knitr: fig.height must be numeric in new version? - latex

I downloaded a new version of the knitr package and suddenly I get the following error when trying to include a plot in an Rnw file:
Error in options[[sprintf("fig.%s", i)]] * options$dpi :
non-numeric argument to binary operator
Calls: knit ... process_group.block -> call_block -> fix_options -> %n%
After a lot of experimenting I found out the problem is with the command fig.height, apparently it is supposed to be numeric. This code produces the error:
\documentclass{article}
\begin{document}
<<cat, fig.cap="dog", fig.height='\\textheight'>>=
plot(c(1, 2, 3), c(2, 1, 3))
#
\end{document}
This code does not:
\documentclass{article}
\begin{document}
<<cat, fig.cap="dog", fig.height=6>>=
plot(c(1, 2, 3), c(2, 1, 3))
#
\end{document}
Unfortunately I have no idea what the unit of measurement for this 6 is.
This code:
\documentclass{article}
\begin{document}
<<cat, fig.cap="dog", fig.height=6cm>>=
plot(c(1, 2, 3), c(2, 1, 3))
#
\end{document}
produces a different error:
Error in base::parse(text = code, srcfile = NULL) :
1:44: unexpected symbol
1: alist( 'cat', fig.cap="dog", fig.height=6cm
^
Calls: knit ... withCallingHandlers -> eval -> parse_only -> <Anonymous>
Execution halted
All in all it seems as if knitr is really expecting a numeric value for fig.height, which is moderately annoying because I have already typed a lot of fig.height='\\textheight' commands throughout my document (back in the days that that still worked) and I also do not really know what the numeric value corresponding to Latex's \textheight is.
Any advise?
UPDATE: I did some more experimenting and it turns out that the numbers you feed to fig.height appear nowhere in the Latex-file, instead they influence the dimension of the cat.pdf image produced by knitr directly. This kind of suggest that leaving out fig.height altogether might be the best strategy. Still this doesn't tell us how to get the [height=\textheight] in the tex document.

Use the chunk option out.height instead of fig.height.

Related

How to implement full LaTeX syntax in Rmarkdown?

I have a two part question regarding the use of LaTeX in Rmarkdown:
1) When working in Rmarkdown (with the intent to render to PDF), is there a rule for when we should just use the double dollar signs ($$) to write something in LaTeX or when we should use the LaTeX syntax to begin and end all of our LaTeX code (e.g. \documentclass{...}, \begin{document}, \end{document}, etc.
I believe I have read that it is okay to just use the latter option, and Rmarkdown will ignore all of the escaped latex commands if the document is rendered to anything other than PDF.
2) The reason I am asking, in this case, is that I am trying to incorporate some labelled matrix multiplication in an Rmarkdown document I am writing. Specifically, I would like to include some matrices that take the form show on this page. Here is the code:
\documentclass{article}
\usepackage{amsmath}
\newenvironment{spmatrix}[1]
{\def\mysubscript{#1}\mathop\bgroup\begin{pmatrix}}
{\end{pmatrix}\egroup_{\textstyle\mathstrut\mysubscript}}
\begin{document}
\begin{equation}
\begin{spmatrix}{A}
a & b \\
c & d
\end{spmatrix}
\begin{spmatrix}{x}
x_1 \\
x_2
\end{spmatrix}
=
\begin{spmatrix}{b}
b_1 \\
b_2
\end{spmatrix}
\end{equation}
\end{document}
How would one implement this code in Rmarkdown? Do you need to move the \usepackage call into the YAML as suggested in other threads discussing the loading of LaTeX packages in Rmarkdown? Is the first line, \documentclass{article} even needed within an Rmarkdown document?
I'm new to all of this, and thusfar, I've been able to get by using the double dollar signs to set off all my LaTeX code for simple equations and even a simple matrix here and there that I've tried to write, but for mathematical expressions that require more formatting, aligning, multi-line proofs, etc., most of the examples I've encountered are on the TEX boards written with syntax similar to the code above. I haven't been able to figure out how to implement these types of examples in Rmarkdown. Any helpful suggestions or pointers where to better understand this issue would be much appreciated!
It is indeed possible to put most body-level LaTeX constructs into the body of your Rmd file. Other backends will ignore these constructs, but the result might look change. So from my point of view your are binding yourself to PDF output. But that might be fine in your case.
Concerning your concrete problem:
amsmath is already included by the default template, which also takes care of \ḑocumentclass and the document environment.
You need to add the environment definition into a separate tex file (in my case preamble.tex) and include that file via the YAML headers.
You can then use the LaTeX constructs as is in the Rmd body.
Putting things together:
---
output:
pdf_document:
keep_tex: yes
includes:
in_header: preamble.tex
---
\begin{equation}
\begin{spmatrix}{A}
a & b \\
c & d
\end{spmatrix}
\begin{spmatrix}{x}
x_1 \\
x_2
\end{spmatrix}
=
\begin{spmatrix}{b}
b_1 \\
b_2
\end{spmatrix}
\end{equation}

