Reader returning duplicated content/content not set to visible

Using the reader endpoint results in duplicate information as it is returning text from elements set not to be visible.

```
curl "https://r.jina.ai/https://github.com/trust-center/privacy" \
-H "Authorization: Bearer MYTOKEN"
```

The html for the duplicate content has `Display: none`.

Is there a way for us to only return content that would be visible on the page?

Start by reproducing the reader endpoint request for the linked GitHub URL and inspect how its HTML is converted to reader content. Check how elements with `display: none` are handled; done means the response no longer includes hidden content or duplicates the visible text. No source file or test is identified in the issue.

Written by the indexing model from the issue text.