Preserve document encoding in TranslateDocument

gems/aws-sdk-translate/CHANGELOG.md

@@ -1,6 +1,8 @@
 Unreleased Changes
 ------------------
 
+* Issue - Preserve Document encoding in `TranslateDocument` for text documents (#2897).
+
 1.57.0 (2023-07-18)
 ------------------
 

gems/aws-sdk-translate/lib/aws-sdk-translate/plugins/translate_document_encoding.rb

@@ -0,0 +1,31 @@
+# frozen_string_literal: true
+
+module Aws::Translate
+  module Plugins
+    # @api private
+    # The translated_documented returned by the translate_document api is modeled
+    # as a blob, since it may contain binary data (eg: word doc).  However,
+    # when the input document is text (eg: text/plain or text/html) the encoding
+    # should be preserved.
+    class TranslateDocumentEncoding < Seahorse::Client::Plugin
+      class Handler < Seahorse::Client::Handler
+        def call(context)
+          # detect encoding
+          document = context.params[:document]
+          encoding =
+            if document[:content_type].start_with?('text/') && document[:content].is_a?(String)
+              document[:content].encoding
+            end
+          resp = @handler.call(context)
+          if encoding
+            resp.translated_document.content = resp.translated_document.content.force_encoding(encoding)
+          end
+          resp
+        end
+      end
+      def add_handlers(handlers, _config)
+        handlers.add(Handler, step: :initialize, operations: [:translate_document])
+      end
+    end
+  end
+end

gems/aws-sdk-translate/spec/translate_document_encoding_spec.rb

@@ -0,0 +1,60 @@
+# frozen_string_literal: true
+
+require_relative 'spec_helper'
+
+module Aws::Translate
+  module Plugins
+    describe TranslateDocumentEncoding do
+      let(:client) { Aws::Translate::Client.new(stub_responses: true) }
+
+      let(:text_document) do
+        { content_type: 'text/plain', content: 'plain text document' }
+      end
+
+      let(:translated_text_document) do
+        Base64.encode64('translated text document')
+      end
+
+      let(:word_document) do
+        {
+          content_type: 'application/vnd.openxmlformats-officedocument.wordprocessingml.document',
+          content: "ZmFrZSB3b3JkIGRvYw==\n"
+        }
+      end
+
+      let(:translated_word_document) do
+        "dHJhbnNsYXRlZCBmYWtlIHdvcmQgZG9j\n"
+      end
+
+      describe '#translate_document' do
+        it 'preserves the encoding on text documents' do
+          client.stub_responses(:translate_document, {
+            translated_document: { content: translated_text_document },
+            source_language_code: 'en',
+            target_language_code: 'es'
+          })
+          resp = client.translate_document(
+            document: text_document,
+            source_language_code: 'en',
+            target_language_code: 'es'
+          )
+          expect(resp.translated_document.content.encoding).to eq(text_document[:content].encoding)
+        end
+
+        it 'does not change the encoding on non-text documents' do
+          client.stub_responses(:translate_document, {
+            translated_document: { content: translated_word_document },
+            source_language_code: 'en',
+            target_language_code: 'es'
+          })
+          resp = client.translate_document(
+            document: word_document,
+            source_language_code: 'en',
+            target_language_code: 'es'
+          )
+          expect(resp.translated_document.content.encoding).to eq(Encoding.find('binary'))
+        end
+      end
+    end
+  end
+end

services.json

@@ -1126,7 +1126,10 @@
     "models": "transfer/2018-11-05"
   },
   "Translate": {
-    "models": "translate/2017-07-01"
+    "models": "translate/2017-07-01",
+    "addPlugins": [
+      "Aws::Translate::Plugins::TranslateDocumentEncoding"
+    ]
   },
   "VPCLattice": {
     "models": "vpc-lattice/2022-11-30"