[Tex/LaTex] Algorithm with beamer package causing nightmares

algorithm2ebeamer

I have an algorithm which works well with \documentclass[twocolumn]{svjour3}. Now I have to present the paper and I would like one slide to be that algorithm, so I do:

\documentclass{beamer}
\usepackage{graphicx}

% definitions 
\newcommand{\kd}{\emph{k}-d\space}
\newcommand{\kgeraf}{\kd GeRaF}
\newcommand{\geraf}{{\tt GeRaF}\xspace}
% \newcommand{\att}[1]{\alert{#1}}
\newcommand{\att}[1]{{#1}}

% shortcuts 
\newcommand{\tab}{\hspace*{2em}}
\newcommand{\head}[1]{{\medskip\noindent\emph{#1}}}
\newtheorem{defn}{Definition}
\usepackage[top=2.5cm,bottom=2.3cm,hmargin=1.5cm]{geometry} % it was hmargin=2.5cm
\usepackage{amssymb,graphicx,url}
\setcounter{tocdepth}{3}
\usepackage[english]{babel}
\usepackage{multirow}
\usepackage[dvipsnames,svgnames]{xcolor}
\usepackage[ruled,vlined,linesnumbered]{algorithm2e}
\usepackage{xspace}
\usepackage{pgfplots}
%\pgfplotsset{compat=1.90} % ize got latex error here
\usepackage{pgfplotstable}
\def\real{{\mathbb R}}

%--------------------------------------------------------------------------------
% Add a period to the end of an abbreviation unless there's one
% already, then \xspace.
\makeatletter
\DeclareRobustCommand\onedot{\futurelet\@let@token\@onedot}
\def\@onedot{\ifx\@let@token.\else.\null\fi\xspace}
%
\def\eg{\emph{e.g}\onedot} \def\Eg{\emph{E.g}\onedot}
\def\ie{\emph{i.e}\onedot} \def\Ie{\emph{I.e}\onedot}
\def\cf{\emph{c.f}\onedot} \def\Cf{\emph{C.f}\onedot}
\def\etc{\emph{etc}\onedot} %\def\vs{\emph{vs}\onedot}
\def\wrt{w.r.t\onedot} \def\dof{d.o.f\onedot}
\def\etal{\emph{et al}\onedot}
\makeatother

%--------------------------------------------------------------------
\begin{document}
\begin{algorithm}[t]
\SetFuncSty{textsc}
\SetDataSty{emph}
\newcommand{\expr}[1]{$\langle${#1}$\rangle$}
\newcommand{\commentsty}[1]{{\color{DarkGreen}#1}}
\SetCommentSty{commentsty}
\SetKwComment{Comment}{$\triangleright$ }{}
\SetKwBlock{Block}{function}{}
\SetKwInOut{Input}{input}
\SetKwInOut{Output}{output}
\SetKw{Return}{return}
\SetKwData{Node}{node}
\SetKwData{Leaf}{leaf}
\SetKwFunction{Build}{build}
\SetKwFunction{Variance}{variance}
\SetKwFunction{Median}{median}
\SetKwFunction{Split}{split}

\Input
{pointset $X$, \#trees $m$, \#split-dimensions $t$, max \#points per leaf $p$
}
\Output{randomized \kd forest $F$}
\Begin
{
    $V \gets$ \expr{\Variance of $X$ in every dimension} \\
    $D \gets$ \expr{$t$ dimensions of maximum variance $V$} \\
    $F \gets \emptyset$  \Comment*[f]{forest} \\
    \For{$i \gets 1$ \KwTo $m$}
    {
        $f \gets$ \expr{random transformation}  \Comment*[f]{isometry, shuffling} \\
        $F \gets F \cup (f,\Build(f(X))$) \Comment*[f]{build on transformed X, store $f$}
    }
    \Return $F$
}

\BlankLine

\Block({$\Build(X)$} \Comment*[f]{recursively build tree (node/leaf)})
{
    \If(\Comment*[f]{termination reached}){$|X| \le p$}{\Return $\Leaf(X)$}
    \Else(\Comment*[f]{split points and recurse})
    {
        $s \gets$ \expr{one of dimensions $D$ at random} \\
        $v \gets $ \expr{\Median of $X$ in dimension $s$} \\
        $(L,R) \gets$ \expr{\Split of $X$ in dimension $s$ at value $v$} \\
        \Return $\Node(c,v,\Build(L),\Build(R))$ \Comment*[f]{build children on $L,R$}\\
    }
}
\caption{\kgeraf: building\label{alg:build}}
\end{algorithm}
\end{document}

which starts producing errors of that nature:

LaTeX Error: Option clash for package geometry.

