#dev 2019-02-02
2019-02-02 UTC
[kevinmarks], [jgmac1106], [chrisaldrich] and [davidmead] joined the channel
# 
[jgmac1106] I don't get much detail (wordpress.com) but it's a nice feed of all my student posts in one place where I can quickly reply
# [jgmac1106] Oops wrong screenshot
# [jgmac1106] Will ask for volunteers to try different themes once I know who gets their sea legs faster
tw2113, [chrisaldrich], KartikPrabhu, [xavierroy], snarfed, leg, [asuh] and frankmeeuwsen joined the channel; frankmeeuwsen left the channel
# 
frankmeeuwsen Goodmorning indieweb
# frankmeeuwsen Time to start the day with a Jekyll problem and RSS. I already posted it on Stack Overflow to get going. https://stackoverflow.com/questions/54491107/local-jekyll-doesnt-generate-proper-rss-feed
frankmee_, frankmeeuwsen, tglobe, [jgmac1106] and [kevinmarks] joined the channel
# [jgmac1106] Does a tool exist that can take json and convert to mf2
# [jgmac1106] [kevinmarks] something unmung can do?
# [jgmac1106] I think it worked(ish)
# [jgmac1106] I lose the source url
# 
[kevinmarks] Hm. What kind of json is the question.
# [kevinmarks] Json does not define property names
# [kevinmarks] I suppose I could make every object into a h-* and every property into a p-*
# [kevinmarks] A bit like xoxo does
# [jgmac1106] is the raw json for one annotation but I am thinking I would do nested h-feeds, with each source being its own h-feed and each annotation an h-entry
# [kevinmarks] So that's a specific format. You could do the kind of thing unmung does by defining an html template.
# [kevinmarks] A lot of the unmung things are "fetch json from somewhere, put it through a jinja2 template with conditional logic in"
# [jgmac1106] What is jinja2?
# 
Loqi jinja2 is a python-based templating language for HTML with rich macro support that adds
{% %} syntax https://indieweb.org/jinja2# [kevinmarks] The node version of jinja is nunjucks
# [jgmac1106] ohh I use nunjucks all the time
# [kevinmarks] So we could make a toolkit for these kind of things on glitch
# [kevinmarks] Look at the unmung code, there are a bunch of templates that you can probably convert to nunjucks fairly easily
# [jgmac1106] so I just need to make a template for what an annotation in an h-feed would look like to start? I am thinking nested h-feeds though know everyone not a fan but I see the entire feed as an h-feed, and the each document as an h-feed, with each annotation as an h-entry
# [jgmac1106] What is unmung?
# Loqi Unmung is a service that turns Atom and RSS feeds into h-feeds https://indieweb.org/unmung
# [jgmac1106] the other idea would be to use p-category to group annotations but I think children makes the most sense
# [jgmac1106] "try to stop tantek exploding" best commit message ever
# [kevinmarks] You can use macros to build up more complex templates, but then you are writing code in nunjucks/jinja2
# [jgmac1106] yeah, today goal will be just to map the annotations, start there to get my brain around it..the rss to hfeed wasn't bad at all...got close
# [jgmac1106] and then make a mess of a code for long enough the judell takes sympathy and finishes project for me, he always said if we just showed him what an h-feed of annotations looks like he would add to export tools
# [kevinmarks] I use feedparser in the middle to get uniform json from feeds, and a bit of python to transform dates
# [jgmac1106] kevinmarks+=
# [jgmac1106] kevinmarks++
tglobe joined the channel
# 
@jgmac1106 ↩️ @toolness making progress on the webmention badges issued my first build set as an h-feed:http://edu407.jgregorymcverry.com/badges/helloworld.html the criteria and evidence deliberately mismatched for a training video (https://quickthoughts.jgregorymcverry.com/s/25iB2k) (twitter.com/_/status/1091678993324171264)
# [kevinmarks] You could map the TextQuoteSelector to a fragmention
tglobe_, swentel, davy__, davy_, [schmarty] and barpthewire joined the channel
# [jgmac1106] What is fragmention?
# Loqi A fragmention is an extension to URL syntax that links and cites a phrase within a document by using a URL fragment consisting of the phrase itself, including whitespace https://indieweb.org/fragmention
# [jgmac1106] all it is just ## insteasd of one in an id?
# [jgmac1106] then each annotation wouldn';t need to be an h-entry and each doc could be which would be better
# [jgmac1106] question on the example on this page: https://github.com/chapmanu/fragmentions why does the ##pursuit jump to <strong>?
# [jgmac1106] duh the word pursuit
barpthewire, tglobe_, tglobe and [kevinmarks] joined the channel
# [kevinmarks] Not ## any more
# [kevinmarks] One # then words separated by %20
# [kevinmarks] As id can't have space in so that is enough to disambiguate
# [jgmac1106] So it's just a normal ID with the text segment,alright.
# [jgmac1106] What would you use to link back to the document a u-url?
[tmiller] joined the channel
# [jgmac1106] @zegnat if I can figure out how to take the html or json export and map it to mf2 just use hypothesis...though their terms of service require every annotation to be in public domain...if that matters
# [jgmac1106] Didn't Karthik make an extension>
# [jgmac1106] yeah#this%20specific
# [jgmac1106] i am there now
# [jgmac1106] funny thing is I am actually now trying to take something that uses RDFa and get it into HTML with mf2
# 
Zegnat Hmm. /fragmention only links to Jonathan’s extension. So I guess we don’t have an implementation for enabling fragmentions on websites that do not themselves use JS for it?