How to diagnose errors in LaTeX generated by Doxygen 1.8.x: LT#LL#FM#cr

I've been using Doxygen successfully to generate PDF documentation for a sizable Fortran 90 project since v1.6. After a recent upgrade to Doxygen 1.8, pdflatex is choking with an error I can't understand. From refman.log:
.
.
.
<use classfate__source_a022bf629bdc1d3059ebd5fb86d13b4f4_icgraph.pdf>
Package pdftex.def Info: classfate__source_a022bf629bdc1d3059ebd5fb86d13b4f4_ic
graph.pdf used on input line 607.
(pdftex.def) Requested size: 350.0pt x 65.42921pt.
)
(./classm__aerosol.tex
! Undefined control sequence.
<recently read> \LT#LL#FM#cr
l.25 ...1833ffa6f2fae54ededb}{ia\-\_\-nsize}), \\*
? ?
Type <return> to proceed, S to scroll future error messages,
R to run without stopping, Q to run quietly,
I to insert something, E to edit your file,
1 or ... or 9 to ignore the next 1 to 9 tokens of input,
H for help, X to quit.
Looking at the first 25 lines of classm__aerosol.tex, nothing obviously matches the error message:
\hypertarget{classm__aerosol}{\section{m\-\_\-aerosol Module Reference}
\label{classm__aerosol}\index{m\-\_\-aerosol#{m\-\_\-aerosol}}
}
Contains general aerosol-\/related constants and routines.
\subsection*{Public Member Functions}
\begin{DoxyCompactItemize}
\item
subroutine \hyperlink{classm__aerosol_aa06c1f39c6bd34f22be92d21535f0320}{aerdis} (I\-A\-E\-R\-O, M\-A\-E\-R\-O, V\-O\-L, A\-R\-E\-A, M\-U, T\-G\-A\-S, R\-H\-O, A\-G\-A\-M\-M\-A, X\-L\-A\-E\-R, D\-M\-E\-A\-N, N\-A\-E\-R, X\-N\-D\-A\-E\-R, L\-S\-D\-A\-E\-R)
\begin{DoxyCompactList}\small\item\em Return aerosol mass given a volume, based on aerosol size distribution function. \end{DoxyCompactList}\item
real(kind=wp) function \hyperlink{classm__aerosol_a2dff4ff413057e8788fba7270a30c093}{lamsed} (V\-O\-L, H, M\-U\-G, R\-H\-O\-A\-E\-R, A\-G\-A\-M\-M\-A, A\-C\-H\-I, A\-F\-E\-O, K\-O, M\-A\-E\-R, F\-M\-A\-E\-R, F\-A\-E\-R\-S\-S, F\-S\-E\-D\-D\-K)
\begin{DoxyCompactList}\small\item\em Calculate aerosol removal constant and interpolation factor between steady-\/state and decaying aerosol correlations. \end{DoxyCompactList}\item
pure real(kind=wp) function \hyperlink{classm__aerosol_a6d0a04004f49c404c67e0aa69dd39ee1}{fdbend} (V\-E\-L, H\-S\-E\-D, T\-G, R\-H\-O\-G, M\-U\-G, R\-H\-O\-P\-A\-R, C\-A\-E\-R\-O, X\-D\-B\-E\-N\-D, N90\-J)
\begin{DoxyCompactList}\small\item\em Find total impaction efficiency for aerosol deposition considering 90-\/degree bends in a flow path. \end{DoxyCompactList}\end{DoxyCompactItemize}
\subsection*{Public Attributes}
\begin{DoxyCompactItemize}
\item
integer, parameter \hyperlink{classm__aerosol_a8f604b7ffe3c1833ffa6f2fae54ededb}{ia\-\_\-nsize} = 30
\item
integer, parameter \hyperlink{classm__aerosol_ae71813ecf0c7768af9d6292efb14774f}{ia\-\_\-nmass} = 10
\item
real(kind=wp), dimension(\hyperlink{classm__aerosol_a8f604b7ffe3c1833ffa6f2fae54ededb}{ia\-\_\-nsize}), \\*
Nothing obviously matches the recently read chunk "\LT#LL#FM#cr" and I don't know enough low-level TeX to translate that into something that might actually be in the source text.
Suspecting this might have been fixed in a later version of Doxygen than the one shipping with Linux Mint (v1.8.1.2), I built & installed v1.8.3.1 from source, updated my doxyfile, blew away the old documentation and regenerated it. I get the same baffling error.
There's nothing obvious in refman.log that would indicate missing or broken LaTeX packages and I'm completely at a loss as to what's causing this.
As this still gets a hit on Google when you search:
doxygen missing $ inserted
I would like to add something.
Do not use a PROJECT_NAME containing underscores (_)!
After a brief look into the doxygen's current documentation (I am using 1.8.4) it does not make that explicit.
this will be difficult to solve unless you provide a bit more information - possibly using \errorcontextlines=9999 as suggested in the comments on the question.
as a first short though, the name of the control sequence that can't be found (i.e. \LT#LL#FM#cr) is one defined by the longtable package (documentation, p. 15) - thus adding:
\usepackage{longtable}
to the preamble of the document might help.
If so, according to the doxygen documentation here, adding the following to your configuration file should do the trick:
EXTRA_PACKAGES=longtable

