#poki Lapo: Toki Pona library
1 messages · Page 4 of 1
no its the DDNS i use normally as my IP is dynamic,, lapo.pona.la would then be pointed at it with a cname
(or lipu-ale.pona.la, which is way better imo :D)
ken
ni li elena tawa mi
or lipu
or is that something already
lipu.pona.la was taken once before but is currently free
in fact lipu.pona.la is currently one of our draft PRs to lapo
nice
sidenote: kvarg and lingonberry jam hit hard
(finnish delicacy: rotten milk and the most acidic berry on the northen hemisphere :D)
dkdhkdjr
hmmmmm
neat to see the tables work, i never tested em
sad to see the image overflows, i never tested em (lmao)
just put some of that sweet max-width: 100% on that sucker
ye ofc
or i guess overflow: scroll works too
@nova gale @heady vigil https://kulupu-lapo.github.io/open/
its missing a lot of vital features but it exists
Nice 👍 the images are still broken tho
ye cuz i still haven't touched em
lmao. Nasi is broken but way less broken than i expected
Nasi is a giant pile of html
:O
yay
since when did you have a fro tend /j
@heady vigil @vague coral i would like to get a window sometime this April where no pull requests are open, so we can apply schema changes to the entirety of the poki
oki
to this end lets try to wrap up the five PRs/drafts still open and ill try to help with that
what are they again
an easy one is:
https://github.com/kulupu-lapo/poki/pull/23
^ i will close this without merging, so that we can repeat this later without conflicts
@heady vigil please look over the changed files in https://github.com/kulupu-lapo/poki/pull/7/files and tell me if anything still needs work before merging
yes, this is an incomplete list of songs, but we can work on more of them after the window
you can merge preemptively
explain what you mean by preemptively
-# a mi jo e https://github.com/hecko-yes/poki/tree/uta-monsuta kin pi tenpo weka,,
-# (mi ken ante e schema kepeken wawa ilo mi la ike ala)
i just need a 👍 on whether the files that are there are all good
do i need to lay out like more restrictive submission requirements for utala musi so that formatting like.. works for this project?
so right now we are sort of doing twice the effort
you are maintaining your list, we are maintaining ours
in an ideal world, we would be able to sort of flip maintenance on its head: Lapo would be responsible, and you could pull from us and formatting should remain perfect
and then you have like, less maintenance
i need to redo the whole utala pona site at some point
but right now i suppose its just, we try to stay in lockstep
oki
i don't think you need to introduce any requirements
(similar to how no one needs to maintain dictionary data themselves anymore, but people might still want new pages for displaying it)
TIL sapaa contributed to the tokirap
dying giggling @ kala pona Tonyu followed by Peter Kropotkin
but ye, we're fixing that soon
and "Contributors to Wikipedia" a a
seeing such silly authors in the frontend is making me like the original-authors idea even more
@vague coral can you rebase your pr
the lentan one
i tried but running into perms issues strangely enough given im the owner
No conflicts with base branch
Changes can be cleanly merged.
sem ea
yea ik
your validation errors include things that have been fixed on main since
to my knowledge
so i want to avoid thinking about them
@heady vigil @jagged burrow @vague coral
idea:
- make a cli command to create a new blank article
- populate that blank article with every metadata field we have
- default to
nullfor a lot of them, maybe the present day for the date - make every(!) field required, but nullable
(counterargument: do this, but only when we're reasonably sure the schema is good?)
@vague coral https://github.com/kulupu-lapo/poki/invitations
so that your future branches can be in the same repo
Sorry, we couldn't find that repository invitation. It is possible that the invitation was revoked or that you are not logged into the invited account.
depending on how much we want new archivers, we could also make that a (quite simple) webapp with github apps and something like wysimark
oh i think markdown is sufficiently intuitive for discord dwellers
its the frontmatter that i think is a slight time sink
i see people copypasting the same file, then potentially forgetting to redo some of the fields, resulting in faulty metadata
which of course we can fix over time, but preventing it by obsoleting file copying sounds better
its just that the CLI command is not that easy to implent so that windows people can also run it easily
i mean there is no python or js preinstalled
seme la ilo penpo li kasi ala e ni?
kasi
thats true
i've been using the schema example from our readme as a template for new files
which works good enough for me
i don't know if we can completely stop contributors from the file copying you describe, but we can at least discourage it by nudging folks to using our template, or providing a CLI tool
Which editor do you use?
@carmine quarry @heady vigil @jagged burrow @vague coral
kate
sublime text
VSCode
so making an extension wont even solve problems :D
- ona li lon esun la ona o lon ala poki anu seme
- ona li kama ante lon tenpo mute la ona o lon ala poki anu seme
- jo
- jo
- jo ala (tan mi)
- sona ala
- jo
- jo ala?
