Ruby Beats Go

Why use Ruby for the glue code and not Go, which has become the de facto choice in this Programming Snapshot column, you ask? Well, a scripting language like Ruby simply can't be beaten for plain vanilla text manipulation and simple, concise, easy-to-extend code that almost looks like a config file. On top of that, Ruby's template system ERB is great for dumping text or picture file names into predefined snippets with a mess of TeX code. There is a good reason why configuration management tools like Puppet and Chef use Ruby as a reference language for dynamically adapting variegated configuration files catering to individual environments.

Listing 2 shows you the Ruby glue code to turn the book description scripts into TeX. It defines a class that user scripts need to include to generate coffee table books, just like Listing 1 does in line 2. Scripts call the built-in Ruby function require_relative bm to find the library from Listing 2 in the same directory and can then immediately get started by calling Bm.new. The returned bm instance then lets you create sophisticated photo books with just a few commands.

01 #!/usr/bin/ruby
02 require 'erb'
03
04 class Bm
05   attr_accessor :name
06
07   def initialize
08     @tmpldir = "tmpl"
09     @photodir = "photos"
10     @tex = ""
11   end
12
13   def add(tmpl, text: "", photos: [])
14     tpath = "#{@tmpldir}/#{tmpl}.tex"
15
16     photos = photos.map {
17      |path| "#{@photodir}/#{path}"}
18
19     t = ERB.new(File.read(tpath))
20     @tex += t.result_with_hash(
21       text: text, photos: photos)
22   end
23
24   def print
25     texf = "#{@name}.tex"
26     File.open(texf, "w") {
27       |f| f.write(@tex) }
28     system("xelatex", texf)
29   end
30 end

Short Paths

To keep things tidy, the predefined TeX templates are all located in a subdirectory named tmpl/. To help the user call them by simple names, such as cover, Listing 2 defines an @tmpldir instance variable that points to the subdirectory containing the templates. When it's time to load a template, line 14 prepends the tmpl path to the provided file name and adds a .tex suffix thereafter. This means that the generated path points to a real file, which the ERB template processor can then read from disk in line 19.

The template processor uses ERB's result_with_hash() method in line 20 to insert the text supplied for the page caption, and the paths to the photo files, into the respective template. The placeholders in the templates are <%= text %> for the page text and <%= photos[0] %> for the first element of the photos array passed in, containing a list of JPEG paths to the photos to be included. For the image paths, there is no need to laboriously type in the entire path each time. As long as the @photodir instance variable is pointing to the directory where all the images are stored (photos in our example), you only need to specify the file name without a path in each case in Listing 1.

Defining the Name

To define the name of the book file to be created, the calling script sets the name attribute of a Bm object as shown in line 5 of Listing 1. To make this work, Listing 2 uses attr_accessor in line 5 to define :name and with it a read-write attribute name for all objects of the class.

This means that an instance of the Bm class in bm can simply call bm.name = "mybook", like it does in line 5 of Listing 1, to set the book name by assignment. Line 25 in Listing 2 uses @name to obtain the name from the instance variable and interpolates it within a string with "#{@name}". This gives you the name of the TeX file with all the piled up TeX code and will send the file to the XeLaTeX LaTeX compiler in the next step.

This means that line 26 only needs to open the .tex file for writing and drop the content of the @tex instance variable there, after having assembled all the necessary TeX instructions in it. The system() call in line 28 then runs XeLaTeX against this, which produces a PDF document of the same name with the finished photo book, if all goes well.