00:24:35DogsRNice_ (Webuser299) joins
00:25:58DogsRNice quits [Ping timeout: 265 seconds]
01:17:57prebala-publici quits [Read error: Connection reset by peer]
01:19:45prebala-publici joins
01:22:02prebala-publici quits [Remote host closed the connection]
01:48:18<HP_Archivist>LoC has it's own policy regarding former WH.gov captures for former admins. I assume people here or at IA saved as much as possible before the current admin updated WH.gov in the past 24 hours?
01:53:27<@JAA>The previous sites are preserved anyway.
01:53:45<@JAA>https://trumpwhitehouse.archives.gov/
01:54:42<@JAA>I'm not sure if whitehouse.gov was archived here before it was moved there and replaced by the Biden administration site, but it was run through ArchiveBot after the move, and I think IA does crawls of .gov in general at the end of term.
01:55:49<HP_Archivist>JAA: Yeah I know there are other means to which it is captured regardless, sounds good though
01:58:09<@JAA>Ah yes, I ran whitehouse.gov through AB pre-election in October. No run between that and the transition.
01:58:57<@JAA>Meh, the WBM calendar for whitehouse.gov doesn't work on the 19th.
01:59:15<@JAA>('e is undefined', like the other day.)
02:18:06Terbium quits [Ping timeout: 265 seconds]
02:18:18Terbium joins
02:18:35@JAA quits [Ping timeout: 265 seconds]
02:18:39JAA (JAA) joins
02:18:39@ChanServ sets mode: +o JAA
02:20:09<HP_Archivist>JAA: So there were no captures in the WBM on Jan 19th for Whitehouse.gov?
02:23:00<@JAA>HP_Archivist: There definitely were, but you can't list them. Accessing via URL works fine though. E.g. https://web.archive.org/web/20210119235526/https://www.whitehouse.gov/
03:57:12<HP_Archivist>JAA: I still don't understand. By list what do you mean?
03:58:03<@JAA>Hover over the date on the calendar. It shows an error instead of a list of snapshots on that date.
03:58:37<@JAA>Well, if the calendar loads in the first place instead of throwing a 500 Internal Server Error.
03:59:01<@JAA>Oh, now the 18th and 19th have disappeared from the calendar entirely.
03:59:09<@JAA>That's https://web.archive.org/web/*/whitehouse.gov just to be clear.
03:59:13<HP_Archivist>No, I'm on it now and it doesn't show any captures or dates for 18th or 19th
03:59:16<HP_Archivist>Yeah exactly
03:59:33<@JAA>Yeah, earlier it did but then produced the 'e is undefined' error in the popup.
03:59:47<HP_Archivist>500 internal error..
03:59:51<HP_Archivist>I'll try later
04:00:27<@JAA>Safe to say that something's broken at the moment. :-)
04:00:55<HP_Archivist>I've learned not to expect perfection from the WayBack Machine heh
04:09:32fionera quits [Ping timeout: 264 seconds]
04:11:00fionera (Fionera) joins
04:13:27qw3rty_ joins
04:16:44qw3rty quits [Ping timeout: 240 seconds]
04:17:38<atphoenix>yea, time machines are still imperfect. Someday maybe it'll be a solved problem, along with fileformats and emulation.
04:46:55DogsRNice_ quits [Read error: Connection reset by peer]
05:31:36Ajay quits [Remote host closed the connection]
05:43:51Ajay joins
07:47:24prebala-publici joins
07:47:24prebala-publici quits [Remote host closed the connection]
07:47:45prebala-publici joins
07:47:47prebala-publici quits [Remote host closed the connection]
07:48:07prebala-publici joins
08:39:14<HP_Archivist>JAA or atphoenix: Has Archive.org been *actually* capturing YouTube videos? Not sure how long this has been going on - or maybe I just never knew - but I went go save this interview in the WBM and the entire video was already crawled
08:39:21<HP_Archivist>And playable in the WBM along with comments
08:39:23<HP_Archivist>https://web.archive.org/web/20210122083609if_/https://www.youtube.com/watch?v=eYVv4uQjhQE
08:41:48<HP_Archivist>I guess was under the assumption that WBM captures of YT videos never worked/playback wasn't possible
08:47:11<atphoenix>I know that SPN can save twitter videos
08:49:14<HP_Archivist>Well like I said, maybe I just never realized, but the whole damn video along with comments is viewable and loads in the WBM
08:49:21<atphoenix>I wonder if this is related to that mechanism. Maybe it starts downloading it upon actual access, if the video is still up?