- jo
- jo ala?
ona li lon esun la ona o lon ala poki anu seme
the link has pdfs freely available
ona li kama ante lon tenpo mute la ona o lon ala poki anu seme
yep
@vague coral picking up jan Lentan's blog from you, gonna try to merge it today
@vague coral check my work: https://github.com/kulupu-lapo/poki/pull/33
leave a 👍 on this message if no more changes are required
@vague coral https://github.com/kulupu-lapo/poki/pull/27/files la, your sources no longer exist
(sorry for pinging so many times, but this is how i guarantee we get things done)
I think it would be interesting to make a toki pona specific spider for crawling the pona web,, we could make it follow links like three sites deep and find text in toki pona which is not included in lapo yet and list these sites somewhere for human inspection
totally doable, if/when someone has the time and effort to put into it
for kulupu Lapo in general, probably not going to count as a required feautre
yea exactly,, can't wait for the summer when i actually have time for Lapo :D
depending on how annoying lipu.pona.la is, i might try to merge everything today
and then we can have fun cleaning up and standardising the schema
"ahh, the sun is shining and weathers are warm (unlikely in finland) it's time to close the blinds and archive some toki pona texts"
The source code and instructions to create your own version of Wiby.
@vague coral bumping this in case you missed it
mi sona
poan
from scratch is more silly
,,mi kama wile ni
@vague coral im taking over the lipu.pona.la PR
whats the license on lipu.pona.la articles
ala
@vague coral whats the source of these duplications?
they shouldn't break anything but they're sus
in that case it seems likely that one of the two in each pair currently does not have a file
you might want to add those manually
im playing around with merging these "messenger-like posts" into one file, we'll see how that goes
@vague coral https://web.archive.org/web/2/https://lipu.pona.la/kita/akesi-li-wile-lon-nena-kepeken-sitelen-lasina the metadata claims this exists but it doesn't
a nasa · ilo wepaka li moku ala tan seme
@vague coral
---
title: toki lili pi lipu pona
authors:
- kijetesantakalu
- telosuwi
- jan Tepo
- jan Nami
- jan Kita
date: 2022-08-08
license: null
sources:
- https://web.archive.org/web/2/https://lipu.pona.la/kijetesantakalu/ale-o-toki
- https://web.archive.org/web/2/https://lipu.pona.la/telosuwi/toki-mi-jan-pi-telo-suwi
- https://web.archive.org/web/2/https://lipu.pona.la/kijetesantakalu/toki
- https://web.archive.org/web/2/https://lipu.pona.la/jan-tepo/jvyi6k2pd6
- https://web.archive.org/web/2/https://lipu.pona.la/jan-nami/toki-a
- https://web.archive.org/web/2/https://lipu.pona.la/jan-nami/lipu-pona-li-toki-pona-taso-ala-toki-pona-taso
- https://web.archive.org/web/2/https://lipu.pona.la/kita/2022-08-07-19-41-15
---
**kijetesantakalu** (2022-08-06):
ale o toki
**telosuwi** (2022-08-06):
toki! mi jan pi telo suwi!
**kijetesantakalu** (2022-08-07):
toki
**jan Tepo** (2022-08-07):
*asdf*
**jan Nami** (2022-08-07):
toki a!
**jan Nami** (2022-08-07):
lipu pona li toki pona taso ala / toki pona taso?
**jan Kita** (2022-08-08):
#meta
· jan Nami o · kama pona ·
· sina [wile sona](/jan-nami/lipu-pona-li-toki-pona-taso-ala-toki-pona-taso) e ni → ma ni la lipu o toki pona taso anu seme ·
· mi pali e ma ni tawa ni → jan li ken pana e lipu sin pi toki pona kepeken pali lili ·
· taso · sina wile pana e toki ante lon poka la ni li pona ·
· (mi o toki e ni lon lipu lawa anu seme a a) ·
· ante la · sina kama tan seme · mi lukin ala e sina lon `ma pona pi toki pona` pi ilo Discord ·
suwi
one of your dumped files started with a <center> and no name and ended up receiving the filenamecenter.md
thats. so funny
lon ala kin
@vague coral go through the pr and suggest changes
also check that ive not lost any posts along the way
lukin lili mi la pona
a taso · "akesi li wile lon nena" la · mi pana e tenpo pali lon lipu la ilo li pana sin e ona e tenpo pana ala · ni li seme tawa sina
so august was the first time it was actuslly posted online, and the june date on lipu.pona.la was just back-dating it? anu seme
ni
@vague coral @heady vigil @jagged burrow @nova gale
catching up to speed:
- poki Lapo currently has no open PRs
- open Lapo builds successfully
this is a good time to do big, sweeping schema edits
wawaaa
its also a good time to make small changes, like browsing open Lapo and fixing any file thats horribly broken
(theres like a thousand of them so youre gonna be browsing for a while)
is there a 1-to-1 correspondence between files that currently have an original-title and files that must now have an original?