How do I use \renewcommand to get BACK my greek letters?

I'm a LaTeX newbie, but I've been doing my homework, and now I have a question that I can't seem to find the answer to.
I create the definition of an equation, let's just say it's this one:
The potential is characterized by a length $\sigma$ and an energy $\epsilon$.
In reality, this equation is more complex, which is why I wanted to try a shortcut. If my equation were this simplistic, I wouldn't try my substitution technique.
I use the \renewcommand to save me some time:
\renewcommand{\sigma}{1}
And this works fabulously and will replace all instances of sigma with 1. Unfortunately though, since \sigma has a global scope, I need to reset it.
I tried a couple different ways:
Attempt 1: -deadlock due to circular reference?
\newcommand{\holdsigma}{\sigma}
\renewcommand{\sigma}{1}
The potential is characterized by a length $\sigma$ and an energy $\epsilon$.
\renewcommand{\sigma}{\holdsigma}
I would think to reset the command, it should look something like this:
\renewcommand{\sigma}{\greek{\sigma}}
but that obviously hasn't worked out for me.
Any idea about how the greek letters are originally defined in the language?
I have to admit that I don't understand why you want to do what you're asking, but this should work:
\documentclass{article}
\begin{document}
Before redefinition, \verb|\sigma| looks like $\sigma$.
% Copy the current definition of \sigma to \oldsigma
\let\oldsigma\sigma
% Redefine \sigma to be '1'
\renewcommand{\sigma}{1}
After redefinition, \verb|\sigma| looks like $\sigma$.
You can still use \verb|\oldsigma| if you want to use the original definition $\oldsigma$.
% Restore the original definition of \sigma
\let\sigma\oldsigma
Now \verb|\sigma| is back to its normal appearance $\sigma$.
\end{document}
To find out how \sigma or any other command is originally defined, you can use \show\sigma. (The answer is that \sigma is defined as \mathchar"11B.) You can type this either in your document itself — compilation will pause and you can type Enter after reading the reply — or you can type this in TeX/LaTeX's interactive mode.
Example with a document:
\documentclass{article}
\begin{document}
What is $\sigma$? % Prints "What is σ" in the DVI/PS/PDF.
\show\sigma % Prints "> \sigma=\mathchar"11B." in the compilation.
Now that we know, let us redefine it.
\renewcommand{\sigma}{1}
Now it is: $\sigma$. % Prints "Now it is: 1." in the DVI/PS/PDF.
OK, let's go back.
\renewcommand{\sigma}{\mathchar"11B}
We again have: $\sigma$. %Prints "We again have: σ." in the DVI/PS/PDF.
\end{document}
Or else at the command prompt, type latex, then type \relax, then type \show\sigma, read what it says, then type x to exit.

Figures occurring after ^ and _ macros (was: LaTeX limitation?)