08:49:58<HP_Archivist>I don't think so. I've tried saving YT videos in the WBM a few years ago. They were never playable like this
08:50:19<HP_Archivist>Unless this video has been tubeup'ed and somehow the YT video id is matched to the url..?
08:50:54<HP_Archivist>One caveat, I can't seem to load responded-to comments or sub-comments
08:51:40<atphoenix>trying to follow a link to another video got me "Sorry, the Wayback Machine does not have this video (0T4S4nreoKo) archived (or not indexed yet).
08:51:41<atphoenix>"
08:52:06<HP_Archivist>^^
08:52:07<HP_Archivist>Hm
08:52:38<atphoenix>but it does load the comments
08:53:57<HP_Archivist>Yeah that's what would happen to me in the past, only comments and the sidebar of suggested videos would be captured
08:54:01<atphoenix>that linked vid was from the 'AD Vids' channel
08:54:03<HP_Archivist>not actual videos
08:54:42<HP_Archivist>Wait, what? That's not the same channel
08:55:12<HP_Archivist>AD Vids is a channel responsible for Tom Snyder interview uploads, I know cause I crawled the channel in #youtubearchive
08:56:38<atphoenix>yes, that channel is what I ended up on when I clicked a related video from your example
08:56:46<atphoenix>the vid id is in the error message
08:57:21<atphoenix>as for your example, I don't see the usual WBM info box at the top
08:57:53<HP_Archivist>Oh, I see what you mean now ^
08:58:05<atphoenix>oh wait...I did a reload of the Tom Snyder video
08:58:05<atphoenix>https://web.archive.org/web/20210122083609if_/https://www.youtube.com/watch?v=eYVv4uQjhQE
08:58:13<atphoenix>it behaves like your first example now
08:58:38<HP_Archivist>And yes, I noticed that too. But if you search the link in the WBM timeline it will come up
08:59:00<atphoenix>it really appears there is some new behavior going on. It seems to have saved the YT video upon first access!
08:59:16<atphoenix>pretty much how we see twitter do it
08:59:43<atphoenix>I mean how we see IA save twitter
08:59:58<HP_Archivist>Huh. I wonder how long this has been possible??
09:00:17<atphoenix>I wasn't aware of this until you mentioned it
09:00:26<HP_Archivist>I mean, it's very slow to load that interview but it does load. and it loads the comments (even more valuable..)
09:00:31<atphoenix>i was only aware of twitter vid
09:00:44<HP_Archivist>Hey JAA do you know anything about this?
09:01:02<atphoenix>are you logged in with an IA browser extension? I am.
09:01:19<HP_Archivist>No, I just loaded it in FF, not logged into anything, no extensions
09:01:24qw3rty_ quits [Ping timeout: 240 seconds]
09:01:26<atphoenix>I'm using the github version
09:01:27<atphoenix>ok
09:02:04<HP_Archivist>It seems that somehow the video file is grabbed, I wonder if they're using ytdl, I doubt on the fly video conversion
09:03:05<atphoenix>I bet the missing WBM timeline bar is missing because of that 'if_' tail. I know there are various tails available.
09:03:15<atphoenix>I get the normal timeline view with https://web.archive.org/web/20210122083609/https://www.youtube.com/watch?v=eYVv4uQjhQE
09:03:26<HP_Archivist>It seems this has been available since 2019? https://forum.videohelp.com/threads/391951-How-to-download-a-YouTube-video-archived-by-Wayback-Machine
09:03:47<HP_Archivist>a quick search for 'save youtube videos waybackmachine' brought up that thread
09:04:15<atphoenix>this is in the 'Live Web Proxy Crawls' collection
09:04:56<HP_Archivist>I don't follow, an actual collection on Archive.org?
09:05:20<atphoenix>https://archive.org/details/liveweb
09:05:47<atphoenix>one of many collections on IA. Most are private. AT's collections are mostly public.
09:06:03<atphoenix>but some private collections allow viewing via WBM
09:06:20<HP_Archivist>Hm
09:06:48<HP_Archivist>I wonder if this works on first attempt on any YT video...
09:07:03<HP_Archivist>I tried that interview using the GH chrome extension
09:07:15<HP_Archivist>But it looks like it was crawled in 2018 going by the timeline
09:08:23<HP_Archivist>ivan - though I think your collection would far exceed what's in the WBM, are you aware that YT videos can be saved and playback is possible?
09:08:37<atphoenix>I'd compare this to what has been happening with the URLs project that isn't getting 'requisites'. It only gets the HTML text of a webpage. But if you visit the page in WBM, WBM will try to find the images either in WBM...or on the source server if the source server is still alive.
