Is Dify's local deployment not equipped with a built-in document parsing engine by default?

I have already installed and run unstructured-api locally, but why is the document parser component still unable to recognize the content in scanned PDF files?
The .env file configuration is as follows:

The workflow in Dify is as follows: