mathematical notation. But it has grown well beyond those roots. We will use it as the method to go from a plain .tex file to a .pdf file.

Kc : LaTeX tools for report generation (i2c4p) from Britt Anderson on Vimeo.

Getting the latex packages you will need from Ubuntu for these exercises. What is LaTeX?

Our emacs editor will be able to export our org file into a tex file. To go the rest of the way to a pdf we will need various libraries and packages available through our Linux installation. Since these will be used by emacs, but not part of emacs, we will use the system package manager to get and install them. It is quite possible to use latex and all the related tools from the command line and it can often be helpful to do so. It is also possible to use these libraries from other editors. Thus we will install them with apt, a package manager for *buntus.

sudo apt install texlive-latex-recommended
sudo apt install texlive-latex-extra

There is also a "full" package as well that you can install instead of the above.

LaTeX could be a whole lecture (maybe even a course) in itself. We will not go into that here, but we will just use it behind the scenes. If you are interested though there are some nice tutorials.

• The not so short introduction (pdf)
• Not as good but really short
• LaTeX for Word Processor Users (pdf)
• A polemic for writing in latex You can do this in emacs too. Look for information on the Auctex.

There are many pdf viewers that work well in the linux world. A very minimal version is xpdf whereas a much more full feature viewer is okular. However, we are working in emacs and it will be convenient to install a pdf-viewer that works in that environment. Emacs comes with a "doc-view" that will render a pdf, but it is not a very pretty view and our ability to interact and edit the pdf is limited. We will use a much more full featured and modern pdf tool called pdf-tools.

The Github Repo for PdfTools provides extensive documentation on use and installation. I have copied the minimal version here to get us going. Note that the pdf-tools instructions describes a package manager called aptitude. You can use apt instead (https://www.tecmint.com/difference-between-apt-and-aptitude/).

It is common for the Ubuntu packages to come in both regular and development versions. As a rule of thumb if you are unsure which to install always pick the development packages (noted by the -dev at the end). This will include some extra files and details, but you will often need those extra files when trying to "make" or assemble code yourself.

Kd : PdfTools (i2c4p) from Britt Anderson on Vimeo.

How to view and annotate pdfs in Emacs. The PdfTools package installed.

You may already have many of the files you will need to build the pdf-tools package, but just in case make sure to install:

1. gcc
2. g++
3. make
4. automake
5. autoconf
6. libpng-dev
7. zlib1g-dev
8. libpoppler-glib-dev
9. libpoppler-private-dev
10. imagemagick

Run M-x package-list-packages to refresh and display available packages. Search for let-alist and tablist. Mark them both with an i and then type x to execute the installation.

Then search for pdf-tools make it with an i and then enter x. This will download the pdf-tools package files to the correct directory (usually ~/.emacs.d/elpa/pdf-tools).

Edit your init.el file and add the line

(pdf-tools-install)

The easiest way to trigger the build will be to save everything. Exit emacs. And then restart emacs.

On your first restart you will be asked whether to rebuild an epdf program. Say yes. Pdf-Tools will then figure out that you are on a -buntu and will ask for your super user password so that it can download from the Ubuntu package system any further packages or software it requires. At the end of that process you will see a message telling you if the compiliation completed successfully. If so, you can open a pdf in emacs just like any other file, C-x C-f.

Test it out with a pdf. If you don't have one handy just down load one via scholar.google.com to your desktop and then open the file from within Emacs.

Go back to the file you wrote in Rmd and which has the for-loop. Now convert that file to an org file (just do a save as: C-x C-w substituting .org for .rmd). Edit the code blocks that you wrote as R chunks to be source blocks in org mode. You have seen examples of this already, but there is a reminder here. Now compile it with the C-c C-e l p command to generate a pdf. Keep tweaking until you get it to work.

Pretty much happens automatically. You may have to play with the header arguments to get exactly the look you want. The header arguments are those things appearing on the +#Begin_Src line.

d <- data.frame(foo=c('a','b','c'), bar=c(1.0/3.0,22,32))
d

  foo        bar
1   a  0.3333333
2   b 22.0000000
3   c 32.0000000

Kf : Demo 2: Code and Tables (i2c4p) from Britt Anderson on Vimeo.

Continuing our demonstration of report generation we now show how to include inline code, tables, and figures.

Try adding this to your prior file just to make sure you can get the table to execute.

An inline result is one that appears at the correct place in the text.

Imagine that you are doing an analysis for a report and instead of having to cut and paste from some other statistical program you can just magically have the result appear in the right location in the text! Try it. Here is an example to adapt into your current file. Try it with a different number of values and a different type of random number (e.g. how about an exponential distribution)?

The mean of 100 mean 0 normally distributed numbers is 0.0498254924850924.

Yes. Of course you can. And you should.

Kg : Report Demo 3: Including Citations (i2c4p) from Britt Anderson on Vimeo.

The final demo of report generation shows you how to get and use bibtex formatted references for managing citations in your report.