Merge pull request #17 from cipherstash/add-rake-tasks-for-generating-tests

Add rake tasks for generating orderable strings test cases

Author: CDThomas

Rakefile

@@ -28,6 +28,18 @@ task :release do
   sh "git release"
 end
 
+require "./lib/cipherstash/client/ordered_string_test_generator"
+
+desc "Generate test cases for orderise_string to be used by other CipherStash clients"
+task :generate_orderise_string_test_cases do
+  CipherStash::Client::OrderedStringTestGenerator.new.generate_orderise_string_test_cases
+end
+
+desc "Generate test cases for string comparison to be used by other CipherStash clients"
+task :generate_string_comparison_test_cases do
+  CipherStash::Client::OrderedStringTestGenerator.new.generate_string_comparison_test_cases
+end
+
 require 'yard'
 
 YARD::Rake::YardocTask.new :doc do |yardoc|

lib/cipherstash/client/ordered_string_test_generator.rb

@@ -0,0 +1,90 @@
+require "securerandom"
+require "json"
+
+require_relative "../index"
+
+module CipherStash
+  class Client
+    # Generates test cases for orderable strings that can be used in automated tests in other
+    # clients (e.g. StashRS). Test cases are written to JSON files.
+    #
+    # Example usage in StashRS:
+    # https://github.com/cipherstash/cipherstash-rs/blob/00ff66d712e6f36a89acfa731b680a69789647cf/packages/cipherstash-client/src/indexer/mapping_indexer.rs#L537-L570
+    #
+    # @private
+    class OrderedStringTestGenerator
+      # The number of test cases to generate per test file. This number is arbitrary. 100 seems good
+      # because GitHub is willing to display the output files in diffs in PRs, but will require
+      # pulling branches down for larger numbers of test cases.
+      NUM_TEST_CASES = 100
+
+      # Max character code for ASCII characters in randomly generated strings in test data.
+      ASCII_CHAR_CODE_MAX = 127
+
+      # The max length of randomly generated strings in test data. This number is also somewhat
+      # arbitrary. The first 80 ASCII characters are considered for ordering, so we want some
+      # strings with a length of at least 80. 200 seems good because this will also test lengths
+      # beyond what actually gets indexed for ordering but doesn't bloat the test files too much.
+      MAX_STRING_LENGTH = 200
+
+      def generate_orderise_string_test_cases
+        orderise_string_cases = (0..(NUM_TEST_CASES - 1)).map do
+          str = random_ascii_string
+          output = index.__send__ :orderise_string, str
+          {input: str, output: output}
+        end
+
+        puts JSON.pretty_generate(orderise_string_cases)
+      end
+
+      def generate_string_comparison_test_cases
+        string_comparison_cases = (0..(NUM_TEST_CASES - 1)).map do
+          str_a = random_ascii_string
+          terms_a = index.__send__ :orderise_string, str_a
+
+          str_b = random_ascii_string
+          terms_b = index.__send__ :orderise_string, str_b
+
+          output = case terms_a <=> terms_b
+            when -1
+              "<"
+            when 0
+              "=="
+            when 1
+              ">"
+            end
+
+          {input: [str_a, str_b], output: output}
+        end
+
+        puts JSON.pretty_generate(string_comparison_cases)
+      end
+
+      private
+
+      def random_ascii_string
+        (0..rand(MAX_STRING_LENGTH - 1)).map { rand(ASCII_CHAR_CODE_MAX + 1).chr }.join
+      end
+
+      def index
+        id = SecureRandom.uuid
+        settings = {
+          "meta" => {
+            "$indexId" => id,
+            "$indexName" => "titleSort",
+            "$prfKey" => SecureRandom.hex(16),
+            "$prpKey" => SecureRandom.hex(16),
+          },
+          "mapping" => {
+            "kind" => "range",
+            "field" => "title",
+            "fieldType"=>"string",
+          }
+        }
+        schema_versions = {:first=>0, :last=>0, :searchable=>true}
+
+        CipherStash::Index.generate(id, settings, schema_versions)
+      end
+    end
+  end
+end