Sorry for that, I should have paid more attention to the scoping in my last answer.
You can wrap all those pgfplotsset options in a new style functionallineplot, which you then only apply to the axis for the functional line. You'll need to change the #1 to ##1 in those options, but that's all the adjustment that's necessary.
\documentclass[10pt,a4paper]{article}
\usepackage{pgfplots}
\pgfplotsset{
functionallineplot/.style={
shift down/.style={
y filter/.code={\pgfmathparse{\pgfmathresult*(##1)}}
},
shift down/.default=1,
every axis plot post/.style={restrict y to domain=0.5:inf},
positive/.style={
no markers,
red
},
negative/.style={
no markers,
blue
},
/tikz/function label/.style={
anchor=east
},
step functionallinenumber/.code={
\stepcounter{functionallinenumber}
},
title entries/.initial={}
}
}
\makeatletter
\newcommand\functionalline[4][\@empty]{
\edef\plots{
\noexpand\addplot [negative, shift down=#3, forget plot] {#2<0};
\noexpand\addplot [positive, shift down=#3, forget plot] {#2>0};
}
\plots
\node at (axis cs:\pgfkeysvalueof{/pgfplots/xmin},#3) [function label] {%
\ifx#1\@empty%
$#2$%
\else%
$#1$%
\fi
};
\pgfplotsinvokeforeach {#4} {
\node at (axis cs:##1,#3) [
fill=white,
inner sep=1pt,
declare function={x=##1;} % Set 'x' to current position
] {%
\pgfkeys{/pgf/fpu}% Use the fpu library, because it doesn't throw an error for divide by zero, but sets result to +/- inf
\pgfmathparse{#2}%
\pgfmathfloatifflags{\pgfmathresult}{0}{\hspace{-0.75ex}0}{x}% Check whether result is zero. The \hspace is necessary because of a bug in the fpu library
\pgfkeys{/pgf/fpu=false}%
};
}
}
\newenvironment{functionallines}[3][]{
\begin{tikzpicture}
\begin{axis}[
functionallineplot,
extra x ticks = {#2},
grid=none,
xticklabel pos=right,
hide y axis,
x axis line style={draw=none},
every tick label/.style={
anchor=base,
yshift=1ex,
gray!50
},
every extra x tick/.style={
every tick label/.style={
anchor=base,
yshift=1ex,
inner xsep=0pt,
fill=white,
text=black
}
},
extra x tick style={grid=major},
xtick pos=right,
major tick length=0pt,
enlarge x limits=false,
enlarge y limits={abs=0.75},
domain=#3,
samples=100,
y dir=reverse, y = -0.5cm,
clip=false,
#1
]
}{
\coordinate (bottom right) at (rel axis cs:1,0);
\coordinate (top right) at (rel axis cs:1,1);
\end{axis}
\draw [-latex] (top right-|current bounding box.west) -- (top right) node [right] {$x$};
\draw (bottom right) -- (bottom right-|current bounding box.west);
\end{tikzpicture}
}
\begin{document}
\begin{figure}[h!tbp]
\centering
\begin{tikzpicture}
\begin{axis}[
scale only axis,
grid=major,
grid style={dashed, gray!30},
axis lines=middle,
inner axis line style={=>},
ymin=-4.2,
ymax=7.2,
xmin=-3.2,
xmax=3.2,
]
\addplot[color=red,thick,samples=50,domain=-3:3] {x^2-2};
\addplot[color=blue!60!white,thick,samples=50,domain=-3:3] {2*x-3};
\addplot[color=blue!30!white,thick,samples=50,domain=-3:3] {2*x-2};
\end{axis}
\end{tikzpicture}
\end{figure}
\begin{figure}[!htbp]
\centering
\begin{functionallines}[
extra x tick labels={$-1$},
xtick={-2,0}]{-1}{-2:0}
\functionalline{2(x+1)}{1}{-1}
\functionalline[f'(x) = 2x + 2]{x+1)}{2}{-1}
\end{functionallines}
\end{figure}
\end{document}
The compat number refers to version numbers for the pgfplots package. By setting the compat you tell pgfplots to use settings/features from that specific version.
If 1.15 doesn't work, that means you don't have version 1.15 of pgfplots installed. The compat setting can not look into the future.
Just use 1.13, or possibly try 1.14.
If you want to update pgfplots, you might be able to use the TeX Live Utility to do that, see How do I update my TeX distribution?, assuming you're using MacTeX.
Best Answer
As user of the
pgfplotssoftware, I do not want to risk that a package update leads to a different outcome of my existing images. Consequently, I would like to have an opportunity to "preserve" my output even if the package author decides to change something for "new" documents. The alternative would be to browse through entire documents to see if any "obvious" change has occurred. I expect that such changes do not happen, even if an isolated view on the image clearly confirms that a change was caused by a bugfix.On the other hand, I as author of
pgfplotsneed the freedom to make changes to the software, either because of new features or in order to fix bugs. I have an interest in maintaining existing workarounds created by users, either those derived becausepgfplotslacked support for something or becausepgfplotshad a bug. A package update should not invalidate old work-arounds, but it should provide a fix for all those users who "started a document from scratch".Choosing a suitable
\pgfplotsset{compat=<version>}flag is a solution which addresses both use-cases, and the fact thatpgfplotssuggests a suitable value for<version>should make it simpler to choose it.However, the choice
compat=newestspoils both use-cases: both suffer seriously.There may be a scenario in which backwards compatibility is of minor importance - this is where
compat=newesthas a use-case. Personally, I do not know what the use-case might be. The choicecompat=newestmeans "I do not care if my old figures change in appearance after the next version upgrade".