I've hit an annoying problem in LaTeX. I've got a tex file of about 1000 lines. I've already got a few figures, but when I try to add another figure, It barfs with:
! Undefined control sequence.
<argument> ... \sf#size \z# \selectfont \#currbox
l.937 \begin{figure}[t]
If I move the figure to other parts of the file, I can get similar errors on different lines:
! Undefined control sequence.
<argument> ... \sf#size \z# \selectfont \#currbox
l.657 \paragraph
{A Centering Algorithm}
If I comment out the figure, all is ok.
%\begin{figure}[t]
% \caption{Example decision tree, from Reiter and Dale [2000]}
% \label{fig:relation-decision-tree}
% \centering
% \includegraphics[keepaspectratio=true]{./relation-decision-tree.eps}
%\end{figure}
If I keep just the begin and end like:
\begin{figure}%[t]
% \caption{Example decision tree, from Reiter and Dale [2000]}
% \label{fig:relation-decision-tree}
% \centering
% \includegraphics[keepaspectratio=true]{./relation-decision-tree.eps}
\end{figure}
I get:
! Undefined control sequence.
<argument> ... \sf#size \z# \selectfont \#currbox
l.942 \end
{figure}
At first, I thought maybe LaTeX has hit some limit, and I tried playing with the ulimits, but that didn't help. Any ideas?
i've got other figures with graphics already. my preamble looks like:
\documentclass[acmcsur,acmnow]{acmtrans2n}
\usepackage{array}
\usepackage{lastpage}
\usepackage{pict2e}
\usepackage{amsmath}
\usepackage{varioref}
\usepackage{epsfig}
\usepackage{graphics}
\usepackage{qtree}
\usepackage{rotating}
\usepackage{tree-dvips}
\usepackage{mdwlist}
\makecompactlist{quote*}{quote}
\usepackage{verbatim}
\usepackage{ulem}
I found, not that it's a problem with \textsuperscript, but that it's with a ^ def I picked up from http://anthony.liekens.net/index.php/LaTeX/SubscriptAndSuperscriptInTextMode . The fix is to put the use of ^ in {}, as in I've put entire sections where I use lots of ^ and _ in {}. Hurrah!
During the end of my Master Thesis I also had the problem that after some amount of figures, I got an error without any special error message. After I read you thread, I also tried something with the packages included and in the end I was successful by taking out the \usepackage{pxfonts} and \usepackage{txfonts}. Yeah, finally.. I almost went crazy.. ;)
If I Google for "latex undefined control sequence" I get this.
I've successfully included a graphic into LaTeX using something like this:
\usepackage{amsmath,amsthm,graphicx}
...
I just wanted to test adding an image to a \LaTeX file:
\includegraphics[scale=0.60]{basic-info.png}
I typeset an entire dissertation of 200 pages with lots of figures in LaTeX and didn't run into a limit like that. I'd bet on a syntax problem first before I'd assume a size issue.
Your error lies elsewhere. I wouldn't be the least surprised if it turned out to be the document class. Try altering your document for \documentclass{article} and see where you get. If that fixes the problem you can complain to the ACM (ROTFLMAO—I've dealth with ACM).
If that doesn't fix, the problem, slip in a \tracingall somewhat before the offending figure or section, put the results into http://pastebin.com/, and let us know.
Please receive the thanks of a (formerly) utterly-confounded graduate student. Quick clarification for other users:
{The quickest ^{way} to put this solution into practice is to bracket all sections of text involving the character ``\^'' as shown here.}

Latex listings-package format option for uppercase keywords

I use the listings package to insert source code. I would like to print all keywords uppercase in the output, regardless of the case in the input.
The manual states that
keywordstyle=[number][*]style
produces just what I want. However the following (almost) minimal example does not work.
if I set keywordstyle to "[1][]{\bfseries}" I end up with "[]" in front of every keyword
and "[*]{\bfseries}" gives me an asterisk in the start of the document.
I also tried "\MakeUppercase" and "{\MakeUppercase}" for keywordstyle which resulted in several errors, the first being:
! Incomplete \iffalse; all text was ignored after line 11
Minimal example:
\documentclass{article}
\usepackage{listings}
\lstdefinelanguage{KA_assembler}
{morekeywords={add,and,or,xor},
keywordstyle=[1][*]{\bfseries},
sensitive=false,
}
\lstset{language=KA_assembler}
\begin{document}
\begin{lstlisting}
and %r1, %r2
xor %r2, %r3
and %r4, %r5
\end{lstlisting}
\end{document}
I use Miktex for compilation of the tex files. So how do I force uppercase for Keywords?
In the manual, the brackets around the * look a bit different then the brackets around number. The reason is that the brackets around * are not meant to be used in the latex code, they just indicate that the presence of the * is optional. So try
keywordstyle=[1]*\bfseries
or
keywordstyle=*\bfseries
- it worked for me.

Resources