galaaz 0.4.9 → 0.4.10
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/README.md +798 -285
- data/blogs/galaaz_ggplot/galaaz_ggplot.Rmd +3 -12
- data/blogs/galaaz_ggplot/galaaz_ggplot.aux +5 -7
- data/blogs/galaaz_ggplot/galaaz_ggplot.html +69 -29
- data/blogs/galaaz_ggplot/galaaz_ggplot.pdf +0 -0
- data/blogs/galaaz_ggplot/galaaz_ggplot_files/figure-html/midwest_rb.png +0 -0
- data/blogs/galaaz_ggplot/galaaz_ggplot_files/figure-html/scatter_plot_rb.png +0 -0
- data/blogs/galaaz_ggplot/galaaz_ggplot_files/figure-latex/midwest_rb.pdf +0 -0
- data/blogs/galaaz_ggplot/galaaz_ggplot_files/figure-latex/scatter_plot_rb.pdf +0 -0
- data/blogs/galaaz_ggplot/midwest.Rmd +1 -9
- data/blogs/gknit/gknit.Rmd +37 -40
- data/blogs/gknit/gknit.html +32 -30
- data/blogs/gknit/gknit.md +36 -37
- data/blogs/gknit/gknit.pdf +0 -0
- data/blogs/gknit/gknit.tex +35 -37
- data/blogs/manual/manual.Rmd +548 -125
- data/blogs/manual/manual.html +509 -286
- data/blogs/manual/manual.md +798 -285
- data/blogs/manual/manual.pdf +0 -0
- data/blogs/manual/manual.tex +2816 -0
- data/blogs/manual/manual_files/figure-latex/diverging_bar.pdf +0 -0
- data/blogs/nse_dplyr/nse_dplyr.Rmd +240 -74
- data/blogs/nse_dplyr/nse_dplyr.html +191 -87
- data/blogs/nse_dplyr/nse_dplyr.md +361 -107
- data/blogs/nse_dplyr/nse_dplyr.pdf +0 -0
- data/blogs/nse_dplyr/nse_dplyr.tex +1373 -0
- data/blogs/ruby_plot/ruby_plot.Rmd +61 -81
- data/blogs/ruby_plot/ruby_plot.html +54 -57
- data/blogs/ruby_plot/ruby_plot.md +48 -67
- data/blogs/ruby_plot/ruby_plot.pdf +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-html/dose_len.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-html/facet_by_delivery.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-html/facet_by_dose.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-html/facets_by_delivery_color.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-html/facets_by_delivery_color2.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-html/facets_with_jitter.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-html/facets_with_points.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-html/final_box_plot.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-html/final_violin_plot.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-html/violin_with_jitter.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-latex/dose_len.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-latex/facet_by_delivery.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-latex/facet_by_dose.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-latex/facets_by_delivery_color.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-latex/facets_by_delivery_color2.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-latex/facets_with_decorations.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-latex/facets_with_jitter.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-latex/facets_with_points.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-latex/final_box_plot.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-latex/final_violin_plot.png +0 -0
- data/blogs/ruby_plot/ruby_plot_files/figure-latex/violin_with_jitter.png +0 -0
- data/lib/R_interface/rdata_frame.rb +0 -12
- data/lib/R_interface/robject.rb +14 -14
- data/lib/R_interface/ruby_extensions.rb +3 -31
- data/lib/R_interface/rvector.rb +0 -12
- data/lib/gknit/knitr_engine.rb +5 -3
- data/lib/util/exec_ruby.rb +22 -61
- data/specs/tmp.rb +26 -12
- data/version.rb +1 -1
- metadata +22 -17
- data/bin/gknit_old_r +0 -236
- data/blogs/dev/dev.Rmd +0 -23
- data/blogs/dev/dev.md +0 -58
- data/blogs/dev/dev2.Rmd +0 -65
- data/blogs/dev/model.rb +0 -41
- data/blogs/dplyr/dplyr.Rmd +0 -29
- data/blogs/dplyr/dplyr.html +0 -433
- data/blogs/dplyr/dplyr.md +0 -58
- data/blogs/dplyr/dplyr.rb +0 -63
- data/blogs/galaaz_ggplot/galaaz_ggplot.log +0 -640
- data/blogs/galaaz_ggplot/galaaz_ggplot.md +0 -431
- data/blogs/galaaz_ggplot/galaaz_ggplot.tex +0 -481
- data/blogs/galaaz_ggplot/midwest.png +0 -0
- data/blogs/galaaz_ggplot/scatter_plot.png +0 -0
- data/blogs/ruby_plot/ruby_plot.tex +0 -1077
Binary file
|
@@ -0,0 +1,1373 @@
|
|
1
|
+
\documentclass[11pt,]{article}
|
2
|
+
\usepackage{lmodern}
|
3
|
+
\usepackage{amssymb,amsmath}
|
4
|
+
\usepackage{ifxetex,ifluatex}
|
5
|
+
\usepackage{fixltx2e} % provides \textsubscript
|
6
|
+
\ifnum 0\ifxetex 1\fi\ifluatex 1\fi=0 % if pdftex
|
7
|
+
\usepackage[T1]{fontenc}
|
8
|
+
\usepackage[utf8]{inputenc}
|
9
|
+
\else % if luatex or xelatex
|
10
|
+
\ifxetex
|
11
|
+
\usepackage{mathspec}
|
12
|
+
\else
|
13
|
+
\usepackage{fontspec}
|
14
|
+
\fi
|
15
|
+
\defaultfontfeatures{Ligatures=TeX,Scale=MatchLowercase}
|
16
|
+
\fi
|
17
|
+
% use upquote if available, for straight quotes in verbatim environments
|
18
|
+
\IfFileExists{upquote.sty}{\usepackage{upquote}}{}
|
19
|
+
% use microtype if available
|
20
|
+
\IfFileExists{microtype.sty}{%
|
21
|
+
\usepackage{microtype}
|
22
|
+
\UseMicrotypeSet[protrusion]{basicmath} % disable protrusion for tt fonts
|
23
|
+
}{}
|
24
|
+
\usepackage[margin=1in]{geometry}
|
25
|
+
\usepackage{hyperref}
|
26
|
+
\hypersetup{unicode=true,
|
27
|
+
pdftitle={Non Standard Evaluation in dplyr with Galaaz},
|
28
|
+
pdfauthor={Rodrigo Botafogo; Daniel Mossé - University of Pittsburgh},
|
29
|
+
pdfborder={0 0 0},
|
30
|
+
breaklinks=true}
|
31
|
+
\urlstyle{same} % don't use monospace font for urls
|
32
|
+
\usepackage{color}
|
33
|
+
\usepackage{fancyvrb}
|
34
|
+
\newcommand{\VerbBar}{|}
|
35
|
+
\newcommand{\VERB}{\Verb[commandchars=\\\{\}]}
|
36
|
+
\DefineVerbatimEnvironment{Highlighting}{Verbatim}{commandchars=\\\{\}}
|
37
|
+
% Add ',fontsize=\small' for more characters per line
|
38
|
+
\usepackage{framed}
|
39
|
+
\definecolor{shadecolor}{RGB}{248,248,248}
|
40
|
+
\newenvironment{Shaded}{\begin{snugshade}}{\end{snugshade}}
|
41
|
+
\newcommand{\AlertTok}[1]{\textcolor[rgb]{0.94,0.16,0.16}{#1}}
|
42
|
+
\newcommand{\AnnotationTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textbf{\textit{#1}}}}
|
43
|
+
\newcommand{\AttributeTok}[1]{\textcolor[rgb]{0.77,0.63,0.00}{#1}}
|
44
|
+
\newcommand{\BaseNTok}[1]{\textcolor[rgb]{0.00,0.00,0.81}{#1}}
|
45
|
+
\newcommand{\BuiltInTok}[1]{#1}
|
46
|
+
\newcommand{\CharTok}[1]{\textcolor[rgb]{0.31,0.60,0.02}{#1}}
|
47
|
+
\newcommand{\CommentTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textit{#1}}}
|
48
|
+
\newcommand{\CommentVarTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textbf{\textit{#1}}}}
|
49
|
+
\newcommand{\ConstantTok}[1]{\textcolor[rgb]{0.00,0.00,0.00}{#1}}
|
50
|
+
\newcommand{\ControlFlowTok}[1]{\textcolor[rgb]{0.13,0.29,0.53}{\textbf{#1}}}
|
51
|
+
\newcommand{\DataTypeTok}[1]{\textcolor[rgb]{0.13,0.29,0.53}{#1}}
|
52
|
+
\newcommand{\DecValTok}[1]{\textcolor[rgb]{0.00,0.00,0.81}{#1}}
|
53
|
+
\newcommand{\DocumentationTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textbf{\textit{#1}}}}
|
54
|
+
\newcommand{\ErrorTok}[1]{\textcolor[rgb]{0.64,0.00,0.00}{\textbf{#1}}}
|
55
|
+
\newcommand{\ExtensionTok}[1]{#1}
|
56
|
+
\newcommand{\FloatTok}[1]{\textcolor[rgb]{0.00,0.00,0.81}{#1}}
|
57
|
+
\newcommand{\FunctionTok}[1]{\textcolor[rgb]{0.00,0.00,0.00}{#1}}
|
58
|
+
\newcommand{\ImportTok}[1]{#1}
|
59
|
+
\newcommand{\InformationTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textbf{\textit{#1}}}}
|
60
|
+
\newcommand{\KeywordTok}[1]{\textcolor[rgb]{0.13,0.29,0.53}{\textbf{#1}}}
|
61
|
+
\newcommand{\NormalTok}[1]{#1}
|
62
|
+
\newcommand{\OperatorTok}[1]{\textcolor[rgb]{0.81,0.36,0.00}{\textbf{#1}}}
|
63
|
+
\newcommand{\OtherTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{#1}}
|
64
|
+
\newcommand{\PreprocessorTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textit{#1}}}
|
65
|
+
\newcommand{\RegionMarkerTok}[1]{#1}
|
66
|
+
\newcommand{\SpecialCharTok}[1]{\textcolor[rgb]{0.00,0.00,0.00}{#1}}
|
67
|
+
\newcommand{\SpecialStringTok}[1]{\textcolor[rgb]{0.31,0.60,0.02}{#1}}
|
68
|
+
\newcommand{\StringTok}[1]{\textcolor[rgb]{0.31,0.60,0.02}{#1}}
|
69
|
+
\newcommand{\VariableTok}[1]{\textcolor[rgb]{0.00,0.00,0.00}{#1}}
|
70
|
+
\newcommand{\VerbatimStringTok}[1]{\textcolor[rgb]{0.31,0.60,0.02}{#1}}
|
71
|
+
\newcommand{\WarningTok}[1]{\textcolor[rgb]{0.56,0.35,0.01}{\textbf{\textit{#1}}}}
|
72
|
+
\usepackage{graphicx,grffile}
|
73
|
+
\makeatletter
|
74
|
+
\def\maxwidth{\ifdim\Gin@nat@width>\linewidth\linewidth\else\Gin@nat@width\fi}
|
75
|
+
\def\maxheight{\ifdim\Gin@nat@height>\textheight\textheight\else\Gin@nat@height\fi}