09:09:03<atphoenix>ivan isn't in here
09:09:29<HP_Archivist>Oh, my bad. Thought he was
09:09:53<atphoenix>so i guess now trying to visit YT pages on WBM might trigger the WBM to save the related videos...
09:10:17<HP_Archivist>But yeah makes sense. I'm still confused how it would grab the video file. YTDL pulls down the highest quality stream possible. I wonder how WBM does it and with what
09:10:46<HP_Archivist>And... that makes archiving very easy...
09:11:20<atphoenix>it looks like ytdl is configurable in terms of the quality it goes for. I haven't tried to tweak those options in my own tests yet.
09:12:03<HP_Archivist>Yeah it is. I've been using YTDL for almost 2 years. And that's what ivan's project uses, but many instances and with a custom script
09:12:11<atphoenix>my guess is IA would use ytdl in the background. I know ytdl works on many other sites besides YT. I use it on twitter vids myself.
09:13:06<HP_Archivist>If they don't use ytdl, how else would they grab the streams, still kind of floored that I can watch YT videos in the WBM heh
09:14:07<HP_Archivist>The entire 'YouTube experience', something that YTDL does not capture, is captured in the WBM is would seem. The html/page, the suggested/recommendation side bar, most comments, the up/down vote on each comment, etc
09:14:42<atphoenix>I don't know what magic they are using to put it all together. Maybe it's just something that happens when warcproxy is used? I don't know enough to say with confidence how they do it.
09:15:32<atphoenix>it's not perfect as you've already noted re:comments. It is still impressive.
09:17:07<HP_Archivist>Yeah, goes beyond my knowledge. Idk if this is a game changer or not when it comes to mass archiving YT. I can say that this adds another tool to our archiving kit as it were, assuming WBM captured videos are not relying on anything externally to still be live on the live page
09:18:09<HP_Archivist>There is a way to test that - I could always upload a test video to my YT account, captured it, then delete the video from my channel
09:18:26<HP_Archivist>See if it still plays in WBM
09:20:21<HP_Archivist>atphoenix: I might try this tomorrow, will let you know what happens. Regardless, cool stuff indeed
09:30:29qw3rty joins
09:48:22<mgrandi>is the WBM having lots of issues lately? it like cannot return a save page request through this library i'm using : "https://web.archive.org:443 "GET /save/https://instagram.com/Merlalias HTTP/1.1" 520 None"
09:58:59<HP_Archivist>Keep trying or try at a later time mgrandi
09:59:16<mgrandi>yeah, my script waits 5 minutes between each attempt
09:59:26<mgrandi>its just been particularly bad this week it seems
09:59:40<HP_Archivist>No surprise given the events this week
09:59:51<HP_Archivist>But yeah, maybe come back and try 30-60 minutes later
09:59:57<HP_Archivist>That's usually what I do and it does work then
10:00:14<mgrandi>yeah, luckily i'm in no rush
10:00:53<mgrandi>the wayback machine could use a real api though, this library claims you just hit archive.org/save/<URL> and then look in the headers and its kinda hacky
10:01:30<HP_Archivist>There is an API
10:01:51<HP_Archivist>I never use it, requires knowledge of python
10:02:32<mgrandi>https://github.com/akamhy/waybackpy/blob/master/waybackpy/utils.py#L195
10:02:56<HP_Archivist>Oh, nice
10:03:15<mgrandi>well that is the unofficial library, but it has the notes about i guess /save does not return JSON
10:03:34<mgrandi>https://archive.org/help/wayback_api.php says it only really returns json for the querying if a site has been saved i guess
10:04:41<mgrandi>and since it just returns the HTML of the page its hard for it to tell if it succeeded or not, or if its just a html error page i guess
10:05:15<HP_Archivist>Idk if this would help but here's IA's documentation for api use https://archive.org/services/docs/api/
10:05:31<HP_Archivist>Like I said, I'm pretty novice when it comes to use of it or python
10:06:16<mgrandi>yeah, while confusing, i think that is referring to the Internet Archive itself and interacting with it
10:06:17<HP_Archivist>I did try to use tubeup through the api 6 months back and successfully had it configured for a bit, was nice to use. But never for wbm
10:06:33<HP_Archivist>Right, yeah, what do you mean, use for wbm?
10:06:50<mgrandi>while the WBM seems to be a separate system
10:07:20<HP_Archivist>^^ good point
10:07:21<mgrandi>my script takes a config file and then tries to do a WBM save of various urls in my configuraiton file
10:08:27<HP_Archivist>So you're trying to do batch URL saves?
