Over at TEDx in Cape Town we’re switching to using Algolia for search on our talks page. We’re busy collecting as much meta data about our talks as possible to make the search experience as rich as possible.
Nice! Be sure to let us know when the search is live and what metadata you found to be the most useful.
Super interesting @craig.wattrus ! We’ve hosted a community event in our office and someone gave a talk about how he made search inside videos - meaning you could search for a word and it would open up the video at the time in which the sentence in which that word was spoken.
It was mind blowing. That person indexed the video transcriptions and then once a user selected the word inside the video it would append the timestamp query parameter to the video of when the sentence started and load it.
Maybe it would be an interesting enhancement as it’s definitely something we haven’t seen before. Let me know if you are interested and I’ll find that person’s contact details and maybe you two can talk and exchange ideas
Hey @craig.wattrus, that is a great project A few years back, we built a demo using Ted as an example on our own (this is not an official TED project). We refreshened it recently and it is available here. There are a few key elements that are outstanding in this demo:
- it is using the search for facet values feature. This is what makes the filter menus searchable.
- it makes searches in multiple indices at the same time.
The search for facet values is already available in the engine, the JS Helper and instantsearch.js but it is not officially out yet. Feel free to ask us anything, we’d be happy to help
Hey Craig - Happy New Year! So great to hear from you and thanks for coming on our Discourse page and sharing this update.
If I’m not mistaken that was @nico_lrx who indexed the video transcriptions Am I right Nicolas?
Hey @dzello, indeed!
@craig.wattrus, I was working on a project indexing the Ted Talks transcripts. I just scraped the transcripts and you can retrieve the url with a get variable to play the video at the exact time for each piece of transcript.
Do not hesitate to contact me if you want more info
I’d be happy to share tips & best practices once you have you data pushed to an index.
@alex : Really cool, I love the UI. I will ping you when I find the time to build the search engine for Ted Talks
Thanks for the input
I’m excited to share this once we’ve got it live!