|
76
|
+
\makeatother
|
77
|
+
% Scale images if necessary, so that they will not overflow the page
|
78
|
+
% margins by default, and it is still possible to overwrite the defaults
|
79
|
+
% using explicit options in \includegraphics[width, height, ...]{}
|
80
|
+
\setkeys{Gin}{width=\maxwidth,height=\maxheight,keepaspectratio}
|
81
|
+
\IfFileExists{parskip.sty}{%
|
82
|
+
\usepackage{parskip}
|
83
|
+
}{% else
|
84
|
+
\setlength{\parindent}{0pt}
|
85
|
+
\setlength{\parskip}{6pt plus 2pt minus 1pt}
|
86
|
+
}
|
87
|
+
\setlength{\emergencystretch}{3em} % prevent overfull lines
|
88
|
+
\providecommand{\tightlist}{%
|
89
|
+
\setlength{\itemsep}{0pt}\setlength{\parskip}{0pt}}
|
90
|
+
\setcounter{secnumdepth}{5}
|
91
|
+
% Redefines (sub)paragraphs to behave more like sections
|
92
|
+
\ifx\paragraph\undefined\else
|
93
|
+
\let\oldparagraph\paragraph
|
94
|
+
\renewcommand{\paragraph}[1]{\oldparagraph{#1}\mbox{}}
|
95
|
+
\fi
|
96
|
+
\ifx\subparagraph\undefined\else
|
97
|
+
\let\oldsubparagraph\subparagraph
|
98
|
+
\renewcommand{\subparagraph}[1]{\oldsubparagraph{#1}\mbox{}}
|
99
|
+
\fi
|
100
|
+
|
101
|
+
%%% Use protect on footnotes to avoid problems with footnotes in titles
|
102
|
+
\let\rmarkdownfootnote\footnote%
|
103
|
+
\def\footnote{\protect\rmarkdownfootnote}
|
104
|
+
|
105
|
+
%%% Change title format to be more compact
|
106
|
+
\usepackage{titling}
|
107
|
+
|
108
|
+
% Create subtitle command for use in maketitle
|
109
|
+
\newcommand{\subtitle}[1]{
|
110
|
+
\posttitle{
|
111
|
+
\begin{center}\large#1\end{center}
|
112
|
+
}
|
113
|
+
}
|
114
|
+
|
115
|
+
\setlength{\droptitle}{-2em}
|
116
|
+
|
117
|
+
\title{Non Standard Evaluation in dplyr with Galaaz}
|
118
|
+
\pretitle{\vspace{\droptitle}\centering\huge}
|
119
|
+
\posttitle{\par}
|
120
|
+
\author{Rodrigo Botafogo \\ Daniel Mossé - University of Pittsburgh}
|
121
|
+
\preauthor{\centering\large\emph}
|
122
|
+
\postauthor{\par}
|
123
|
+
\predate{\centering\large\emph}
|
124
|
+
\postdate{\par}
|
125
|
+
\date{10/05/2019}
|
126
|
+
|
127
|
+
% usar portugues do Brasil
|
128
|
+
% \usepackage[brazilian]{babel}
|
129
|
+
\usepackage[utf8]{inputenc}
|
130
|
+
|
131
|
+
\usepackage{geometry}
|
132
|
+
\geometry{a4paper, top=1in}
|
133
|
+
|
134
|
+
% needed for kableExtra
|
135
|
+
\usepackage{longtable}
|
136
|
+
\usepackage{multirow}
|
137
|
+
\usepackage[table]{xcolor}
|
138
|
+
\usepackage{wrapfig}
|
139
|
+
\usepackage{float}
|
140
|
+
\usepackage{colortbl}
|
141
|
+
\usepackage{pdflscape}
|
142
|
+
\usepackage{tabu}
|
143
|
+
\usepackage{threeparttable}
|
144
|
+
\usepackage[normalem]{ulem}
|
145
|
+
|
146
|
+
\usepackage{bbm}
|
147
|
+
\usepackage{booktabs}
|
148
|
+
\usepackage{expex}
|
149
|
+
|
150
|
+
\usepackage{graphicx}
|
151
|
+
|
152
|
+
\usepackage{fancyhdr}
|
153
|
+
% set the header and foot style
|
154
|
+
% style 'fancy' adds the section name on the header
|
155
|
+
% and the page number on the footer
|
156
|
+
\pagestyle{fancy}
|
157
|
+
|
158
|
+
% style 'fancyhf' leaves header and footer empty
|
159
|
+
%\fancyhf{}
|
160
|
+
|
161
|
+
% sets the left head element to \rightmark, which contains the
|
162
|
+
% current section (\leftmark is the current chapter)
|
163
|
+
%\fancyhead[L]{\rightmark} .
|
164
|
+
|
165
|
+
% sets the right head element to the page number.
|
166
|
+
% \fancyhead[R]{\thepage}
|
167
|
+
|
168
|
+
% lets the head rule disappear.
|
169
|
+
% \renewcommand{\headrulewidth}{0pt}
|
170
|
+
% Possible selectors for the optional argument of \fancyhead/\fancyfoot
|
171
|
+
% are L (left), C (center) or R (right) for the position of the element
|
172
|
+
% and E (even) or O (odd) to distinguish even and odd pages. If you omit
|
173
|
+
% E/O the element is set for all pages.
|
174
|
+
|
175
|
+
% \usepackage{lipsum}
|
176
|
+
|
177
|
+
% make available command lastpage
|
178
|
+
\usepackage{lastpage}
|
179
|
+
|
180
|
+
% default fontsize 11pt better to add
|
181
|
+
% fontsize on the yaml header
|
182
|
+
% \usepackage[fontsize=11pt]{scrextend}
|
183
|
+
|
184
|
+
% comandos para formatar uma tabela
|
185
|
+
\usepackage{array}
|
186
|
+
\newcolumntype{L}[1]{>{\raggedright\let\newline\\\arraybackslash\hspace{0pt}}m{#1}}
|
187
|
+
\newcolumntype{C}[1]{>{\centering\let\newline\\\arraybackslash\hspace{0pt}}m{#1}}
|
188
|
+
\newcolumntype{R}[1]{>{\raggedleft\let\newline\\\arraybackslash\hspace{0pt}}m{#1}}
|
189
|
+
|
190
|
+
% necessário if we need to import other latex documents
|
191
|
+
\usepackage{import}
|
192
|
+
|
193
|
+
% Command to import an R variable to latex
|
194
|
+
\newcommand{\RtoLatex}[2]{\newcommand{#1}{#2}}
|
195
|
+
|
196
|
+
%
|
197
|
+
%\newcommand{\atraso}[1]{\color{red} \textbf {Tempo desde a Assinatura do Contrato: #1 dias}}
|
198
|
+
|
199
|
+
\begin{document}
|
200
|
+
\maketitle
|
201
|
+
|
202
|
+
{
|
203
|
+
\setcounter{tocdepth}{2}
|
204
|
+
\tableofcontents
|
205
|
+
}
|
206
|
+
\hypertarget{introduction}{%
|
207
|
+
\section{Introduction}\label{introduction}}
|
208
|
+
|
209
|
+
In this post we will see how to program with \emph{dplyr} in Galaaz.
|
210
|
+
|
211
|
+
\hypertarget{but-first-what-is-galaaz}{%
|
212
|
+
\subsection{But first, what is
|
213
|
+
Galaaz??}\label{but-first-what-is-galaaz}}
|
214
|
+
|
215
|
+
Galaaz is a system for tightly coupling Ruby and R. Ruby is a powerful
|
216
|
+
language, with a large community, a very large set of libraries and
|
217
|
+
great for web development. However, it lacks libraries for data science,
|
218
|
+
statistics, scientific plotting and machine learning. On the other hand,
|
219
|
+
R is considered one of the most powerful languages for solving all of
|
220
|
+
the above problems. Maybe the strongest competitor to R is Python with
|
221
|
+
libraries such as NumPy, Pandas, SciPy, SciKit-Learn and many more.
|
222
|
+
|
223
|
+
With Galaaz we do not intend to re-implement any of the scientific
|
224
|
+
libraries in R. However, we allow for very tight coupling between the
|
225
|
+
two languages to the point that the Ruby developer does not need to know
|
226
|
+
that there is an R engine running. Also, from the point of view of the R
|
227
|
+
user/developer Galaaz looks a lot like R, with just minor syntactic
|
228
|
+
difference, so there is almost no learning courve for the R developer.
|
229
|
+
And as we will see in this post, programming with \emph{dplyr} is easier
|
230
|
+
in Galaaz than in R.
|
231
|
+
|
232
|
+
R users are probably quite knowledgeable about \emph{dplyr}, for the
|
233
|
+
Ruby developer, \emph{dplyr} and the \emph{tidyverse} libraries are a
|
234
|
+
set of libraries for data manipulation in R, developed by Hardley
|
235
|
+
Wickham, chief scientis at RStudio and a prolific R coder and writer.
|
236
|
+
|
237
|
+
For the coupling of Ruby and R we use new technologies provided by
|
238
|
+
Oracle: GraalVM, TruffleRuby and FastR:
|
239
|
+
|
240
|
+
\begin{verbatim}
|
241
|
+
GraalVM is a universal virtual machine for running applications
|
242
|
+
written in JavaScript, Python 3, Ruby, R, JVM-based languages like Java,
|
243
|
+
Scala, Kotlin, and LLVM-based languages such as C and C++.
|
244
|
+
|
245
|
+
GraalVM removes the isolation between programming languages and enables
|
246
|
+
interoperability in a shared runtime. It can run either standalone or in
|
247
|
+
the context of OpenJDK, Node.js, Oracle Database, or MySQL.
|
248
|
+
|
249
|
+
GraalVM allows you to write polyglot applications with a seamless way to
|
250
|
+
pass values from one language to another. With GraalVM there is no copying
|
251
|
+
or marshaling necessary as it is with other polyglot systems. This lets
|
252
|
+
you achieve high performance when language boundaries are crossed. Most
|
253
|
+
of the time there is no additional cost for crossing a language boundary
|
254
|
+
at all.
|
255
|
+
|
256
|
+
Often developers have to make uncomfortable compromises that require them
|
257
|
+
to rewrite their software in other languages. For example:
|
258
|
+
|
259
|
+
* “That library is not available in my language. I need to rewrite it.”
|
260
|
+
* “That language would be the perfect fit for my problem, but we cannot
|
261
|
+
run it in our environment.”
|
262
|
+
* “That problem is already solved in my language, but the language is
|
263
|
+
too slow.”
|
264
|
+
|
265
|
+
With GraalVM we aim to allow developers to freely choose the right language
|
266
|
+
for the task at hand without making compromises.
