My guess is that to create a nice user experience you have to decide whether to only support one or both.
The text_links have the limitations described in the readme. The user should be warned IMO when
1) no annotation links were found but text links were found
2) annotation links were found but the number of links different from the text links
3) no links were found
The annotation_links seem to be extracted without issues when present. I did not extensively test this, but I have a few test PDFs used in the tests that have different characteristics.
mojomonger
commented
1 year ago
My guess is that to create a nice user experience you have to decide whether to only support one or both. The text_links have the limitations described in the readme. The user should be warned IMO when 1) no annotation links were found but text links were found 2) annotation links were found but the number of links different from the text links 3) no links were found The annotation_links seem to be extracted without issues when present. I did not extensively test this, but I have a few test PDFs used in the tests that have different characteristics.