files that have original-title or translators, yes
the sources for the lipu mama are either not listed at all or grouped in with the translation's sources. so you'd have to some manual work here
are any original-sources listed anywhere?
sina ken pana ala e tan pi lipu mama · mi pana tan ni taso → "o lukin · nasin mi li ken e ijo sin pi lipu mama kepeken nasa ala"
okay so my biggest problem right now is that
i don't actually know of any non stupid, non manual ways to update the schema for a thousand yaml frontmatters, which also have comments in them that itd be nice to keep
and presumably do so for not the last time ever
mi weka li kama lukin ala e toki
okay so ruamel works, but it does something else thats annoying: replacing every null with nothing
because i guess i still don't understand whether field: and field: null are the same thing in yaml
and im not sure its worth it rn
it would be super semantic and easy to do this in go
it depends on the parser,, for example the already mentioned GoYAML understands "null" as "nil" and empty as the empty variable of whatever type the field normally is OOB
I can do the poke everything commit if we get the schema locked down
we have two big changes we need to make:
title:
translators:
-
original-title:
authors:
-
becomes
title:
authors:
-
original:
title:
authors:
-
and all dates require yyyy-mm-dd, with an added date-specificity enum field for those which are less specific
but i wouldn't mind just a general cleanup, e.g. deciding if lists are always indented or unindented
What about games? Games are interactive and while they might have plaintext segments(dialogue, story, intro, ending) it might not be possible to just flatten a game into plaintext (some parts might happen in a different order, some parts might change depending on player actions, some text might only make sense in the context of actions the player just did, and so on.)
Maybe extracting the plaintext from the source code just in the order it appears?
toki_lawa_ike23 = "toki a, %jan_tawa o. o sona e ni: sina kama moli tan luka mi."
if you're able to convert the game into a script (think like a shakespeare play!) then yea! we can store it
but most games aren't quite so linear and dialogue-heavy
help is always welcome! you can take one of the sources that is currently unclaimed and start converting it to markdown and writing appropriate metadata for it. we can guide you through the process
should i also make all fields required but nullable
id be down for that
you're only doing automatic manipulation of files rn, not manual, right?
it would be good to have the option for back-and-forth feedback
without making your work ridiculously difficult
lipu wawa :p
mi wile kama e ni → lipu li tan toki ante la ken pana li lon lipu pi toki [pona] li lon ala lipu mama la mi toki a e ni (la jan li sona e ni → ona li kepeken la esun li utala e ona)
nasin ni li pona ala pona tawa sinjale
license: CC-BY-4.0 AND LicenseRef-AllRightsReserved
mi pana e poki tawa ilo alasa [Tantivy] la ona li alasa pona kepeken tenpo lili a :3
,,anu pona ala · mi alasa e "Kita" taso la mi lukin e lipu pi [kalama sin] · mi alasa e "jan Kita" la ala
-aaa ona li pana e ijo sewi luka luka taso
Where are the unclaimed sources?
the list of them, anyway
yea ofc,, im not that crazy
do you think 268435456 unique commits is permanent (as in permalink) enough?
please elaborate
whats the context for this
i need permanent links to files for something and seven characters (default for git rev-parse --short) on commit hash allows for that many unique commits before confilcting
i just hope you arn't committing 1000 files, 1 file per commit, lol
but otherwise like sure???
i mean i'd still have space for 268,436x that :D
If you pass
--abbrev-committo thegit logcommand, the output will use shorter values but keep them unique; it defaults to using seven characters but makes them longer if necessary to keep the SHA-1 unambiguous
https://git-scm.com/book/en/v2/Git-Tools-Revision-Selection
taso tenpo
kulijo mute
@carmine quarry are titles optional or required?
required
Also, i think we should at least start throwing warnings on missing descriptions
read poki/utils/validate-schemas.ts for details
we really shouldnt; very few works actually have descriptions, and its unlikely we'll be showing them in the frontend
its mostly like what, some utala.pona.la submissions?
i don't know,, theys would be nice in search results etc.
short of reading the entire library, it seems unlikely we'll ever describe every work
Should we make dates
date:
iso-date:
precision:
or
date:
precision:
or something else
the latter
more flat is more good
-# lipu luka luka ale "taso" li lon la ken,, :p
<@&768320337896275998>_asi o · mi o ni anu ante
aa ike ilo !