|
267
|
+
\end{verbatim}
|
268
|
+
|
269
|
+
Interested readers should also check out the following sites:
|
270
|
+
|
271
|
+
\begin{itemize}
|
272
|
+
\tightlist
|
273
|
+
\item
|
274
|
+
\href{https://www.graalvm.org/}{GraalVM Home}
|
275
|
+
\item
|
276
|
+
\href{https://github.com/oracle/truffleruby}{TruffleRuby}
|
277
|
+
\item
|
278
|
+
\href{https://github.com/oracle/fastr}{FastR}
|
279
|
+
\item
|
280
|
+
\href{https://medium.com/graalvm/faster-r-with-fastr-4b8db0e0dceb}{Faster
|
281
|
+
R with FastR}
|
282
|
+
\item
|
283
|
+
\href{https://medium.freecodecamp.org/how-to-make-beautiful-ruby-plots-with-galaaz-320848058857}{How
|
284
|
+
to make Beautiful Ruby Plots with Galaaz}
|
285
|
+
\item
|
286
|
+
\href{https://towardsdatascience.com/ruby-plotting-with-galaaz-an-example-of-tightly-coupling-ruby-and-r-in-graalvm-520b69e21021}{Ruby
|
287
|
+
Plotting with Galaaz: An example of tightly coupling Ruby and R in
|
288
|
+
GraalVM}
|
289
|
+
\item
|
290
|
+
\href{https://towardsdatascience.com/how-to-do-reproducible-research-in-ruby-with-gknit-c26d2684d64e}{How
|
291
|
+
to do reproducible research in Ruby with gKnit}
|
292
|
+
\item
|
293
|
+
\href{https://r4ds.had.co.nz/}{R for Data Science}
|
294
|
+
\item
|
295
|
+
\href{https://adv-r.hadley.nz/}{Advanced R}
|
296
|
+
\end{itemize}
|
297
|
+
|
298
|
+
\hypertarget{programming-with-dplyr}{%
|
299
|
+
\subsection{Programming with dplyr}\label{programming-with-dplyr}}
|
300
|
+
|
301
|
+
This post will follow closely the work done in
|
302
|
+
\url{https://dplyr.tidyverse.org/articles/programming.html}, by Hardley
|
303
|
+
Wickham. In it, Hardley states:
|
304
|
+
|
305
|
+
\begin{quote}
|
306
|
+
Most dplyr functions use non-standard evaluation (NSE). This is a
|
307
|
+
catch-all term that means they don't follow the usual R rules of
|
308
|
+
evaluation. Instead, they capture the expression that you typed and
|
309
|
+
evaluate it in a custom way. This has two main benefits for dplyr code:
|
310
|
+
\end{quote}
|
311
|
+
|
312
|
+
\begin{quote}
|
313
|
+
Operations on data frames can be expressed succinctly because you don't
|
314
|
+
need to repeat the name of the data frame. For example, you can write
|
315
|
+
filter(df, x == 1, y == 2, z == 3) instead of df{[}df\$x == 1 \& df\$y
|
316
|
+
==2 \& df\$z == 3, {]}.
|
317
|
+
\end{quote}
|
318
|
+
|
319
|
+
\begin{quote}
|
320
|
+
dplyr can choose to compute results in a different way to base R. This
|
321
|
+
is important for database backends because dplyr itself doesn't do any
|
322
|
+
work, but instead generates the SQL that tells the database what to do.
|
323
|
+
\end{quote}
|
324
|
+
|
325
|
+
\begin{quote}
|
326
|
+
Unfortunately these benefits do not come for free. There are two main
|
327
|
+
drawbacks:
|
328
|
+
\end{quote}
|
329
|
+
|
330
|
+
\begin{quote}
|
331
|
+
Most dplyr arguments are not referentially transparent. That means you
|
332
|
+
can't replace a value with a seemingly equivalent object that you've
|
333
|
+
defined elsewhere. In other words, this code:
|
334
|
+
\end{quote}
|
335
|
+
|
336
|
+
\begin{Shaded}
|
337
|
+
\begin{Highlighting}[]
|
338
|
+
\NormalTok{df <-}\StringTok{ }\KeywordTok{data.frame}\NormalTok{(}\DataTypeTok{x =} \DecValTok{1}\OperatorTok{:}\DecValTok{3}\NormalTok{, }\DataTypeTok{y =} \DecValTok{3}\OperatorTok{:}\DecValTok{1}\NormalTok{)}
|
339
|
+
\KeywordTok{print}\NormalTok{(}\KeywordTok{filter}\NormalTok{(df, x }\OperatorTok{==}\StringTok{ }\DecValTok{1}\NormalTok{))}
|
340
|
+
\CommentTok{#> # A tibble: 1 x 2}
|
341
|
+
\CommentTok{#> x y}
|
342
|
+
\CommentTok{#> <int> <int>}
|
343
|
+
\CommentTok{#> 1 1 3}
|
344
|
+
\end{Highlighting}
|
345
|
+
\end{Shaded}
|
346
|
+
|
347
|
+
\begin{quote}
|
348
|
+
Is not equivalent to this code:
|
349
|
+
\end{quote}
|
350
|
+
|
351
|
+
\begin{Shaded}
|
352
|
+
\begin{Highlighting}[]
|
353
|
+
\NormalTok{my_var <-}\StringTok{ }\NormalTok{x}
|
354
|
+
\CommentTok{#> Error in eval(expr, envir, enclos): object 'x' not found}
|
355
|
+
\KeywordTok{filter}\NormalTok{(df, my_var }\OperatorTok{==}\StringTok{ }\DecValTok{1}\NormalTok{)}
|
356
|
+
\CommentTok{#> Error: object 'my_var' not found}
|
357
|
+
\end{Highlighting}
|
358
|
+
\end{Shaded}
|
359
|
+
|
360
|
+
\begin{quote}
|
361
|
+
This makes it hard to create functions with arguments that change how
|
362
|
+
dplyr verbs are computed.
|
363
|
+
\end{quote}
|
364
|
+
|
365
|
+
In this post we will see that programming with \emph{dplyr} in Galaaz
|
366
|
+
does not require knowledge of non-standard evaluation in R and can be
|
367
|
+
accomplished by utilizing normal Ruby constructs.
|
368
|
+
|
369
|
+
\hypertarget{writing-expressions-in-galaaz}{%
|
370
|
+
\section{Writing Expressions in
|
371
|
+
Galaaz}\label{writing-expressions-in-galaaz}}
|
372
|
+
|
373
|
+
Galaaz extends Ruby to work with expressions, similar to R's expressions
|
374
|
+
build with `quote' (base R) or `quo' (tidyverse). Expressions in this
|
375
|
+
context are like mathematical expressions or formulae. For instance, in
|
376
|
+
mathematics, the expression \(y = sin(x)\) describes a function but
|
377
|
+
cannot be computed unless the value of \(x\) is bound to some value.
|
378
|
+
|
379
|
+
Let's take a look at some of those expressions in Ruby:
|
380
|
+
|
381
|
+
\hypertarget{expressions-from-operators}{%
|
382
|
+
\subsection{Expressions from
|
383
|
+
operators}\label{expressions-from-operators}}
|
384
|
+
|
385
|
+
The code bellow creates an expression summing two symbols. Note that :a
|
386
|
+
and :b are Ruby symbols and are not bound to any value at the time of
|
387
|
+
expression definition:
|
388
|
+
|
389
|
+
\begin{Shaded}
|
390
|
+
\begin{Highlighting}[]
|
391
|
+
\NormalTok{exp1 = }\StringTok{:a}\NormalTok{ + }\StringTok{:b}
|
392
|
+
\NormalTok{puts exp1}
|
393
|
+
\end{Highlighting}
|
394
|
+
\end{Shaded}
|
395
|
+
|
396
|
+
\begin{verbatim}
|
397
|
+
## a + b
|
398
|
+
\end{verbatim}
|
399
|
+
|
400
|
+
We can build any complex mathematical expression such as:
|
401
|
+
|
402
|
+
\begin{Shaded}
|
403
|
+
\begin{Highlighting}[]
|
404
|
+
\NormalTok{exp2 = (}\StringTok{:a}\NormalTok{ + }\StringTok{:b}\NormalTok{) * }\FloatTok{2.0}\NormalTok{ + }\StringTok{:c}\NormalTok{ ** }\DecValTok{2}\NormalTok{ / }\StringTok{:z}
|
405
|
+
\NormalTok{puts exp2}
|
406
|
+
\end{Highlighting}
|
407
|
+
\end{Shaded}
|
408
|
+
|
409
|
+
\begin{verbatim}
|
410
|
+
## (a + b) * 2 + c^2L/z
|
411
|
+
\end{verbatim}
|
412
|
+
|
413
|
+
The `L' after two indicates that 2 is an integer.
|
414
|
+
|
415
|
+
It is also possible to use inequality operators in building expressions:
|
416
|
+
|
417
|
+
\begin{Shaded}
|
418
|
+
\begin{Highlighting}[]
|
419
|
+
\NormalTok{exp3 = (}\StringTok{:a}\NormalTok{ + }\StringTok{:b}\NormalTok{) >= }\StringTok{:z}
|
420
|
+
\NormalTok{puts exp3}
|
421
|
+
\end{Highlighting}
|
422
|
+
\end{Shaded}
|
423
|
+
|
424
|
+
\begin{verbatim}
|
425
|
+
## a + b >= z
|
426
|
+
\end{verbatim}
|
427
|
+
|
428
|
+
Expressions' definition can also make use of normal Ruby variables
|
429
|
+
without any problem:
|
430
|
+
|
431
|
+
\begin{Shaded}
|
432
|
+
\begin{Highlighting}[]
|
433
|
+
\NormalTok{x = }\DecValTok{20}
|
434
|
+
\NormalTok{y = }\DecValTok{30}
|
435
|
+
\NormalTok{exp_var = (}\StringTok{:a}\NormalTok{ + }\StringTok{:b}\NormalTok{) * x <= }\StringTok{:z}\NormalTok{ - y}
|
436
|
+
\NormalTok{puts exp_var}
|
437
|
+
\end{Highlighting}
|
438
|
+
\end{Shaded}
|
439
|
+
|
440
|
+
\begin{verbatim}
|
441
|
+
## (a + b) * 20L <= z - 30L
|
442
|
+
\end{verbatim}
|
443
|
+
|
444
|
+
Galaaz provides both symbolic representations for operators, such as
|
445
|
+
(\textgreater{}, \textless{}, !=) as functional notation for those
|
446
|
+
operators such as (.gt, .ge, etc.). So the same expression written above
|
447
|
+
can also be written as
|
448
|
+
|
449
|
+
\begin{Shaded}
|
450
|
+
\begin{Highlighting}[]
|
451
|
+
\NormalTok{exp4 = (}\StringTok{:a}\NormalTok{ + }\StringTok{:b}\NormalTok{).ge }\StringTok{:z}
|
452
|
+
\NormalTok{puts exp4}
|
453
|
+
\end{Highlighting}
|
454
|
+
\end{Shaded}
|
455
|
+
|
456
|
+
\begin{verbatim}
|
457
|
+
## a + b >= z
|
458
|
+
\end{verbatim}
|
459
|
+
|
460
|
+
Two type of expression, however, can only be created with the functional
|
461
|
+
representation of the operators, those are expressions involving `==',
|
462
|
+
and `='. In order to write an expression involving `==' we need to use
|
463
|
+
the method `.eq' and for `=' we need the function `.assign'
|
464
|
+
|
465
|
+
\begin{Shaded}
|
466
|
+
\begin{Highlighting}[]
|
467
|
+
\NormalTok{exp5 = (}\StringTok{:a}\NormalTok{ + }\StringTok{:b}\NormalTok{).eq }\StringTok{:z}
|
468
|
+
\NormalTok{puts exp5}
|
469
|
+
\end{Highlighting}
|
470
|
+
\end{Shaded}
|
471
|
+
|
472
|
+
\begin{verbatim}
|
473
|
+
## a + b == z
|
474
|
+
\end{verbatim}
|
475
|
+
|
476
|
+
\begin{Shaded}
|
477
|
+
\begin{Highlighting}[]
|
478
|
+
\NormalTok{exp6 = }\StringTok{:y}\NormalTok{.assign }\StringTok{:a}\NormalTok{ + }\StringTok{:b}
|
479
|
+
\NormalTok{puts exp6}
|
480
|
+
\end{Highlighting}
|
481
|
+
\end{Shaded}
|
482
|
+
|
483
|
+
\begin{verbatim}
|
484
|
+
## y <- a + b
|
485
|
+
\end{verbatim}
|
486
|
+
|
487
|
+
In general we think that using the functional notation is preferable to
|
488
|
+
using the symbolic notation as otherwise, we end up writing invalid
|
489
|
+
expressions such as
|
490
|
+
|
491
|
+
\begin{Shaded}
|
492
|
+
\begin{Highlighting}[]
|
493
|
+
\NormalTok{exp_wrong = (}\StringTok{:a}\NormalTok{ + }\StringTok{:b}\NormalTok{) == }\StringTok{:z}
|
494
|
+
\NormalTok{puts exp_wrong}
|
495
|
+
\end{Highlighting}
|
496
|
+
\end{Shaded}
|
497
|
+
|
498
|
+
\begin{verbatim}
|
499
|
+
## Message:
|
500
|
+
## Error in function (x, y, num.eq = TRUE, single.NA = TRUE, attrib.as.set = TRUE, :
|
501
|
+
## object 'a' not found (RError)
|
502
|
+
## Translated to internal error
|
503
|
+
\end{verbatim}
|
504
|
+
|
505
|
+
and it might be difficult to understand what is going on here. The
|
506
|
+
problem lies with the fact that when using `==' we are comparing
|
507
|
+
expression (:a + :b) to expression :z with `=='. When the comparison is
|
508
|
+
executed, the system tries to evaluate :a, :b and :z, and those symbols
|
509
|
+
at this time are not bound to anything and we get a ``object `a' not
|
510
|
+
found'' message. If we only use functional notation, this type of error
|
511
|
+
will not occur.