# [jgmac1106] but that is just for your own site correct
# [jgmac1106] oh yeah you are no JS so that is tough....could always try ctrl+c
# Zegnat [kevinmarks], is the wiki outdated? https://indieweb.org/fragmention#URL_encoding says to use + rather than %20
[schmarty] joined the channel
# [jgmac1106] yeah I didn't even think it was possible for a browser to jump to page that doesn't support it...that is cool it works this way
tglobe joined the channel
# [kevinmarks] Hmm. I'll check the wiki when I'm on my computer. I thought we edited the old stuff out
[tantek] and [schmarty] joined the channel
# [jgmac1106] can you nest <details>
# [jgmac1106] I guess I can use seperate, ones and style them over a bit if they are in a thread, probably better
# [kevinmarks] Yes, they nest well if you want an outline style layout
# [kevinmarks] Chrome shows an arrow by default, Firefox doesn't, so you may need to style it a bit.
tglobe and tglobe_ joined the channel
# [jgmac1106] okay will go back and nest away
# [jgmac1106] was trying to think how you display reply to annotations: https://gist.github.com/jgmac1106/726a8399dce96e28e0ae9cfd1ad288c3
# [jgmac1106] that is my first attempt, interesting though as this would cause an h-entry to be nested in an h-entry is that wrong?
# [jgmac1106] then the next question do the js fragmention tools work with: https://mozilla.github.io/pdf.js/?
# [jgmac1106] if not isn't much of a point for me to do this as everything annotated is a pdf
# Loqi Dat is a protocol for transferring files peer-to-peer https://indieweb.org/Dat
# [jgmac1106] no the thing I really want to start experimenting with is to see if I can set dispaly sizes using some kind of css animation timer and then I could use media queries to do all kinds of weird animations cell by cell, making things as big or small, overlaying grids.....I have no idea if you can do this but I see these really cool animations in my head of stuff growing and collapsing
# [jgmac1106] little bit of dis-in-dat
# [jgmac1106] I love idea of Beaker as a fully remixable read/write experiment that is what I think is so cool
# [jgmac1106] read/write environment, just taking any page and doing what you please,..cool playground
# [jgmac1106] ..well that used to work
[Rose] and snarfed joined the channel
# [jgmac1106] Just thinking nested details and summary would be a nice display for threaded webmentions
[kevinmarks] joined the channel
# [kevinmarks] There are dat proxies
# [kevinmarks] I have a dat page, and dat slides for a talk
# [kevinmarks] Add webmention to fritter?
# [kevinmarks] I pointed dat at my local git dir of the site
[metbril] joined the channel
snarfed joined the channel
DenSchub joined the channel
# 
[metbril] The basic concept of Grav is just that. But planning a growing site (in the number of posts that is) is challenging. If GitHub hosting would allow (easy) saving webmentions and micropub, that would be a big step forward. Micropub from voxpelli does it, but it needs heroku which is no-ssl for the free plans. Choices, choices...
# 
petermolnar this was one of the reasons that put me off grav: the cache invalidation included image regeneration, which hogged up my server for half an hour
snarfed, snarfed1 and [tantek] joined the channel
# Zegnat For anyone looking into creating “seemless” data back-and-forths from their website to the backend, this looks interesting: http://websocketd.com/
[cleverdevil] joined the channel
# 
[cleverdevil] So, I have made some great progress on a NNCaaS today.
# [cleverdevil] Features: lookup "pseudonyms" (as I call them) based upon a domain name, cache for domains with auto-timeout of the cache after a configurable period, the ability to search h-card "name" declarations.
# [cleverdevil] Oh, and the ability to send in content in a special format and have it automatically parsed and re-configured to various targets.
# [cleverdevil] It looks for a "mention" of the format `@
{somedomain.com}` and then will replace all instances of that with all of the different variants by syndication target automatically, in both plain text, and HTML.# [cleverdevil] I'll try and deploy it in a bit if anyone wants to play with it.
# 
gRegorLove What's NNCaas?
# gRegorLove oh, nickname cache
# gRegorLove nice!
# [cleverdevil] I originally decided to use MongoDB, but it looks like its going to be a lot cheaper for me to run if I just piggyback on my existing MySQL RDS instance, so I am porting it over now.
# [cleverdevil] Won't take too long 🙂
# [cleverdevil] Once its done, I'll deploy and let people poke around!
# [cleverdevil] Meh, I'll just run a little local MongoDB on one of my EC2 instances.
# [cleverdevil] Good enough 🙂
snarfed and [asuh] joined the channel
tantek, snarfed and [tantek] joined the channel
# [cleverdevil] OK, check it out - https://pseudonym.cleverdevil.io/domains/cleverdevil.io/all
# [cleverdevil] You can try any domain that has rel-me links and, optionally, an h-card with a 'name' property.
# [cleverdevil] You can also search the database - https://pseudonym.cleverdevil.io/search/jonathan
# [cleverdevil] It searches against full word terms, though, not partial.
# [cleverdevil] You can also send an HTTP POST to https://pseudonym.cleverdevil.io/content/format with a "content" parameter and it'll auto replace mentions
snarfed joined the channel
# [cleverdevil] `http post https://pseudonym.cleverdevil.io/content/format content="Hello @
{eddiehinkle.com}, I hope you're well."`# [cleverdevil] results in:
snarfed joined the channel
# [cleverdevil] With that, I am taking a break 🙂
[eddie] and oodani joined the channel