Merge modsulator gem into the rails app

This reduces the overhead of having to maintain two projects
independently.  Since this app is the only consumer of the modsulator
gem, this approach makes sense.

.gitignore

@@ -1,8 +1,6 @@
 *.rbc
 capybara-*.html
 .rspec
-/log
-/tmp
 /db/*.sqlite3
 /db/*.sqlite3-journal
 /public/system
@@ -15,6 +13,13 @@ pickle-email-*.html
 # TODO Comment out these rules if you are OK with secrets being uploaded to the repo
 config/initializers/secret_token.rb
 
+# Ignore all logfiles and tempfiles.
+/log/*
+/tmp/*
+.tmp
+!/log/.keep
+!/tmp/.keep
+
 # dotenv
 # TODO Comment out this rule if environment variables can be committed
 .env

Gemfile

@@ -5,16 +5,18 @@ git_source(:github) do |repo_name|
   "https://github.com/#{repo_name}.git"
 end
 
-gem 'modsulator', '~> 1.2'
-gem 'stanford-mods-normalizer'
-gem 'honeybadger'
-
 # Bundle edge Rails instead: gem 'rails', github: 'rails/rails'
 gem 'rails', '~> 5.0.2'
 
 # Use Puma as the app server
 gem 'puma', '~> 3.0'
 
+gem 'stanford-mods-normalizer'
+gem 'roo', '>= 2.7.1'
+gem 'honeybadger'
+gem 'deprecation'
+
+
 group :development, :test do
   # Call 'byebug' anywhere in the code to stop execution and get a debugger console
   gem 'byebug', platform: :mri

Gemfile.lock

@@ -73,7 +73,7 @@ GEM
       thor (~> 0.19.4)
       tins (~> 1.6)
     crass (1.0.4)
-    deprecation (0.99.0)
+    deprecation (1.0.0)
       activesupport
     diff-lcs (1.3)
     dlss-capistrano (3.4.1)
@@ -104,13 +104,6 @@ GEM
     mini_mime (1.0.1)
     mini_portile2 (2.3.0)
     minitest (5.11.3)
-    modsulator (1.3.0)
-      activesupport
-      deprecation (~> 0)
-      equivalent-xml (>= 0.6.0)
-      nokogiri
-      roo (>= 2.7.1)
-      stanford-mods-normalizer (~> 0.1)
     net-scp (1.2.1)
       net-ssh (>= 2.6.5)
     net-ssh (5.0.2)
@@ -210,13 +203,14 @@ DEPENDENCIES
   capistrano-passenger
   capistrano-rails
   coveralls
+  deprecation
   dlss-capistrano
   equivalent-xml (>= 0.6.0)
   honeybadger
   listen (~> 3.0.5)
-  modsulator (~> 1.2)
   puma (~> 3.0)
   rails (~> 5.0.2)
+  roo (>= 2.7.1)
   rspec-rails (~> 3.5)
   spring
   spring-watcher-listen (~> 2.0.0)

Rakefile

@@ -4,3 +4,11 @@
 require_relative 'config/application'
 
 Rails.application.load_tasks
+
+Rake::Task['spec'].clear
+RSpec::Core::RakeTask.new(:spec) do |t|
+  t.pattern = 'spec/**/*_spec.rb'
+
+  # The modsulator integration_tests are very slow
+  t.exclude_pattern = 'spec/integration_tests/*_spec.rb'
+end

app/controllers/modsulator_controller.rb

@@ -5,6 +5,6 @@ def create
   end
 
   def version
-    render plain: Gem.loaded_specs['modsulator'].version.version
+    render plain: '2.0.0'
   end
 end

app/models/modsulator_sheet.rb

@@ -0,0 +1,61 @@
+# File "modsulator_sheet.rb" - a class to load and validate metadata spreadsheets (.xlsx or .csv) for input
+# to Modsulator.
+
+require 'json'
+require 'roo'
+
+# This class provides methods to parse Stanford's MODS spreadsheets into either an array of hashes, or a JSON string.
+class ModsulatorSheet
+  attr_reader :file, :filename
+
+  # Creates a new ModsulatorSheet. When called with temporary files, the filename must be specified separately, hence the
+  # second argument.
+  # @param [File]    file        The input spreadsheet
+  # @param [String]  filename    The filename of the input spreadsheet.
+  def initialize(file, filename)
+    @file = file
+    @filename = filename
+  end
+
+
+  # Loads the input spreadsheet into an array of hashes. This spreadsheet should conform to the Stanford MODS template format,
+  # which has three header rows. The first row is a kind of "super header", the second row is an intermediate header and the
+  # third row is the header row that names the fields. The data rows are in the fourth row onwards.
+  #
+  # @return [Array<Hash>]      An array with one entry per data row in the spreadsheet. Each entry is a hash, indexed by
+  #                            the spreadsheet headers.
+  def rows
+    # Parse the spreadsheet, automatically finding the header row by looking for "druid" and "sourceId" and leave the
+    # header row itself out of the resulting array. Everything preceding the header row is discarded.
+    @rows ||= spreadsheet.parse(header_search: ['druid', 'sourceId'], clean: true)
+  end
+
+
+  # Opens a spreadsheet based on its filename extension.
+  #
+  # @return [Roo::CSV, Roo::Excel, Roo::Excelx]   A Roo object, whose type depends on the extension of the given filename.
+  def spreadsheet
+    @spreadsheet ||= case File.extname(@filename)
+                     when '.csv' then Roo::Spreadsheet.open(@file, extension: :csv)
+                     when '.xls' then Roo::Spreadsheet.open(@file, extension: :xls)
+                     when '.xlsx' then Roo::Spreadsheet.open(@file, extension: :xlsx)
+                     else fail "Unknown file type: #{@filename}"
+    end
+  end
+
+
+  # Get the headers used in the spreadsheet
+  def headers
+    rows.first.keys
+  end
+
+
+  # Convert the loaded spreadsheet to a JSON string.
+  # @return [String]  A JSON string.
+  def to_json
+    json_hash = {}
+    json_hash['filename'] = File.basename(filename)
+    json_hash['rows'] = rows
+    json_hash.to_json
+  end
+end