|
512
|
+
|
513
|
+
\hypertarget{expressions-with-r-methods}{%
|
514
|
+
\subsection{Expressions with R
|
515
|
+
methods}\label{expressions-with-r-methods}}
|
516
|
+
|
517
|
+
It is often necessary to create an expression that uses a method or
|
518
|
+
function. For instance, in mathematics, it's quite natural to write an
|
519
|
+
expressin such as \(y = sin(x)\). In this case, the `sin' function is
|
520
|
+
part of the expression and should not immediately be executed. When we
|
521
|
+
want the function to be part of the expression, we call the function
|
522
|
+
preceeding it by the letter E, such as `E.sin(x)'
|
523
|
+
|
524
|
+
\begin{Shaded}
|
525
|
+
\begin{Highlighting}[]
|
526
|
+
\NormalTok{exp7 = }\StringTok{:y}\NormalTok{.assign E.sin(}\StringTok{:x}\NormalTok{)}
|
527
|
+
\NormalTok{puts exp7}
|
528
|
+
\end{Highlighting}
|
529
|
+
\end{Shaded}
|
530
|
+
|
531
|
+
\begin{verbatim}
|
532
|
+
## y <- sin(x)
|
533
|
+
\end{verbatim}
|
534
|
+
|
535
|
+
Expressions can also be written using `.' notation:
|
536
|
+
|
537
|
+
\begin{Shaded}
|
538
|
+
\begin{Highlighting}[]
|
539
|
+
\NormalTok{exp8 = }\StringTok{:y}\NormalTok{.assign }\StringTok{:x}\NormalTok{.sin}
|
540
|
+
\NormalTok{puts exp8}
|
541
|
+
\end{Highlighting}
|
542
|
+
\end{Shaded}
|
543
|
+
|
544
|
+
\begin{verbatim}
|
545
|
+
## y <- sin(x)
|
546
|
+
\end{verbatim}
|
547
|
+
|
548
|
+
When a function has multiple arguments, the first one can be used before
|
549
|
+
the `.':
|
550
|
+
|
551
|
+
\begin{Shaded}
|
552
|
+
\begin{Highlighting}[]
|
553
|
+
\NormalTok{exp9 = }\StringTok{:x}\NormalTok{.c(}\StringTok{:y}\NormalTok{)}
|
554
|
+
\NormalTok{puts exp9}
|
555
|
+
\end{Highlighting}
|
556
|
+
\end{Shaded}
|
557
|
+
|
558
|
+
\begin{verbatim}
|
559
|
+
## c(x, y)
|
560
|
+
\end{verbatim}
|
561
|
+
|
562
|
+
\hypertarget{evaluating-an-expression}{%
|
563
|
+
\subsection{Evaluating an Expression}\label{evaluating-an-expression}}
|
564
|
+
|
565
|
+
Expressions can be evaluated by calling function `eval' with a binding.
|
566
|
+
A binding can be provided with a list:
|
567
|
+
|
568
|
+
\begin{Shaded}
|
569
|
+
\begin{Highlighting}[]
|
570
|
+
\NormalTok{exp = (}\StringTok{:a}\NormalTok{ + }\StringTok{:b}\NormalTok{) * }\FloatTok{2.0}\NormalTok{ + }\StringTok{:c}\NormalTok{ ** }\DecValTok{2}\NormalTok{ / }\StringTok{:z}
|
571
|
+
\NormalTok{puts exp.eval(R.list(}\StringTok{a: }\DecValTok{10}\NormalTok{, }\StringTok{b: }\DecValTok{20}\NormalTok{, }\StringTok{c: }\DecValTok{30}\NormalTok{, }\StringTok{z: }\DecValTok{40}\NormalTok{))}
|
572
|
+
\end{Highlighting}
|
573
|
+
\end{Shaded}
|
574
|
+
|
575
|
+
\begin{verbatim}
|
576
|
+
## [1] 82.5
|
577
|
+
\end{verbatim}
|
578
|
+
|
579
|
+
\ldots{} with a data frame:
|
580
|
+
|
581
|
+
\begin{Shaded}
|
582
|
+
\begin{Highlighting}[]
|
583
|
+
\NormalTok{df = R.data__frame(}
|
584
|
+
\StringTok{a: }\NormalTok{R.c(}\DecValTok{1}\NormalTok{, }\DecValTok{2}\NormalTok{, }\DecValTok{3}\NormalTok{),}
|
585
|
+
\StringTok{b: }\NormalTok{R.c(}\DecValTok{10}\NormalTok{, }\DecValTok{20}\NormalTok{, }\DecValTok{30}\NormalTok{),}
|
586
|
+
\StringTok{c: }\NormalTok{R.c(}\DecValTok{100}\NormalTok{, }\DecValTok{200}\NormalTok{, }\DecValTok{300}\NormalTok{),}
|
587
|
+
\StringTok{z: }\NormalTok{R.c(}\DecValTok{1000}\NormalTok{, }\DecValTok{2000}\NormalTok{, }\DecValTok{3000}\NormalTok{))}
|
588
|
+
|
589
|
+
\NormalTok{puts exp.eval(df)}
|
590
|
+
\end{Highlighting}
|
591
|
+
\end{Shaded}
|
592
|
+
|
593
|
+
\begin{verbatim}
|
594
|
+
## [1] 32 64 96
|
595
|
+
\end{verbatim}
|
596
|
+
|
597
|
+
\hypertarget{using-galaaz-to-call-r-functions}{%
|
598
|
+
\section{Using Galaaz to call R
|
599
|
+
functions}\label{using-galaaz-to-call-r-functions}}
|
600
|
+
|
601
|
+
Galaaz tries to emulate as closely as possible the way R functions are
|
602
|
+
called and migrating from R to Galaaz should be quite easy requiring
|
603
|
+
only minor syntactic changes to an R script. In this post, we do not
|
604
|
+
have enough space to write a complete manual on Galaaz (a short manual
|
605
|
+
can be found at: \url{https://www.rubydoc.info/gems/galaaz/0.4.9}), so
|
606
|
+
we will present only a few examples scripts using Galaaz.