I have searched for the past 3 hours in the site and applied the numerous answers, but I just can't make this thing compile! Can you?

Best Answer

You have several errors, I fixed the most important ones. All the \def should better be \newcommand.

Don't load geometry, it doesn't make sense, as beamer already loads it with its own options
Don't load xcolor, rather pass the options to the class as shown
Don't load graphicx
Don't load url, it's already coming with hyperref
Since you need a frame, it must be fragile because you perform macro definitions in it
Use \footnotesize or the algorithm will not fit
Use \protect\kgeraf in captions; probably the definition of \kd should be $k$\mbox{-}d

Here's a working version.

\documentclass[xcolor={dvipsnames,svgnames}]{beamer}

% definitions 
\newcommand{\kd}{\emph{k}-d\space}
\newcommand{\kgeraf}{\kd GeRaF}
\newcommand{\geraf}{{\tt GeRaF}\xspace}
% \newcommand{\att}[1]{\alert{#1}}
\newcommand{\att}[1]{{#1}}

% shortcuts 
\newcommand{\tab}{\hspace*{2em}}
\newcommand{\head}[1]{{\medskip\noindent\emph{#1}}}
\newtheorem{defn}{Definition}
%\usepackage[top=2.5cm,bottom=2.3cm,hmargin=1.5cm]{geometry} % it was hmargin=2.5cm
\usepackage{amssymb,url}
\setcounter{tocdepth}{3}
\usepackage[english]{babel}
\usepackage{multirow}
\usepackage[ruled,vlined,linesnumbered]{algorithm2e}
\usepackage{xspace}
\usepackage{pgfplots}
%\pgfplotsset{compat=1.90} % ize got latex error here
\usepackage{pgfplotstable}
\def\real{{\mathbb R}}

%--------------------------------------------------------------------------------
% Add a period to the end of an abbreviation unless there's one
% already, then \xspace.
\makeatletter
\DeclareRobustCommand\onedot{\futurelet\@let@token\@onedot}
\def\@onedot{\ifx\@let@token.\else.\null\fi\xspace}
%
\def\eg{\emph{e.g}\onedot} \def\Eg{\emph{E.g}\onedot}
\def\ie{\emph{i.e}\onedot} \def\Ie{\emph{I.e}\onedot}
\def\cf{\emph{c.f}\onedot} \def\Cf{\emph{C.f}\onedot}
\def\etc{\emph{etc}\onedot} %\def\vs{\emph{vs}\onedot}
\def\wrt{w.r.t\onedot} \def\dof{d.o.f\onedot}
\def\etal{\emph{et al}\onedot}
\makeatother

%--------------------------------------------------------------------
\begin{document}

\begin{frame}[fragile]
\footnotesize
\begin{algorithm}[H]
\SetFuncSty{textsc}
\SetDataSty{emph}
\newcommand{\expr}[1]{$\langle${#1}$\rangle$}
\newcommand{\commentsty}[1]{{\color{DarkGreen}#1}}
\SetCommentSty{commentsty}
\SetKwComment{Comment}{$\triangleright$ }{}
\SetKwBlock{Block}{function}{}
\SetKwInOut{Input}{input}
\SetKwInOut{Output}{output}
\SetKw{Return}{return}
\SetKwData{Node}{node}
\SetKwData{Leaf}{leaf}
\SetKwFunction{Build}{build}
\SetKwFunction{Variance}{variance}
\SetKwFunction{Median}{median}
\SetKwFunction{Split}{split}

\Input
{pointset $X$, \#trees $m$, \#split-dimensions $t$, max \#points per leaf $p$
}
\Output{randomized \kd forest $F$}
\Begin
{
    $V \gets$ \expr{\Variance of $X$ in every dimension} \\
    $D \gets$ \expr{$t$ dimensions of maximum variance $V$} \\
    $F \gets \emptyset$  \Comment*[f]{forest} \\
    \For{$i \gets 1$ \KwTo $m$}
    {
        $f \gets$ \expr{random transformation}  \Comment*[f]{isometry, shuffling} \\
        $F \gets F \cup (f,\Build(f(X))$) \Comment*[f]{build on transformed X, store $f$}
    }
    \Return $F$
}