10:09:08<mgrandi>i guess, basically i'm looping over them and using that library ^
10:09:52<mgrandi>or at least automate them, without having me to go to the save page now site every time
12:22:56Jake5 (Jake) joins
12:24:24Jake quits [Ping timeout: 240 seconds]
12:24:24Jake5 is now known as Jake
13:14:00<mgrandi>so apparently doing `web.archive.org/save/https://www.instagram.com/lilsimsie/` is 520ing every time , but putting that url into the SPN page works fine?
13:14:08<mgrandi>ill try again tomorrow i guess
14:15:51Arcorann quits [Ping timeout: 265 seconds]
14:42:51prebala-publici quits [Client Quit]
14:43:10prebala-publici joins
15:03:35<@JAA>HP_Archivist: No idea. I know that there are some videos in the WBM, but I've always seen it as a '1 % of the time, it works every time' situation.
15:04:13<@JAA>mgrandi: There is a proper API for SPN2, but you'd have to ask for access via email I believe.
16:06:27<atphoenix>JAA, was your reply to HP_Archivist missing a word after the comma?
16:25:51<@JAA>atphoenix: Uh, no.
16:26:45<atphoenix>this is the part that was unclear: '1 % of the time, it works every time'
16:27:45<@JAA>https://www.youtube.com/watch?v=k8djas5pIvk
16:39:28rsn_ joins
16:43:46rsn quits [Ping timeout: 252 seconds]
17:09:17<atphoenix>ah, ok. I didn't catch the reference
17:13:29<@JAA>Looks like the WBM has issues with snapshots from around the 18th and 19th currently in general. I'm getting 504s and whatnot.
17:24:10OrIdow6^2 (OrIdow6) joins
17:25:01OrIdow6 quits [Ping timeout: 252 seconds]
18:44:33DogsRNice (Webuser299) joins
20:32:34<mgrandi>I'm fine with not having a full api, but this url is giving me a 500 every time , and then twitter urls have a 30% chance of being corrupted and not saving the actual twitter page
20:54:13Arcorann (Arcorann) joins
21:36:45<mgrandi>yeah, tried it again with that instagram link and it just 520s and says job failed
21:38:28<@JAA>mgrandi: SPN and SPN2 are completely separate systems. When you use /save/URL, that's SPN. When you 'put that URL into the SPN page', that's SPN2.
21:39:08<@JAA>SPN basically just retrieves that URL while writing to WARC. SPN2 runs a browser for the retrieval including scripts and whatnot.
21:39:12<mgrandi>i guess this library is using spn1 then...is there any way to have an 'api like interface' for spn2? like calling web.archive.org/save ?
21:39:16<@JAA>So yeah, they behave very differently.
21:39:32<@JAA>Twitter only works correctly on one of them as well.
21:39:51<mgrandi>(twitter has that issue with spn2 as well, or its only spn2, i'm not sure)
21:40:41<@JAA>Well yeah, the SPN2 API is available but not public. I guess you could reverse-engineer what happens in a browser, but I wouldn't recommend that.
21:40:48<mgrandi>ugh
21:40:58<mgrandi>well i don't really need a full api, the SPN1 one isn't really an api either
21:41:36<@JAA>I guess you could also script the interaction with /save/ with a headless browser or something.
21:42:02<@JAA>But the proper API is definitely easiest.
21:42:38<mgrandi>is it hard to get access to this? i'm like, saving on the order of 10 urls a day maybe
21:46:23<@JAA>No idea. Can't hurt to ask though. I'd assume that if you tell them that, they'd be fine with it.
21:46:51<@JAA>Can't imagine they want to lock that down since it's easier to monitor etc. than people trying to emulate the browser stuff.
22:03:46OrIdow6^2 is now known as OrIdow6
22:06:37Arcorann quits [Ping timeout: 265 seconds]
22:07:12<OrIdow6>That video was captured by IA's Youtube collection (https://archive.org/details/youtubecrawl?sort=-publicdate)
22:07:58<HP_Archivist>JAA: No worries. Just surprised to have found that YT videos, some, are playable and captured in WBM
22:10:19<@arkiver>yeah they are playable if archived and indexed
22:26:20<mgrandi>who do i ask @JAA ?
22:46:52<@JAA>mgrandi: info@archive.org, cf. https://blog.archive.org/2019/10/23/the-wayback-machines-save-page-now-is-new-and-improved/
23:21:24Doranwen quits [Ping timeout: 240 seconds]
23:23:40Doranwen joins
23:23:41Doranwen quits [Changing host]
23:23:41Doranwen (Doranwen) joins