|
607
|
+
|
608
|
+
Basically, to call an R function from Ruby with Galaaz, one only needs
|
609
|
+
to preceed the function with `R.'. For instance, to create a vector in
|
610
|
+
R, the `c' function is used. From Galaaz, a vector can be created by
|
611
|
+
using `R.c':
|
612
|
+
|
613
|
+
\begin{Shaded}
|
614
|
+
\begin{Highlighting}[]
|
615
|
+
\NormalTok{vec = R.c(}\FloatTok{1.0}\NormalTok{, }\DecValTok{2}\NormalTok{, }\DecValTok{3}\NormalTok{)}
|
616
|
+
\NormalTok{puts vec}
|
617
|
+
\end{Highlighting}
|
618
|
+
\end{Shaded}
|
619
|
+
|
620
|
+
\begin{verbatim}
|
621
|
+
## [1] 1 2 3
|
622
|
+
\end{verbatim}
|
623
|
+
|
624
|
+
A list is created in R with the `list' function, so in Galaaz we do:
|
625
|
+
|
626
|
+
\begin{Shaded}
|
627
|
+
\begin{Highlighting}[]
|
628
|
+
\NormalTok{list = R.list(}\StringTok{a: }\FloatTok{1.0}\NormalTok{, }\StringTok{b: }\DecValTok{2}\NormalTok{, }\StringTok{c: }\DecValTok{3}\NormalTok{)}
|
629
|
+
\NormalTok{puts list}
|
630
|
+
\end{Highlighting}
|
631
|
+
\end{Shaded}
|
632
|
+
|
633
|
+
\begin{verbatim}
|
634
|
+
## $a
|
635
|
+
## [1] 1
|
636
|
+
##
|
637
|
+
## $b
|
638
|
+
## [1] 2
|
639
|
+
##
|
640
|
+
## $c
|
641
|
+
## [1] 3
|
642
|
+
\end{verbatim}
|
643
|
+
|
644
|
+
Note that we can use named arguments in our list. The same code in R
|
645
|
+
would be:
|
646
|
+
|
647
|
+
\begin{Shaded}
|
648
|
+
\begin{Highlighting}[]
|
649
|
+
\NormalTok{lst =}\StringTok{ }\KeywordTok{list}\NormalTok{(}\DataTypeTok{a =} \DecValTok{1}\NormalTok{, }\DataTypeTok{b =}\NormalTok{ 2L, }\DataTypeTok{c =}\NormalTok{ 3L)}
|
650
|
+
\KeywordTok{print}\NormalTok{(lst)}
|
651
|
+
\end{Highlighting}
|
652
|
+
\end{Shaded}
|
653
|
+
|
654
|
+
\begin{verbatim}
|
655
|
+
## $a
|
656
|
+
## [1] 1
|
657
|
+
##
|
658
|
+
## $b
|
659
|
+
## [1] 2
|
660
|
+
##
|
661
|
+
## $c
|
662
|
+
## [1] 3
|
663
|
+
\end{verbatim}
|
664
|
+
|
665
|
+
Now, let's say that `x' is an angle of 45\(^\circ\) and we acttually
|
666
|
+
want to create the expression \(y = sin(45^\circ)\), which is
|
667
|
+
\(y = 0.850...\). In this case, we will use `R.sin':
|
668
|
+
|
669
|
+
\begin{Shaded}
|
670
|
+
\begin{Highlighting}[]
|
671
|
+
\NormalTok{exp10 = }\StringTok{:y}\NormalTok{.assign R.sin(}\DecValTok{45}\NormalTok{)}
|
672
|
+
\NormalTok{puts exp10}
|
673
|
+
\end{Highlighting}
|
674
|
+
\end{Shaded}
|
675
|
+
|
676
|
+
\begin{verbatim}
|
677
|
+
## y <- 0.850903524534118
|
678
|
+
\end{verbatim}
|
679
|
+
|
680
|
+
\hypertarget{filtering-using-expressions}{%
|
681
|
+
\section{Filtering using
|
682
|
+
expressions}\label{filtering-using-expressions}}
|
683
|
+
|
684
|
+
Now that we know how to write expression and call R functions let's do
|
685
|
+
some data manipulation in Galaaz. Let's first start by creating the same
|
686
|
+
data frame that we created previously in section ``Programming with
|
687
|
+
dplyr'':
|
688
|
+
|
689
|
+
\begin{Shaded}
|
690
|
+
\begin{Highlighting}[]
|
691
|
+
\NormalTok{df = R.data__frame(}\StringTok{x: }\NormalTok{(}\DecValTok{1}\NormalTok{..}\DecValTok{3}\NormalTok{), }\StringTok{y: }\NormalTok{(}\DecValTok{3}\NormalTok{..}\DecValTok{1}\NormalTok{))}
|
692
|
+
\NormalTok{puts df}
|
693
|
+
\end{Highlighting}
|
694
|
+
\end{Shaded}
|
695
|
+
|
696
|
+
\begin{verbatim}
|
697
|
+
## x y
|
698
|
+
## 1 1 3
|
699
|
+
## 2 2 2
|
700
|
+
## 3 3 1
|
701
|
+
\end{verbatim}
|
702
|
+
|
703
|
+
The `filter' function can be called on this data frame either by using
|
704
|
+
`R.filter(df, \ldots{})' or by using dot notation. We prefer to use dot
|
705
|
+
notation as shown bellow. The argument to `filter' in Galaaz should be
|
706
|
+
an expression. Note that if we gave to filter a Ruby expression such as
|
707
|
+
`x == 1', we would get an error, since there is no variable `x' defined
|
708
|
+
and if `x' was a variable then `x == 1' would either be `true' or
|
709
|
+
`false'. Our goal is to filter our data frame returning all rows in
|
710
|
+
which the `x' value is equal to 1. To express this we want: `:x.eq 1',
|
711
|
+
where :x will be interpreted by filter as the `x' column.
|
712
|
+
|
713
|
+
\begin{Shaded}
|
714
|
+
\begin{Highlighting}[]
|
715
|
+
\NormalTok{puts df.filter(}\StringTok{:x}\NormalTok{.eq }\DecValTok{1}\NormalTok{)}
|
716
|
+
\end{Highlighting}
|
717
|
+
\end{Shaded}
|
718
|
+
|
719
|
+
\begin{verbatim}
|
720
|
+
## x y
|
721
|
+
## 1 1 3
|
722
|
+
\end{verbatim}
|
723
|
+
|
724
|
+
In R, and when coding with `tidyverse', arguments to a function are
|
725
|
+
usually not \emph{referencially transparent}. That is, you can't replace
|
726
|
+
a value with a seemingly equivalent object that you've defined
|
727
|
+
elsewhere. In other words, this code
|
728
|
+
|
729
|
+
\begin{Shaded}
|
730
|
+
\begin{Highlighting}[]
|
731
|
+
\NormalTok{my_var <-}\StringTok{ }\NormalTok{x}
|
732
|
+
\KeywordTok{filter}\NormalTok{(df, my_var }\OperatorTok{==}\StringTok{ }\DecValTok{1}\NormalTok{)}
|
733
|
+
\end{Highlighting}
|
734
|
+
\end{Shaded}
|
735
|
+
|
736
|
+
Generates the following error: "object `x' not found.
|
737
|
+
|
738
|
+
However, in Galaaz, arguments are referencially transparent as can be
|
739
|
+
seen by the code bellow. Note initally that `my\_var = :x' will not give
|
740
|
+
the error ``object `x' not found'' since `:x' is treated as an
|
741
|
+
expression and assigned to my\_var. Then when doing (my\_var.eq 1),
|
742
|
+
my\_var is a variable that resolves to `:x' and it becomes equivalent to
|
743
|
+
(:x.eq 1) which is what we want.
|
744
|
+
|
745
|
+
\begin{Shaded}
|
746
|
+
\begin{Highlighting}[]
|
747
|
+
\NormalTok{my_var = }\StringTok{:x}
|
748
|
+
\NormalTok{puts df.filter(my_var.eq }\DecValTok{1}\NormalTok{)}
|
749
|
+
\end{Highlighting}
|
750
|
+
\end{Shaded}
|
751
|
+
|
752
|
+
\begin{verbatim}
|
753
|
+
## x y
|
754
|
+
## 1 1 3
|
755
|
+
\end{verbatim}
|
756
|
+
|
757
|
+
As stated by Hardley
|
758
|
+
|
759
|
+
\begin{quote}
|
760
|
+
dplyr code is ambiguous. Depending on what variables are defined where,
|
761
|
+
filter(df, x == y) could be equivalent to any of:
|
762
|
+
\end{quote}
|
763
|
+
|
764
|
+
\begin{verbatim}
|
765
|
+
df[df$x == df$y, ]
|
766
|
+
df[df$x == y, ]
|
767
|
+
df[x == df$y, ]
|
768
|
+
df[x == y, ]
|
769
|
+
\end{verbatim}
|
770
|
+
|
771
|
+
In galaaz this ambiguity does not exist, filter(df, x.eq y) is not a
|
772
|
+
valid expression as expressions are build with symbols. In doing
|
773
|
+
filter(df, :x.eq y) we are looking for elements of the `x' column that
|
774
|
+
are equal to a previously defined y variable. Finally in filter(df,
|
775
|
+
:x.eq :y) we are looking for elements in which the `x' column value is
|
776
|
+
equal to the `y' column value. This can be seen in the following two
|
777
|
+
chunks of code:
|
778
|
+
|
779
|
+
\begin{Shaded}
|
780
|
+
\begin{Highlighting}[]
|
781
|
+
\NormalTok{y = }\DecValTok{1}
|
782
|
+
\NormalTok{x = }\DecValTok{2}
|
783
|
+
|
784
|
+
\CommentTok{# looking for values where the 'x' column is equal to the 'y' column}
|
785
|
+
\NormalTok{puts df.filter(}\StringTok{:x}\NormalTok{.eq }\StringTok{:y}\NormalTok{)}
|
786
|
+
\end{Highlighting}
|
787
|
+
\end{Shaded}
|
788
|
+
|
789
|
+
\begin{verbatim}
|
790
|
+
## x y
|
791
|
+
## 1 2 2
|
792
|
+
\end{verbatim}
|
793
|
+
|
794
|
+
\begin{Shaded}
|
795
|
+
\begin{Highlighting}[]
|
796
|
+
\CommentTok{# looking for values where the 'x' column is equal to the 'y' variable}
|
797
|
+
\CommentTok{# in this case, the number 1}
|
798
|
+
\NormalTok{puts df.filter(}\StringTok{:x}\NormalTok{.eq y)}
|
799
|
+
\end{Highlighting}
|
800
|
+
\end{Shaded}
|
801
|
+
|
802
|
+
\begin{verbatim}
|
803
|
+
## x y
|
804
|
+
## 1 1 3
|
805
|
+
\end{verbatim}
|
806
|
+
|
807
|
+
\hypertarget{writing-a-function-that-applies-to-different-data-sets}{%
|
808
|
+
\section{Writing a function that applies to different data
|
809
|
+
sets}\label{writing-a-function-that-applies-to-different-data-sets}}
|
810
|
+
|
811
|
+
Let's suppose that we want to write a function that receives as the
|
812
|
+
first argument a data frame and as second argument an expression that
|
813
|
+
adds a column to the data frame that is equal to the sum of elements in
|
814
|
+
column `a' plus `x'.
|
815
|
+
|
816
|
+
Here is the intended behaviour using the `mutate' function of `dplyr':
|
817
|
+
|
818
|
+
\begin{verbatim}
|
819
|
+
mutate(df1, y = a + x)
|
820
|
+
mutate(df2, y = a + x)
|
821
|
+
mutate(df3, y = a + x)
|
822
|
+
mutate(df4, y = a + x)
|
823
|
+
\end{verbatim}
|
824
|
+
|
825
|
+
The naive approach to writing an R function to solve this problem is:
|
826
|
+
|
827
|
+
\begin{verbatim}
|
828
|
+
mutate_y <- function(df) {
|
829
|
+
mutate(df, y = a + x)
|
830
|
+
}
|
831
|
+
\end{verbatim}
|
832
|
+
|
833
|
+
Unfortunately, in R, this function can fail silently if one of the
|
834
|
+
variables isn't present in the data frame, but is present in the global
|
835
|
+
environment. We will not go through here how to solve this problem in R.
|
836
|
+
|
837
|
+
In Galaaz the method mutate\_y bellow will work fine and will never fail
|
838
|
+
silently.