@carmine quarry
mu kala suli pi sike nanpa MAML 😔
a a a
pona ala
jan li ante e lipu la ona li ken ala pana e ken sama nasin CC-BY-4.0
ona li pakala pali
ona li pana e pali ona anu seme
a a a mi kala
mu kala a
kala ale li lon
mi kala li mi lon 
a. mi kala
aaa. ken??? msa
kala li kama lon
mi wile pana e sona ale · nasin mi li pana e sona ni → "jan wan li pana e ken · taso jan ante li pana ala"
kala
tenpo seme la mi ken kama pana e lipu sin
kala ala
@nova gale will you manage to make schema edits within the next week
ike la mi open e pali :p
li pini e pali :p :p
(pi sitelen wan taso)
A library / monolingual corpus of Toki Pona texts. - Create utmws-13-a-o-lukin-e-lon-ni.md · kulupu-lapo/poki@949c450
if i were to guess, what you assumed would be a valid way to write a multiline string
was not
taso ! mi lukin lon lipu ! https://yaml-multiline.info/
Find the right syntax for your YAML multiline strings.
a mi lukin e ike
taso seme la ilo a li moli
🐟
the thread flooded with kala
yea likely
i also wanted to ask if we could raise an error if
- we have unquoted strings
- we have quoted keys
- we have some non-standard amount of whitespace in lists
oh and also if there is non-schema keys
its a balancing act between "whats annoying for the data gathering person?" and "whats annoying for the frontend person?"
mi la forcing people to make changes to a pr because they failed to follow a quote style is unreasonable
but for adding non schema keys - would be reasonable
ilo li ken kama e ante ni anu seme
Continuous Integration [1] goes far beyond testing and building code nowadays. Test code coverage and other reports might get created, documentation might get updated and metrics and statistics might get refreshed. This article shows how these results can be pushed back into the repository using GitHub Actions [3].
a poka la https://stackoverflow.com/a/46063271
yes, but doing so goes back to the question of automatically editing the schema without fucking up the text or the comments - and that is something soko ni is looking into
poki Lapo li kama poki kon tawa Lapo
the "not fucking up anything" would be made way easier with the restrictions i mentioned, as we could even roll our own parser
im 👍ing you trying that, we just gotta not make it too annoying for ppl
then how about a new schema completely? we could loan more TOML than YAML for example
toml could actually be better suited for lapo now that i think about it
one option could also be just making our own parser for a yaml subset that woud a) be so simple you cant mess it up b) be compatible with yaml so it would work correct in md readers c) be easy to parse
the toml idea is, imo, bad. yaml frontmatter being standard is really nice for the end user, be that open Lapo, or potentially maybe utala.pona.la or liputenpo.org
mi ken alasa e pakala lon lipu ale lon wile :p
the own parser is like... ngehhh??? if you succeed at it cool fine. but writing lower level stuff is a good way to derail a project
ni la sina pona e ike seme · sina pona ala e ike ni mi
title: UTMWS 13: a, o lukin e lon ni!
always use quotes and youre good, no?
ike
yes
-# taso mi wile alaaa
yaml parsers are greedy with keys
o kepeken nasin INI tbh
INI == TOML
opinion: 90% of these errors would be fixed by making contributors use a good template file
which is like, much less work
taso ini li weka e wile pi poki toki
the dream is having an html form + server hooked to github api to open PRs
-a taso ijo mute li ken ala lon nimi wan la anpa
mi toki e nasin ante lon tenpo pini
taso lon la mi wile ala e ante :p
uuu anu · lipu pi ilo jan taso · ona la jan li ken pana e ijo li ken kama jo e yaml
nena li ken tawa ijo pi tenpo mute,,
poka la · ni li wile ala wile lon poki pi tan lipu https://genius.com/Jan-usawi-7-likujo-lyrics
ken
pilin mi la ilo [Claude] li ken pali pona e ona :p
,,ona li ken ala
ona li pakala sama mi
ni li wile mi · taso ona li tawa nasin JSON li tawa ala nasin YAML · https://github.com/json-editor/json-editor
that wouldn't be that hard with something like https://www.wysimark.com/ and htmx
the hard part is having a host, imo
poki pi ilo Linku:
poki pi ma Seka:
kulupu pi lipu wile :3
https://www.youtube.com/playlist?list=PLuYLhuXt4HrRFeV7wSFL-1LRMtj5353pq
https://www.youtube.com/playlist?list=PLuYLhuXt4HrRIt7XTnyygXomDqlW3xPFK
all songs included in my toki pona medley kijetesumikyoku (except NASA, which is not on youtube. you can find it here: https://djemynai.bandcamp.com/track/na...