\BlankLine

\Block({$\Build(X)$} \Comment*[f]{recursively build tree (node/leaf)})
{
    \If(\Comment*[f]{termination reached}){$|X| \le p$}{\Return $\Leaf(X)$}
    \Else(\Comment*[f]{split points and recurse})
    {
        $s \gets$ \expr{one of dimensions $D$ at random} \\
        $v \gets $ \expr{\Median of $X$ in dimension $s$} \\
        $(L,R) \gets$ \expr{\Split of $X$ in dimension $s$ at value $v$} \\
        \Return $\Node(c,v,\Build(L),\Build(R))$ \Comment*[f]{build children on $L,R$}\\
    }
}
\caption{\protect\kgeraf: building}\label{alg:build}
\end{algorithm}

\end{frame}
\end{document}

Related Solutions

[Tex/LaTex] Algorithm with chapter number

Since you are using both the algorithm2e and algorithm packages, and both of them implement some same commands/environments, you had to use the algo2e option for algorithm2e to prevent a clash and this overrides the settings for the counter that you set as options for algorith2e, since the environment that will be used is the one from algorithm. You have three possible solutions:

If you don't need the algorithm package, load only algorithm2e and not algorithm and load the former without the algo2e option:

\documentclass[11pt]{book}
\usepackage[ruled,vlined,linesnumbered,resetcount,algochapter]{algorithm2e}

\begin{document}

\chapter{chapter1}
\begin{algorithm}[H]
\centering
test
\caption{algo}
\label{algo1}
\end{algorithm}

\end{document}

enter image description here

If you really need both algorithm2e and algorithm in your document, and you want to use the algorithm environment (from the algorithm package) add the following lines to your preamble, after loading the packages:

\makeatletter 
\renewcommand\thealgorithm{\thechapter.\arabic{algorithm}} 
\@addtoreset{algorithm}{chapter} 
\makeatother

A complete example:

\documentclass[11pt]{book}
\usepackage[ruled,vlined,linesnumbered,algo2e,resetcount,algochapter]{algorithm2e}
\usepackage{algorithm}

\makeatletter 
\renewcommand\thealgorithm{\thechapter.\arabic{algorithm}} 
\@addtoreset{algorithm}{chapter} 
\makeatother

\begin{document}

\chapter{chapter1}
\begin{algorithm}[H]
\centering
test
\caption{algo}
\label{algo1}
\end{algorithm}

\end{document}

enter image description here

If you really need both algorithm2e and algorithm in your document, and you want to use the algorithm environment (from the algorithm2e package), use the algorithm2e environment (the algo2e option renames algorithm to algorithm2e):

\documentclass[11pt]{book}
\usepackage[ruled,vlined,linesnumbered,algo2e,resetcount,algochapter]{algorithm2e}
\usepackage{algorithm}

\begin{document}

\chapter{chapter1}
\begin{algorithm2e}[H]
\centering
test
\caption{algo}
\label{algo1}
\end{algorithm2e}

\end{document}

enter image description here

By the way (not related to the question), you are loading packages more than once; this should be avoided.

[Tex/LaTex] Big algorithm causing big mess

From the algorithm2e documentation:

The optional arguments [Hhtbp] works like those of figure environment. The H argument forces the algorithm to stay in place. If used, an algorithm is no more a floating object. Caution: algorithms cannot be cut, so if there is not enough place to put an algorithm with H option at a given spot, LaTeX will place a blank and put the algorithm on the following page.

Give more sensible placement options for a large floating environment: tbp.

But now the float is still too big for the page (which causes it to be pushed far later in the document as you mentioned in a comment). You have three options:

Reword your explanations to shorten it.
Split it into smaller pieces for presentation.
Set it in a smaller font size to fit.

Because options 1 and 2 are dependent on your algorithm and field, I've shown option 3 here. But I think option 1 or 2 would be preferable for a real algorithm/document.

All I did was change your line

\begin{algorithm}[H]

\begin{algorithm}[tbp]\footnotesize

and here is the result (page 1 of 1 for this MWE, no leading blank page and no overfull boxes or float too large warnings in the log file):

enter image description here

Best Answer

Related Solutions

[Tex/LaTex] Algorithm with chapter number

[Tex/LaTex] Big algorithm causing big mess

Related Question