Closed tohuynh closed 4 years ago
Agree on all fronts.
Similar to the backend PR linked, I would make it a try catch. Or if you wanted to be really fancy, have a transcript renderer map. Where given the transcript format string it returns the content ready to display.
Similar to the backend PR linked, I would make it a try catch. Or if you wanted to be really fancy, have a transcript renderer map. Where given the transcript format string it returns the content ready to display.
Thanks for this recommendation. I'm using the transcript renderer map. Here's a preview of what speaker-turn blocks looks like.
Any thoughts?
This already looks awesome. I only have nitpicky things.
Nitpicky things:
Mind if I send this for feedback by UX peeps too?
No, not at all.
This one is for timestamped-sentences
. I could use any input on that too.
I think for timestamped-sentences
can we keep the current render? Having a card per sentence makes it feel like a lot is going on.
Ok, will switch back.
Resolved in #57
In light of #101, the front-end of transcripts should be updated to show new
timestamped-speaker-turns
format.I'm thinking maybe group the sentences together for a speaker turn block. That is, just add a horizontal line between each speaker turn block. If there's a speaker, show that above each speaker turn block.
I think the transcript under Search Transcript can remain the same, regardless of formats.