|
839
|
+
|
840
|
+
\begin{Shaded}
|
841
|
+
\begin{Highlighting}[]
|
842
|
+
\KeywordTok{def}\NormalTok{ mutate_y(df)}
|
843
|
+
\NormalTok{ df.mutate(}\StringTok{:y}\NormalTok{.assign }\StringTok{:a}\NormalTok{ + }\StringTok{:x}\NormalTok{)}
|
844
|
+
\KeywordTok{end}
|
845
|
+
\end{Highlighting}
|
846
|
+
\end{Shaded}
|
847
|
+
|
848
|
+
Here we create a data frame that has only one column named `x':
|
849
|
+
|
850
|
+
\begin{Shaded}
|
851
|
+
\begin{Highlighting}[]
|
852
|
+
\NormalTok{df1 = R.data__frame(}\StringTok{x: }\NormalTok{(}\DecValTok{1}\NormalTok{..}\DecValTok{3}\NormalTok{))}
|
853
|
+
\NormalTok{puts df1}
|
854
|
+
\end{Highlighting}
|
855
|
+
\end{Shaded}
|
856
|
+
|
857
|
+
\begin{verbatim}
|
858
|
+
## x
|
859
|
+
## 1 1
|
860
|
+
## 2 2
|
861
|
+
## 3 3
|
862
|
+
\end{verbatim}
|
863
|
+
|
864
|
+
Note that method mutate\_y will fail independetly from the fact that
|
865
|
+
variable `a' is defined and in the scope of the method. Variable `a' has
|
866
|
+
no relationship with the symbol `:a' used in the definition of
|
867
|
+
`mutate\_y' above:
|
868
|
+
|
869
|
+
\begin{Shaded}
|
870
|
+
\begin{Highlighting}[]
|
871
|
+
\NormalTok{a = }\DecValTok{10}
|
872
|
+
\NormalTok{mutate_y(df1)}
|
873
|
+
\end{Highlighting}
|
874
|
+
\end{Shaded}
|
875
|
+
|
876
|
+
\begin{verbatim}
|
877
|
+
## Message:
|
878
|
+
## Error in mutate_impl(.data, dots) :
|
879
|
+
## Evaluation error: object 'a' not found.
|
880
|
+
## In addition: Warning message:
|
881
|
+
## In mutate_impl(.data, dots) :
|
882
|
+
## mismatched protect/unprotect (unprotect with empty protect stack) (RError)
|
883
|
+
## Translated to internal error
|
884
|
+
\end{verbatim}
|
885
|
+
|
886
|
+
\hypertarget{different-expressions}{%
|
887
|
+
\section{Different expressions}\label{different-expressions}}
|
888
|
+
|
889
|
+
Let's move to the next problem as presented by Hardley where trying to
|
890
|
+
write a function in R that will receive two argumens, the first a
|
891
|
+
variable and the second an expression is not trivial. Bellow we create a
|
892
|
+
data frame and we want to write a function that groups data by a
|
893
|
+
variable and summarises it by an expression:
|
894
|
+
|
895
|
+
\begin{Shaded}
|
896
|
+
\begin{Highlighting}[]
|
897
|
+
\KeywordTok{set.seed}\NormalTok{(}\DecValTok{123}\NormalTok{)}
|
898
|
+
|
899
|
+
\NormalTok{df <-}\StringTok{ }\KeywordTok{data.frame}\NormalTok{(}
|
900
|
+
\DataTypeTok{g1 =} \KeywordTok{c}\NormalTok{(}\DecValTok{1}\NormalTok{, }\DecValTok{1}\NormalTok{, }\DecValTok{2}\NormalTok{, }\DecValTok{2}\NormalTok{, }\DecValTok{2}\NormalTok{),}
|
901
|
+
\DataTypeTok{g2 =} \KeywordTok{c}\NormalTok{(}\DecValTok{1}\NormalTok{, }\DecValTok{2}\NormalTok{, }\DecValTok{1}\NormalTok{, }\DecValTok{2}\NormalTok{, }\DecValTok{1}\NormalTok{),}
|
902
|
+
\DataTypeTok{a =} \KeywordTok{sample}\NormalTok{(}\DecValTok{5}\NormalTok{),}
|
903
|
+
\DataTypeTok{b =} \KeywordTok{sample}\NormalTok{(}\DecValTok{5}\NormalTok{)}
|
904
|
+
\NormalTok{)}
|
905
|
+
|
906
|
+
\KeywordTok{as.data.frame}\NormalTok{(df) }
|
907
|
+
\end{Highlighting}
|
908
|
+
\end{Shaded}
|
909
|
+
|
910
|
+
\begin{verbatim}
|
911
|
+
## g1 g2 a b
|
912
|
+
## 1 1 1 2 1
|
913
|
+
## 2 1 2 4 3
|
914
|
+
## 3 2 1 5 4
|
915
|
+
## 4 2 2 3 2
|
916
|
+
## 5 2 1 1 5
|
917
|
+
\end{verbatim}
|
918
|
+
|
919
|
+
\begin{Shaded}
|
920
|
+
\begin{Highlighting}[]
|
921
|
+
\NormalTok{d2 <-}\StringTok{ }\NormalTok{df }\OperatorTok{%>%}
|
922
|
+
\StringTok{ }\KeywordTok{group_by}\NormalTok{(g1) }\OperatorTok{%>%}
|
923
|
+
\StringTok{ }\KeywordTok{summarise}\NormalTok{(}\DataTypeTok{a =} \KeywordTok{mean}\NormalTok{(a))}
|
924
|
+
|
925
|
+
\KeywordTok{as.data.frame}\NormalTok{(d2) }
|
926
|
+
\end{Highlighting}
|
927
|
+
\end{Shaded}
|
928
|
+
|
929
|
+
\begin{verbatim}
|
930
|
+
## g1 a
|
931
|
+
## 1 1 3
|
932
|
+
## 2 2 3
|
933
|
+
\end{verbatim}
|
934
|
+
|
935
|
+
\begin{Shaded}
|
936
|
+
\begin{Highlighting}[]
|
937
|
+
\NormalTok{d2 <-}\StringTok{ }\NormalTok{df }\OperatorTok{%>%}
|
938
|
+
\StringTok{ }\KeywordTok{group_by}\NormalTok{(g2) }\OperatorTok{%>%}
|
939
|
+
\StringTok{ }\KeywordTok{summarise}\NormalTok{(}\DataTypeTok{a =} \KeywordTok{mean}\NormalTok{(a))}
|
940
|
+
|
941
|
+
\KeywordTok{as.data.frame}\NormalTok{(d2) }
|
942
|
+
\end{Highlighting}
|
943
|
+
\end{Shaded}
|
944
|
+
|
945
|
+
\begin{verbatim}
|
946
|
+
## g2 a
|
947
|
+
## 1 1 2.666667
|
948
|
+
## 2 2 3.500000
|
949
|
+
\end{verbatim}
|
950
|
+
|
951
|
+
As shown by Hardley, one might expect this function to do the trick:
|
952
|
+
|
953
|
+
\begin{Shaded}
|
954
|
+
\begin{Highlighting}[]
|
955
|
+
\NormalTok{my_summarise <-}\StringTok{ }\ControlFlowTok{function}\NormalTok{(df, group_var) \{}
|
956
|
+
\NormalTok{ df }\OperatorTok{%>%}
|
957
|
+
\StringTok{ }\KeywordTok{group_by}\NormalTok{(group_var) }\OperatorTok{%>%}
|
958
|
+
\StringTok{ }\KeywordTok{summarise}\NormalTok{(}\DataTypeTok{a =} \KeywordTok{mean}\NormalTok{(a))}
|
959
|
+
\NormalTok{\}}
|
960
|
+
|
961
|
+
\CommentTok{# my_summarise(df, g1)}
|
962
|
+
\CommentTok{#> Error: Column `group_var` is unknown}
|
963
|
+
\end{Highlighting}
|
964
|
+
\end{Shaded}
|
965
|
+
|
966
|
+
In order to solve this problem, coding with dplyr requires the
|
967
|
+
introduction of many new concepts and functions such as `quo', `quos',
|
968
|
+
`enquo', `enquos', `!!' (bang bang), `!!!' (triple bang). Again, we'll
|
969
|
+
leave to Hardley the explanation on how to use all those functions.
|
970
|
+
|
971
|
+
Now, let's try to implement the same function in galaaz. The next code
|
972
|
+
block first prints the `df' data frame define previously in R (to access
|
973
|
+
an R variable from Galaaz, we use the tilda operator `\textasciitilde{}'
|
974
|
+
applied to the R variable name as symbol, i.e., `:df'. We then create
|
975
|
+
the `my\_summarize' method and call it passing the R data frame and the
|
976
|
+
group by variable `:g1':
|
977
|
+
|
978
|
+
\begin{Shaded}
|
979
|
+
\begin{Highlighting}[]
|
980
|
+
\NormalTok{puts ~}\StringTok{:df}
|
981
|
+
\NormalTok{print }\StringTok{"\textbackslash{}n"}
|
982
|
+
|
983
|
+
\KeywordTok{def}\NormalTok{ my_summarize(df, group_var)}
|
984
|
+
\NormalTok{ df.group_by(group_var).}
|
985
|
+
\NormalTok{ summarize(}\StringTok{a: :a}\NormalTok{.mean)}
|
986
|
+
\KeywordTok{end}
|
987
|
+
|
988
|
+
\NormalTok{puts my_summarize(}\StringTok{:df}\NormalTok{, }\StringTok{:g1}\NormalTok{).as__data__frame}
|
989
|
+
\end{Highlighting}
|
990
|
+
\end{Shaded}
|
991
|
+
|
992
|
+
\begin{verbatim}
|
993
|
+
## g1 g2 a b
|
994
|
+
## 1 1 1 2 1
|
995
|
+
## 2 1 2 4 3
|
996
|
+
## 3 2 1 5 4
|
997
|
+
## 4 2 2 3 2
|
998
|
+
## 5 2 1 1 5
|
999
|
+
##
|
1000
|
+
## g1 a
|
1001
|
+
## 1 1 3
|
1002
|
+
## 2 2 3
|
1003
|
+
\end{verbatim}
|
1004
|
+
|
1005
|
+
It works!!! Well, let's make sure this was not just some coincidence
|
1006
|
+
|
1007
|
+
\begin{Shaded}
|
1008
|
+
\begin{Highlighting}[]
|
1009
|
+
\NormalTok{puts my_summarize(}\StringTok{:df}\NormalTok{, }\StringTok{:g2}\NormalTok{).as__data__frame}
|
1010
|
+
\end{Highlighting}
|
1011
|
+
\end{Shaded}
|
1012
|
+
|
1013
|
+
\begin{verbatim}
|
1014
|
+
## g2 a
|
1015
|
+
## 1 1 2.666667
|
1016
|
+
## 2 2 3.500000
|
1017
|
+
\end{verbatim}
|
1018
|
+
|
1019
|
+
Great, everything is fine! No magic, no new functions, no complexities,
|
1020
|
+
just normal, standard Ruby code. If you've ever done NSE in R, this
|
1021
|
+
certainly feels much safer and easy to implement.
|
1022
|
+
|
1023
|
+
\hypertarget{different-input-variables}{%
|
1024
|
+
\section{Different input variables}\label{different-input-variables}}
|
1025
|
+
|
1026
|
+
In the previous section we've managed to get rid of all NSE formulation
|
1027
|
+
for a simple example, but does this remain true for more complex
|
1028
|
+
examples, or will the Galaaz way prove inpractical for more complex
|
1029
|
+
code?
