Introduction

What is plotcpp?

plotcpp is a 2D plotting library for modern C++. Its purpose is to provide a variety of data visualization devices that are easy to use and integrate into your code. plotcpp is aimed at everyday/casual plotting needs as well as visualisation and debugging.

How does it work?

plotcpp provides Figures, which are the most general visualisation abstraction. A Figure contains one or more plots, which in turn can be configured in many obvious ways and ultimately built/rendered. A plot can be a pie chart, a histogram, a function plot, etc.

Internally, plotcpp represents all visual elements using an SVG description. This allows plots to be scalable as well as to potentially support an infinity of visual styles while looking nice.

Contents

1. Requirements and dependencies
2. Building plotcpp
3. Using plotcpp

Requirements and dependencies

At the moment, plotcpp requires at least C++20. It is hard to find a balance between supporting a broader set of compiler standards and providing modern language features. As stated before, this library is targeted for use in modern C++ projects.

Dependencies

plotcpp is based on the following dependencies:

• A C++20 compliant compiler.
• libxml2
• fmt: Neither clang nor gcc support <format> yet, so fmt is a temporary workaround.
• librsvg2
• GLFW

Building plotcpp

To build plotcpp you need to make sure you have all dependencies installed on your system (see dependencies). In Fedora, you can achieve this by running

sudo dnf install libxml2 libxml2-devel fmt-devel librsvg2 librsvg2-devel glfw3 glfw3-devel

Other distributions may provide these dependencies through different packages.

First, compile the shared library by running make

make

Lastly, install the headers and the shared library:

sudo make install

From this point, you can use plotcpp in your code, provided that you link your executable with libplotcpp.

Using plotcpp

Generating the documentation

plotcpp uses Doxygen for documentation. You only need to install Doxygen if you want to generate the documentation. To do that, just run

make doc

This will generate a series of html files located at plotcpp/doc/html. To access the documentation, open plotcpp/doc/html/index.html on your browser. These pages will be useful if you need to understand the plotcpp API to include plots in your program.

Using plotcpp in your program

In order to use plotcpp inside your program you will need to #include the plotcpp headers. For example, to include the Plot2D.hpp header you need the following:

#include <plotcpp/Plot2D.hpp>

To compile your program you will need to link the shared library you installed earlier. Use the -lplotcpp linker flag to achieve that. In some distributions, the libxml2 headers are not located at /usr/include/libxml2. Normally, the libxml2 package contains a helper binary that provides such information. Try running xml2-config.This is an example of a command to do all this:

clang++ -std=c++20 `xml2-config --cflags` -lplotcpp <sources> -o <target>

Supported figure types

• Plot2D
• BarPlot
• HistogramPlot
• GroupFigure

Plot2D

The Plot2D figure can plot single variable functions. A number of features are lacking for this figure type:

• Smart subdivision of axes

BarPlot

The BarPlot figure can represent (stacked) segments of data.

A number of features are lacking for this figure type:

• Smart subdivision of axes
• Annotated values

HistogramPlot

The HistogramPlot is a BarPlot that represents the histogram of a series of values using a custom number of bins.

Since this figure is based on the BarPlot it will be lacking the same features.

GroupFigure

A GroupFigure is simply a group of figures (or subplots). The number of rows and columns are specified as template arguments GroupFigure<rows, cols>. Subplots are added by calling Subplot():

plotcpp::Plot2D top;
plotcpp::Plot2D bottom;

// ...

plotcpp::GroupFigure<2, 1> group;
group.Subplot(&top, 0, 0);
group.Subplot(&bottom, 1, 0);
group.Build();

// ...

Subplots added to the group can still be modified independently, but be aware that calling Build() on the group will modify the size of the plot.

Planned features

• Annotated heatmap
• Pie chart