If @seconddayout is like me, the issue is with the YouTube video transcripts. Although it looks like one can select across <div>s...
...when Hypothesis grabs the text, it only gets the part in the first <div>:
I ran into this same sort of problem when I was creating an equivalent to droppdf for podcasts (see unchecked-transcript). To solve the problem, I put the timestamp into a CSS :before declaration. See this example for what it looks like in practice—notably the first Hypothesis annotation that crosses three <div>s.
This pull request is my best guess at how this solution would work with droppdf. Unfortunately, I couldn't get all of the droppdf prerequisites to install on my machine, so I wasn't able to do an end-to-end test of the code changes.
Addresses https://github.com/dwhly-proj/droppdf/issues/129
If @seconddayout is like me, the issue is with the YouTube video transcripts. Although it looks like one can select across <div>s...
...when Hypothesis grabs the text, it only gets the part in the first <div>:
I ran into this same sort of problem when I was creating an equivalent to droppdf for podcasts (see unchecked-transcript). To solve the problem, I put the timestamp into a CSS
:beforedeclaration. See this example for what it looks like in practice—notably the first Hypothesis annotation that crosses three <div>s.This pull request is my best guess at how this solution would work with droppdf. Unfortunately, I couldn't get all of the droppdf prerequisites to install on my machine, so I wasn't able to do an end-to-end test of the code changes.