Live data from Hacker News

Show HN: YouTube Full Text Search – Search all of a channel from the commandline

github.com

121–130 of 172 posts

Re: Show HN: YouTube Full Text Search – Search all of a channel from the commandline

#121
post #3

I love that a third party is stepping up here, but it's incomprehensible to me that Google doesn't do this themselves. They're a search company, and they own YouTube. The YouTube data — including the subtitle files — is already sitting there on their servers; they don't have to scrape it, they just have to index it. What are they even doing? Fun thing to try: do a Google search with "site:youtube.com" in it. You get…

Its Google, the obvious seems to elude them even when its sitting in front of them.

Re: Show HN: YouTube Full Text Search – Search all of a channel from the commandline

#122

Earlier quoted context omitted.

Actually, letting people search video text would enable less watching and that is probably the reason they aren't interested in it.

Exactly! Google is in the attention business, the more you stay the more they make

> Google is in the attention business

This is probably the best description of their business model.

Re: Show HN: YouTube Full Text Search – Search all of a channel from the commandline

#123
post #3

I love that a third party is stepping up here, but it's incomprehensible to me that Google doesn't do this themselves. They're a search company, and they own YouTube. The YouTube data — including the subtitle files — is already sitting there on their servers; they don't have to scrape it, they just have to index it. What are they even doing? Fun thing to try: do a Google search with "site:youtube.com" in it. You get…

> Google search with "site:youtube.com" in it. You get basically nothing

That is not my experience! I regularly resort to this when the crappy inbuilt youtube search, which prefers to throw out algorithmic recommendations over returning actual search results, fails to come up with the goods.

Do you really get no results for, say: https://www.google.com/search?&q=intitle%3A%22thomas+brinkma... ?

Re: Show HN: YouTube Full Text Search – Search all of a channel from the commandline

#124
post #3

I love that a third party is stepping up here, but it's incomprehensible to me that Google doesn't do this themselves. They're a search company, and they own YouTube. The YouTube data — including the subtitle files — is already sitting there on their servers; they don't have to scrape it, they just have to index it. What are they even doing? Fun thing to try: do a Google search with "site:youtube.com" in it. You get…

I believe Bard has the capability of searching youtube transcripts

Re: Show HN: YouTube Full Text Search – Search all of a channel from the commandline

#125
Does this rely on an API key?

There have been earlier tools which permitted command-line / terminal access to Youtube, one of the best being mps-youtube:

https://github.com/mps-youtube/mps-youtube>

That permitted searches for terms and channels (though not subtitle text within channel AFAIK), for music specifically, and compiliation of either temporary or saved playlists, with the option to play through a full selection of videos. It also offered either full-video or audio-only playback.

Google killed it by throttling API-key access.

Discussed previously: https://news.ycombinator.com/item?id=32919545> https://news.ycombinator.com/item?id=28571421>

Re: Show HN: YouTube Full Text Search – Search all of a channel from the commandline

#126
post #3

I love that a third party is stepping up here, but it's incomprehensible to me that Google doesn't do this themselves. They're a search company, and they own YouTube. The YouTube data — including the subtitle files — is already sitting there on their servers; they don't have to scrape it, they just have to index it. What are they even doing? Fun thing to try: do a Google search with "site:youtube.com" in it. You get…

Some of us recall that when Google+ launched it lacked any search whatsoever.

That this was the case with a company whose name is synonymous with online search was ... simply mind-boggling.

The platform eventually did get search (actually a few different implementations), which varied between mostly useless to actually reasonably functional, though I'll note that HN's Algolia-based search is vastly more useful on an ongoing basis.

G+'s content, to the extent it survives at all, is largely on the Internet Archive's Wayback Machine which ... lacks search.

Re: Show HN: YouTube Full Text Search – Search all of a channel from the commandline

#127
post #3

I love that a third party is stepping up here, but it's incomprehensible to me that Google doesn't do this themselves. They're a search company, and they own YouTube. The YouTube data — including the subtitle files — is already sitting there on their servers; they don't have to scrape it, they just have to index it. What are they even doing? Fun thing to try: do a Google search with "site:youtube.com" in it. You get…

Have you tried searching through your call history on a Google phone? It's awful. You'd think they'd have a solid search built in but it's nearly useless. Especially considering you're usually searching for that number you only dialed once and isn't in your contacts and is for some reason excluded from your recents, so you go into your call history (strangely hidden behind a menu) and...there's no search function. WT…

As much as I like to generally, I wouldn't blame Google for this, rather I blame the entire field of "UX".

Re: Show HN: YouTube Full Text Search – Search all of a channel from the commandline

#128
This is great!

I had made something like this for my own use, but it was way more complicated. It took a user account, downloaded all the liked videos, ran it through some model and vectorized it, and then you could use natural language search to describe a scene in your history of liked videos and it would show you the timestamp and the thumbnail (with the link to start watching from the timestamp).

I ended up taking it offline because I didn't use it much, it was expensive and I couldn't see a path to monetization.

This solution being posted however, is really elegant, in particular because it is very resource-efficient.

Re: Show HN: YouTube Full Text Search – Search all of a channel from the commandline

#129

Earlier quoted context omitted.

In what cases is it unwise to rely on rowid over a id field?

I don’t think there are any. They are one and the same — if you create an integer primary key named id it is aliased to rowid : https://www.sqlite.org/rowidtable.html

I think the integer autoincrement primary key is more explicit / less mysterious than the implicit rowid. Even if most of us have run into that explanation in the sqlite manual.

Re: Show HN: YouTube Full Text Search – Search all of a channel from the commandline

#130
post #99

Earlier quoted context omitted.

Exactly! Google is in the attention business, the more you stay the more they make

Best case for Google is you click through and watch 10 videos without ever finding what you're after, so you try again the next day.

Watch 10 videos and then fall down a rabbit hole and forget what you were even there to look for - the even more best case!
Post reply on HN