|
1030
|
+
|
1031
|
+
In the next example Hardley proposes us to write a function that given
|
1032
|
+
an expression such as `a' or `a * b', calculates three summaries. What
|
1033
|
+
we want a function that does the same as these R statements:
|
1034
|
+
|
1035
|
+
\begin{verbatim}
|
1036
|
+
summarise(df, mean = mean(a), sum = sum(a), n = n())
|
1037
|
+
#> # A tibble: 1 x 3
|
1038
|
+
#> mean sum n
|
1039
|
+
#> <dbl> <int> <int>
|
1040
|
+
#> 1 3 15 5
|
1041
|
+
|
1042
|
+
summarise(df, mean = mean(a * b), sum = sum(a * b), n = n())
|
1043
|
+
#> # A tibble: 1 x 3
|
1044
|
+
#> mean sum n
|
1045
|
+
#> <dbl> <int> <int>
|
1046
|
+
#> 1 9 45 5
|
1047
|
+
\end{verbatim}
|
1048
|
+
|
1049
|
+
Let's try it in galaaz:
|
1050
|
+
|
1051
|
+
\begin{Shaded}
|
1052
|
+
\begin{Highlighting}[]
|
1053
|
+
\KeywordTok{def}\NormalTok{ my_summarise2(df, expr)}
|
1054
|
+
\NormalTok{ df.summarize(}
|
1055
|
+
\StringTok{mean: }\NormalTok{E.mean(expr),}
|
1056
|
+
\StringTok{sum: }\NormalTok{E.sum(expr),}
|
1057
|
+
\StringTok{n: }\NormalTok{E.n}
|
1058
|
+
\NormalTok{ )}
|
1059
|
+
\KeywordTok{end}
|
1060
|
+
|
1061
|
+
\NormalTok{puts my_summarise2((~}\StringTok{:df}\NormalTok{), }\StringTok{:a}\NormalTok{)}
|
1062
|
+
\NormalTok{puts my_summarise2((~}\StringTok{:df}\NormalTok{), }\StringTok{:a}\NormalTok{ * }\StringTok{:b}\NormalTok{)}
|
1063
|
+
\end{Highlighting}
|
1064
|
+
\end{Shaded}
|
1065
|
+
|
1066
|
+
\begin{verbatim}
|
1067
|
+
## mean sum n
|
1068
|
+
## 1 3 15 5
|
1069
|
+
## mean sum n
|
1070
|
+
## 1 9 45 5
|
1071
|
+
\end{verbatim}
|
1072
|
+
|
1073
|
+
Once again, there is no need to use any special theory or functions. The
|
1074
|
+
only point to be careful about is the use of `E' to build expressions
|
1075
|
+
from functions `mean', `sum' and `n'.
|
1076
|
+
|
1077
|
+
\hypertarget{different-input-and-output-variable}{%
|
1078
|
+
\section{Different input and output
|
1079
|
+
variable}\label{different-input-and-output-variable}}
|
1080
|
+
|
1081
|
+
Now the next challenge presented by Hardley is to vary the name of the
|
1082
|
+
output variables based on the received expression. So, if the input
|
1083
|
+
expression is `a', we want our data frame columns to be named `mean\_a'
|
1084
|
+
and `sum\_a'. Now, if the input expression is `b', columns should be
|
1085
|
+
named `mean\_b' and `sum\_b'.
|
1086
|
+
|
1087
|
+
\begin{verbatim}
|
1088
|
+
mutate(df, mean_a = mean(a), sum_a = sum(a))
|
1089
|
+
#> # A tibble: 5 x 6
|
1090
|
+
#> g1 g2 a b mean_a sum_a
|
1091
|
+
#> <dbl> <dbl> <int> <int> <dbl> <int>
|
1092
|
+
#> 1 1 1 1 3 3 15
|
1093
|
+
#> 2 1 2 4 2 3 15
|
1094
|
+
#> 3 2 1 2 1 3 15
|
1095
|
+
#> 4 2 2 5 4 3 15
|
1096
|
+
#> # … with 1 more row
|
1097
|
+
|
1098
|
+
mutate(df, mean_b = mean(b), sum_b = sum(b))
|
1099
|
+
#> # A tibble: 5 x 6
|
1100
|
+
#> g1 g2 a b mean_b sum_b
|
1101
|
+
#> <dbl> <dbl> <int> <int> <dbl> <int>
|
1102
|
+
#> 1 1 1 1 3 3 15
|
1103
|
+
#> 2 1 2 4 2 3 15
|
1104
|
+
#> 3 2 1 2 1 3 15
|
1105
|
+
#> 4 2 2 5 4 3 15
|
1106
|
+
#> # … with 1 more row
|
1107
|
+
\end{verbatim}
|
1108
|
+
|
1109
|
+
In order to solve this problem in R, Hardley needs to introduce some
|
1110
|
+
more new functions and notations: `quo\_name' and the `:=' operator from
|
1111
|
+
package `rlang'
|
1112
|
+
|
1113
|
+
Here is our Ruby code:
|
1114
|
+
|
1115
|
+
\begin{Shaded}
|
1116
|
+
\begin{Highlighting}[]
|
1117
|
+
\KeywordTok{def}\NormalTok{ my_mutate(df, expr)}
|
1118
|
+
\NormalTok{ mean_name = }\StringTok{"mean_}\OtherTok{#\{}\NormalTok{expr.to_s}\OtherTok{\}}\StringTok{"}
|
1119
|
+
\NormalTok{ sum_name = }\StringTok{"sum_}\OtherTok{#\{}\NormalTok{expr.to_s}\OtherTok{\}}\StringTok{"}
|
1120
|
+
|
1121
|
+
\NormalTok{ df.mutate(mean_name => E.mean(expr),}
|
1122
|
+
\NormalTok{ sum_name => E.sum(expr))}
|
1123
|
+
\KeywordTok{end}
|
1124
|
+
|
1125
|
+
\NormalTok{puts my_mutate((~}\StringTok{:df}\NormalTok{), }\StringTok{:a}\NormalTok{)}
|
1126
|
+
\NormalTok{puts my_mutate((~}\StringTok{:df}\NormalTok{), }\StringTok{:b}\NormalTok{)}
|
1127
|
+
\end{Highlighting}
|
1128
|
+
\end{Shaded}
|
1129
|
+
|
1130
|
+
\begin{verbatim}
|
1131
|
+
## g1 g2 a b mean_a sum_a
|
1132
|
+
## 1 1 1 2 1 3 15
|
1133
|
+
## 2 1 2 4 3 3 15
|
1134
|
+
## 3 2 1 5 4 3 15
|
1135
|
+
## 4 2 2 3 2 3 15
|
1136
|
+
## 5 2 1 1 5 3 15
|
1137
|
+
## g1 g2 a b mean_b sum_b
|
1138
|
+
## 1 1 1 2 1 3 15
|
1139
|
+
## 2 1 2 4 3 3 15
|
1140
|
+
## 3 2 1 5 4 3 15
|
1141
|
+
## 4 2 2 3 2 3 15
|
1142
|
+
## 5 2 1 1 5 3 15
|
1143
|
+
\end{verbatim}
|
1144
|
+
|
1145
|
+
It really seems that ``Non Standard Evaluation'' is actually quite
|
1146
|
+
standard in Galaaz! But, you might have noticed a small change in the
|
1147
|
+
way the arguments to the mutate method were called. In a previous
|
1148
|
+
example we used df.summarise(mean: E.mean(:a), \ldots{}) where the
|
1149
|
+
column name was followed by a `:' colom. In this example, we have
|
1150
|
+
df.mutate(mean\_name =\textgreater{} E.mean(expr), \ldots{}) and
|
1151
|
+
variable mean\_name is not followed by `:' but by `=\textgreater{}'.
|
1152
|
+
This is standard Ruby notation.
|
1153
|
+
|
1154
|
+
{[}explain\ldots{}.{]}
|
1155
|
+
|
1156
|
+
\hypertarget{capturing-multiple-variables}{%
|
1157
|
+
\section{Capturing multiple
|
1158
|
+
variables}\label{capturing-multiple-variables}}
|
1159
|
+
|
1160
|
+
Moving on with new complexities, Hardley proposes us to solve the
|
1161
|
+
problem in which the summarise function will receive any number of
|
1162
|
+
grouping variables.
|
1163
|
+
|
1164
|
+
This again is quite standard Ruby. In order to receive an undefined
|
1165
|
+
number of paramenters the paramenter is preceded by '*':
|
1166
|
+
|
1167
|
+
\begin{Shaded}
|
1168
|
+
\begin{Highlighting}[]
|
1169
|
+
\KeywordTok{def}\NormalTok{ my_summarise3(df, *group_vars)}
|
1170
|
+
\NormalTok{ df.group_by(*group_vars).}
|
1171
|
+
\NormalTok{ summarise(}\StringTok{a: }\NormalTok{E.mean(}\StringTok{:a}\NormalTok{))}
|
1172
|
+
\KeywordTok{end}
|
1173
|
+
|
1174
|
+
\NormalTok{puts my_summarise3((~}\StringTok{:df}\NormalTok{), }\StringTok{:g1}\NormalTok{, }\StringTok{:g2}\NormalTok{).as__data__frame}
|
1175
|
+
\end{Highlighting}
|
1176
|
+
\end{Shaded}
|
1177
|
+
|
1178
|
+
\begin{verbatim}
|
1179
|
+
## g1 g2 a
|
1180
|
+
## 1 1 1 2
|
1181
|
+
## 2 1 2 4
|
1182
|
+
## 3 2 1 3
|
1183
|
+
## 4 2 2 3
|
1184
|
+
\end{verbatim}
|
1185
|
+
|
1186
|
+
\hypertarget{why-does-r-require-nse-and-galaaz-does-not}{%
|
1187
|
+
\section{Why does R require NSE and Galaaz does
|
1188
|
+
not?}\label{why-does-r-require-nse-and-galaaz-does-not}}
|
1189
|
+
|
1190
|
+
NSE introduces a number of new concepts, such as `quoting',
|
1191
|
+
`quasiquotation', `unquoting' and `unquote-splicing', while in Galaaz
|
1192
|
+
none of those concepts are needed. What gives?
|
1193
|
+
|
1194
|
+
R is an extremely flexible language and it has lazy evaluation of
|
1195
|
+
parameters. When in R a function is called as `summarise(df, a = b)',
|
1196
|
+
the summarise function receives the litteral `a = b' parameter and can
|
1197
|
+
work with this as if it were a string. In R, it is not clear what a and
|
1198
|
+
b are, they can be expressions or they can be variables, it is up to the
|
1199
|
+
function to decide what `a = b' means.
|
1200
|
+
|
1201
|
+
In Ruby, there is no lazy evaluation of parameters and `a' is always a
|
1202
|
+
variable and so is `b'. Variables assume their value as soon as they are
|
1203
|
+
used, so `x = a' is immediately evaluate and variable `x' will receive
|
1204
|
+
the value of variable `a' as soon as the Ruby statement is executed.
|
1205
|
+
Ruby also provides the notion of a symbol; `:a' is a symbol and does not
|
1206
|
+
evaluate to anything. Galaaz uses Ruby symbols to build expressions that
|
1207
|
+
are not bound to anything: `:a.eq :b' is clearly an expression and has
|
1208
|
+
no relationship whatsoever with the statment `a = b'. By using symbols,
|
1209
|
+
variables and expressions all the possible ambiguities that are found in
|
1210
|
+
R are eliminated in Galaaz.
|
1211
|
+
|
1212
|
+
The main problem that remains, is that in R, functions are not clearly
|
1213
|
+
documented as what type of input they are expecting, they might be
|
1214
|
+
expecting regular variables or they might be expecting expressions and
|
1215
|
+
the R function will know how to deal with an input of the form `a = b',
|
1216
|
+
now for the Ruby developer it might not be immediately clear if it
|
1217
|
+
should call the function passing the value `true' if variable `a' is
|
1218
|
+
equal to variable `b' or if it should call the function passing the
|
1219
|
+
expression `:a.eq :b'.
