[B] Replace invalid chars in gdoc ingestion

[Fixes #409]
ManifoldScholar · Jun 28, 2017 · 60c08b5 · 60c08b5
1 parent 9f84e05
commit 60c08b5
Show file tree

Hide file tree

Showing 2 changed files with 5 additions and 4 deletions.
diff --git a/api/app/services/ingestor/strategy/google_doc/strategy.rb b/api/app/services/ingestor/strategy/google_doc/strategy.rb
@@ -49,7 +49,8 @@ def self.inspector(ingestion, pointer)
 
         def fetch(pointer)
           contents = self.class.session.drive.export_file(pointer.id, "text/html")
-          @ingestion.write("index.html", contents)
+          encoded_contents = contents.encode("utf-8", invalid: :replace, undef: :replace)
+          @ingestion.write("index.html", encoded_contents)
           self
         end
 

diff --git a/api/spec/services/ingestor/strategy/google_doc/strategy_spec.rb b/api/spec/services/ingestor/strategy/google_doc/strategy_spec.rb
@@ -37,8 +37,8 @@
         expect(@text.page_list).to eq []
       end
 
-      it "has no stylesheets" do
-        expect(@text.stylesheets.length).to eq 0
+      it "has one stylesheet" do
+        expect(@text.stylesheets.length).to eq 1
       end
 
       it "has one entry in the spine" do
@@ -49,7 +49,7 @@
         expect(@text.text_sections.length).to eq 1
       end
 
-      it "has ond ingestion sources" do
+      it "has one ingestion sources" do
         expect(@text.ingestion_sources.length).to eq 1
       end