#dev 2021-08-28
2021-08-28 UTC
jeremycherfas, nertzy_ and gerben joined the channel
hendursa1 joined the channel
# 
@RobbiNespu Implementing #webmention #indieweb (client side) on #hugo SSG
https://robbinespu.gitlab.io/posts/webmentions-hugo/ (twitter.com/_/status/1431538537875316736)
# 
capjamesg[d] Odd question: does anyone know anything about ideal cloud computing for compute-heavy operations (preferably that will not break the bank!).
# capjamesg[d] DO seems a bit pricey.
tetov-irc, rockorager and jamietanna joined the channel
# 
jamietanna capjamesg[d] give hetzner.cloud or Scaleway a go for potentially cheaper - but often to get compute heavy, it's gonna be fairly expensive :(
rockorager and christin joined the channel
# capjamesg[d] Hoping the Discord-IRC bridge sends a link ^
# capjamesg[d] !tell snarfed ^
hendursaga joined the channel
# capjamesg[d] sknebel I am just thinking about scale.
# capjamesg[d] The way things are going I could feasibly index a few thousand pages per hour with the PythonAnywhere server I am using.
# capjamesg[d] (But that server is limited by CPU usage so I'm only using it to relieve my computer of some work)
# capjamesg[d] This is indexing from IndieMap WARC files, not the web itself.
# capjamesg[d] Say I wanted to index 100 IndieWeb sites. That would take at least 10 days 😄 And that's from a file, not the web 🙂
# capjamesg[d] So would I.
# capjamesg[d] Ingesting from WARC is fast(ish) compared to actually crawling documents.
# capjamesg[d] Because then request time / processing needs to be taken into account.
# capjamesg[d] With a bit of Python wizardry I think I can get it down to about 15 mins for 5,000 documents.
rockorager joined the channel
# capjamesg[d] And my image search engine would take everything to another level.
# capjamesg[d] Because then images need to be downloaded, checked, compared, and optimized before being indexed. But I'm not even going to attempt indexing images.
# 
rockorager capjamesg[d]: Linode has $100 credits so you could test that out for free for a few months (I think you have 60 days to use the $100)
# capjamesg[d] Good call rockorager. Thanks!
# capjamesg[d] Good question. It's more preference and about not wanting to leave my desktop running almost maxed out while these operations go on.
# 
[jeremycherfas] I’ve got myself in a bit of a muddle since GitHub deprecated passwords. I have a PAT working fine on my desktop, but not on my laptop. And when I use the PAT on the laptop I cannot authenticate. Is there a simple option? Am I best off to create a new PAT for the laptop only?
chenghiz_ joined the channel
# [jeremycherfas] I guess I had better try that next. Tomorrow. I’ve had enough excitement for one day.
[chrisaldrich] and maxwelljoslyn[d] joined the channel
# capjamesg[d] Good example sknebel. I have tried not to use too many "luxurious" resources.
# capjamesg[d] (i.e. using smaller language models vs. larger ones with only slightly more accuracy)
jamietan1a joined the channel
# capjamesg[d] jeena.net is now in the index.
# capjamesg[d] There are at least 7,000 records collectively now.
# 
doosboox capjamesg[d]: how big is your db vs the amount of data you've parsed?
# capjamesg[d] Let's see.
# capjamesg[d] I need to index HTML for featured snippet support.
# capjamesg[d] That doesn't work for multiple suites quite yet because the logic was oriented around my site and its structure. But I will index HTML anyway for when that support comes (i.e. for "who is" queries to show h-cards).
# capjamesg[d] So I expect to have quite a big DB.
# capjamesg[d] 58MB.
# capjamesg[d] With 8329 records.
# 
@JamieTanna Right now I'm attending the #IndieWeb #IndieAuth pop-up to look at further improving the specifications, and making it easier for folks to implement and integrate (https://www.jvt.me/mf2/2021/08/1jpqv/) (twitter.com/_/status/1431682508140302337)
# jamietanna aaronpk rack-oauth2 (a common Ruby OAuth2 client) checks errors based on status code https://github.com/nov/rack-oauth2/blob/a8f1d2acd698cc13b6e102ba6fbf5a202d3b9d57/lib/rack/oauth2/client.rb#L150-L158
[dmitshur] joined the channel
# 
GWG The spec says 'expires_in' where did we get 'expires_at'? I'd prefer it. I convert to it now. https://github.com/indieweb/indieauth/issues/81
# 
Zegnat (For an example of me messing up, see https://github.com/indieweb/indieauth/pull/80 where I had to open a secondary PR because I only changed the public file)
KartikPrabhu joined the channel
tetov-irc and Rattroupe joined the channel
# 
[fluffy] Are there any codified standards for what an IndieAuth profile should include? I see that @rattroupe has added support for outgoing profiles but I don’t see any authoritative/canonical list of what fields are expected to be presented, just a vague-ish example in the IndieAuth spec itself.