|
1220
|
+
|
1221
|
+
\hypertarget{advanced-dplyr-features}{%
|
1222
|
+
\section{Advanced dplyr features}\label{advanced-dplyr-features}}
|
1223
|
+
|
1224
|
+
In the blog: Programming with dplyr by using dplyr
|
1225
|
+
(\url{https://www.r-bloggers.com/programming-with-dplyr-by-using-dplyr/})
|
1226
|
+
Iñaki Úcar shows surprise that some R users are trying to code in dplyr
|
1227
|
+
avoiding the use of NSE. For instance he says:
|
1228
|
+
|
1229
|
+
\begin{quote}
|
1230
|
+
Take the example of seplyr. It stands for standard evaluation dplyr, and
|
1231
|
+
enables us to program over dplyr without having ``to bring in (or study)
|
1232
|
+
any deep-theory or heavy-weight tools such as rlang/tidyeval''.
|
1233
|
+
\end{quote}
|
1234
|
+
|
1235
|
+
For me, there isn't really any surprise that users are trying to avoid
|
1236
|
+
dplyr deep-theory. R users frequently are not programmers and learning
|
1237
|
+
to code is already hard business, on top of that, having to learn how to
|
1238
|
+
`quote' or `enquo' or `quos' or `enquos' is not necessarily a `piece of
|
1239
|
+
cake'. So much so, that `tidyeval' has some more advanced functions that
|
1240
|
+
instead of using quoted expressions, uses strings as arguments.
|
1241
|
+
|
1242
|
+
In the following examples, we show the use of functions `group\_by\_at',
|
1243
|
+
`summarise\_at' and `rename\_at' that receive strings as argument. The
|
1244
|
+
data frame used in `starwars' that describes features of characters in
|
1245
|
+
the Starwars movies:
|
1246
|
+
|
1247
|
+
\begin{Shaded}
|
1248
|
+
\begin{Highlighting}[]
|
1249
|
+
\NormalTok{puts (~}\StringTok{:starwars}\NormalTok{).head.as__data__frame}
|
1250
|
+
\end{Highlighting}
|
1251
|
+
\end{Shaded}
|
1252
|
+
|
1253
|
+
\begin{verbatim}
|
1254
|
+
## name height mass hair_color skin_color eye_color birth_year
|
1255
|
+
## 1 Luke Skywalker 172 77 blond fair blue 19.0
|
1256
|
+
## 2 C-3PO 167 75 <NA> gold yellow 112.0
|
1257
|
+
## 3 R2-D2 96 32 <NA> white, blue red 33.0
|
1258
|
+
## 4 Darth Vader 202 136 none white yellow 41.9
|
1259
|
+
## 5 Leia Organa 150 49 brown light brown 19.0
|
1260
|
+
## 6 Owen Lars 178 120 brown, grey light blue 52.0
|
1261
|
+
## gender homeworld species
|
1262
|
+
## 1 male Tatooine Human
|
1263
|
+
## 2 <NA> Tatooine Droid
|
1264
|
+
## 3 <NA> Naboo Droid
|
1265
|
+
## 4 male Tatooine Human
|
1266
|
+
## 5 female Alderaan Human
|
1267
|
+
## 6 male Tatooine Human
|
1268
|
+
## films
|
1269
|
+
## 1 Revenge of the Sith, Return of the Jedi, The Empire Strikes Back, A New Hope, The Force Awakens
|
1270
|
+
## 2 Attack of the Clones, The Phantom Menace, Revenge of the Sith, Return of the Jedi, The Empire Strikes Back, A New Hope
|
1271
|
+
## 3 Attack of the Clones, The Phantom Menace, Revenge of the Sith, Return of the Jedi, The Empire Strikes Back, A New Hope, The Force Awakens
|
1272
|
+
## 4 Revenge of the Sith, Return of the Jedi, The Empire Strikes Back, A New Hope
|
1273
|
+
## 5 Revenge of the Sith, Return of the Jedi, The Empire Strikes Back, A New Hope, The Force Awakens
|
1274
|
+
## 6 Attack of the Clones, Revenge of the Sith, A New Hope
|
1275
|
+
## vehicles starships
|
1276
|
+
## 1 Snowspeeder, Imperial Speeder Bike X-wing, Imperial shuttle
|
1277
|
+
## 2
|
1278
|
+
## 3
|
1279
|
+
## 4 TIE Advanced x1
|
1280
|
+
## 5 Imperial Speeder Bike
|
1281
|
+
## 6
|
1282
|
+
\end{verbatim}
|
1283
|
+
|
1284
|
+
The grouped\_mean function bellow will receive a grouping variable and
|
1285
|
+
calculate summaries for the value\_variables given:
|
1286
|
+
|
1287
|
+
\begin{Shaded}
|
1288
|
+
\begin{Highlighting}[]
|
1289
|
+
\NormalTok{grouped_mean <-}\StringTok{ }\ControlFlowTok{function}\NormalTok{(data, grouping_variables, value_variables) \{}
|
1290
|
+
\NormalTok{ data }\OperatorTok{%>%}
|
1291
|
+
\StringTok{ }\KeywordTok{group_by_at}\NormalTok{(grouping_variables) }\OperatorTok{%>%}
|
1292
|
+
\StringTok{ }\KeywordTok{mutate}\NormalTok{(}\DataTypeTok{count =} \KeywordTok{n}\NormalTok{()) }\OperatorTok{%>%}
|
1293
|
+
\StringTok{ }\KeywordTok{summarise_at}\NormalTok{(}\KeywordTok{c}\NormalTok{(value_variables, }\StringTok{"count"}\NormalTok{), mean, }\DataTypeTok{na.rm =} \OtherTok{TRUE}\NormalTok{) }\OperatorTok{%>%}
|
1294
|
+
\StringTok{ }\KeywordTok{rename_at}\NormalTok{(value_variables, }\KeywordTok{funs}\NormalTok{(}\KeywordTok{paste0}\NormalTok{(}\StringTok{"mean_"}\NormalTok{, .)))}
|
1295
|
+
\NormalTok{ \}}
|
1296
|
+
|
1297
|
+
\NormalTok{gm =}\StringTok{ }\NormalTok{starwars }\OperatorTok{%>%}\StringTok{ }
|
1298
|
+
\StringTok{ }\KeywordTok{grouped_mean}\NormalTok{(}\StringTok{"eye_color"}\NormalTok{, }\KeywordTok{c}\NormalTok{(}\StringTok{"mass"}\NormalTok{, }\StringTok{"birth_year"}\NormalTok{))}
|
1299
|
+
|
1300
|
+
\KeywordTok{as.data.frame}\NormalTok{(gm) }
|
1301
|
+
\end{Highlighting}
|
1302
|
+
\end{Shaded}
|
1303
|
+
|
1304
|
+
\begin{verbatim}
|
1305
|
+
## eye_color mean_mass mean_birth_year count
|
1306
|
+
## 1 black 76.28571 33.00000 10
|
1307
|
+
## 2 blue 86.51667 67.06923 19
|
1308
|
+
## 3 blue-gray 77.00000 57.00000 1
|
1309
|
+
## 4 brown 66.09231 108.96429 21
|
1310
|
+
## 5 dark NaN NaN 1
|
1311
|
+
## 6 gold NaN NaN 1
|
1312
|
+
## 7 green, yellow 159.00000 NaN 1
|
1313
|
+
## 8 hazel 66.00000 34.50000 3
|
1314
|
+
## 9 orange 282.33333 231.00000 8
|
1315
|
+
## 10 pink NaN NaN 1
|
1316
|
+
## 11 red 81.40000 33.66667 5
|
1317
|
+
## 12 red, blue NaN NaN 1
|
1318
|
+
## 13 unknown 31.50000 NaN 3
|
1319
|
+
## 14 white 48.00000 NaN 1
|
1320
|
+
## 15 yellow 81.11111 76.38000 11
|
1321
|
+
\end{verbatim}
|
1322
|
+
|
1323
|
+
The same code with Galaaz, becomes:
|
1324
|
+
|
1325
|
+
\begin{Shaded}
|
1326
|
+
\begin{Highlighting}[]
|
1327
|
+
\KeywordTok{def}\NormalTok{ grouped_mean(data, grouping_variables, value_variables)}
|
1328
|
+
\NormalTok{ data.}
|
1329
|
+
\NormalTok{ group_by_at(grouping_variables).}
|
1330
|
+
\NormalTok{ mutate(}\StringTok{count: }\NormalTok{E.n).}
|
1331
|
+
\NormalTok{ summarise_at(E.c(value_variables, }\StringTok{"count"}\NormalTok{), ~}\StringTok{:mean}\NormalTok{, }\StringTok{na__rm: }\DecValTok{true}\NormalTok{).}
|
1332
|
+
\NormalTok{ rename_at(value_variables, E.funs(E.paste0(}\StringTok{"mean_"}\NormalTok{, value_variables)))}
|
1333
|
+
\KeywordTok{end}
|
1334
|
+
|
1335
|
+
\NormalTok{puts grouped_mean((~}\StringTok{:starwars}\NormalTok{), }\StringTok{"eye_color"}\NormalTok{, E.c(}\StringTok{"mass"}\NormalTok{, }\StringTok{"birth_year"}\NormalTok{)).as__data__frame}
|
1336
|
+
\end{Highlighting}
|
1337
|
+
\end{Shaded}
|
1338
|
+
|
1339
|
+
\begin{verbatim}
|
1340
|
+
## eye_color mean_mass mean_birth_year count
|
1341
|
+
## 1 black 76.28571 33.00000 10
|
1342
|
+
## 2 blue 86.51667 67.06923 19
|
1343
|
+
## 3 blue-gray 77.00000 57.00000 1
|
1344
|
+
## 4 brown 66.09231 108.96429 21
|
1345
|
+
## 5 dark NaN NaN 1
|
1346
|
+
## 6 gold NaN NaN 1
|
1347
|
+
## 7 green, yellow 159.00000 NaN 1
|
1348
|
+
## 8 hazel 66.00000 34.50000 3
|
1349
|
+
## 9 orange 282.33333 231.00000 8
|
1350
|
+
## 10 pink NaN NaN 1
|
1351
|
+
## 11 red 81.40000 33.66667 5
|
1352
|
+
## 12 red, blue NaN NaN 1
|
1353
|
+
## 13 unknown 31.50000 NaN 3
|
1354
|
+
## 14 white 48.00000 NaN 1
|
1355
|
+
## 15 yellow 81.11111 76.38000 11
|
1356
|
+
\end{verbatim}
|
1357
|
+
|
1358
|
+
\hypertarget{conclusion}{%
|
1359
|
+
\section{Conclusion}\label{conclusion}}
|
1360
|
+
|
1361
|
+
Ruby and Galaaz provide a nice framework for developing code that uses R
|
1362
|
+
functions. Although R is a very powerful and flexible language,
|
1363
|
+
sometimes, too much flexibility makes life harder for the casual user.
|
1364
|
+
We believe however, that even for the advanced user, Ruby integrated
|
1365
|
+
with R throught Galaaz, makes a powerful environment for data analysis.
|
1366
|
+
In this blog post we showed how Galaaz consistent syntax eliminates the
|
1367
|
+
need for complex constructs such as quoting, enquoting, quasiquotation,
|
1368
|
+
etc. This simplification comes from the fact that expressions and
|
1369
|
+
variables are clearly separated objects, which is not the case in the R
|
1370
|
+
language.
|
1371
|
+
|
1372
|
+
|
1373
|
+
\end{document}
|