yea i got no idea what this is and why
plz no json 😭
also that's probably the opposite of what we want
about this,, i wrote the following parser
def get_frontmatter(data: str) -> dict[str, list[str] | str]:
f = re.search(r"^---\n(.*)\n---(?:\n|$)", data, re.S)
if not isinstance(f, re.Match):
raise ParseError("Couldn't find frontmatter!")
frontmatter = f.group()
out: dict[str, list[str] | str] = {}
keys = re.finditer(r"^(.*?):(?: \"(.*)\"$|\n((?: - \".*\"\n?)+))", frontmatter, re.M)
for key in keys:
groups = key.groups()
if groups[0] and groups[1]:
out[str(groups[0])] = str(groups[1])
elif groups[0] and groups[2]:
out[groups[0]] = [str(item) for item in re.finditer(r"^ - \"(.*)\"$", groups[2], re.M)]
else:
raise ParseError("Invalid syntax!")
return out
mainly in regexp in fifteen minutes. It can parse the YAML subset i described (only strings and lists of strings, list is described using strictly by
- Newline
- Space
- Dash
- Another space
and all strings are quoted
i can also remove the all strings are quoted thing quite easily
it's name is snaml or "still not a markup language" and it is fully compatible with yaml
just popping in again to say mi olin e sina to everyone working on this<3 this is something that i want to exist, and I think that if I didn't already feel like I have 10 times the things I want to do than I have time to do
, I would be contributing
ala · kon wile mi li ni → o kepeken nasin [Markdown] lon poki pi nasin [YAML]
title: mu
authors:
- mu
- mu tu
li kama
# title
mu
# authors
- mu
- mu tu
ike mute >:3
ona li ken ala ken moku e toki pi jan taso
- plaintext/2019/11/toki-nasa.md
- plaintext/2021/04/toki-nasa.md
- plaintext/2021/05/toki-nasa.md
- plaintext/2021/08/toki-nasa.md
- plaintext/2021/09/toki-nasa.md
- plaintext/2021/10/toki-nasa.md
- plaintext/2021/12/toki-nasa.md
- plaintext/2022/02/toki-nasa.md
a a
aaa ona ale pi wan ala li tan lipu tenpo
mi pali e lipu kulupu tan ni tu lon poki mi jan-misali-medleys
kulupu pi soweli musi la · lipu 14 17 li lon · lipu 21 18 wile li lon ala · lipu 1 li ken wile ala
kulupu pi ante ike la · lipu 21 23 li lon · lipu 27 25 wile li lon ala · lipu 2 li wile ala
mi lukin e insa ilo la · nullable en nullish en optional li ante seme a
mi alasa ni · suli la ale li pona · taso · mi ken awen e toki pi jan taso · mi ken pana e poki sin lon ma wile mi lon pini taso ala · mi ken ala ni tu lon tenpo sama
taaaso · pilin mi la mi ken ni taso
if "\ntranslators: " in front:
front = front.replace("\ntranslators: ", "\nauthors: ").replace("\nauthors: ", "\noriginal-authors: ")
li ken alasa e pakala
pali mi (mi pana lon nasin CC0)
i can prolly do it tonight
ona li ken weka ala · taso ni la poki pi mama open li sin li kama lon pini pi poki sona
awen la ni li wile mi
@nova gale @vague coral if you two work independently, will this lead to duplicate work being done and frustration, or are you chill if one of you gets ahead?
¯_(ツ)_/¯ · mi la soko li suli nasa e pali pi lili lon · taso mi pana e nasin pi suli nasa ala la pali mi li pini · li kama kon
(kin mi o pali ante,,)
im giving you both a carte blanche to try stuff, open a PR, show the work
"a tan li ken wile lon tenpo ale anu seme · seme la tan li ken lon ala"
kepeken nanpa wan
tan li lon ala
(lon la tan li lon li weka · taso nimi linjuwi li wile e alasa)
kalama ni https://www.youtube.com/watch?v=MryDuToQMsg la mi o nimi ala nimi kalama pi jan Tom Lehrer · kalama li sama · taso toki li sama ala
Inspired by Tom Lehrer's Elements, this song consists of every official word in Toki Pona, plus a few more.
ni, nanpa, nasa, nasin, nena, nimi, noka, namako
sin, sike, sina, seme, sewi, seli, jaki, sijelo
en, esun, suno, supa, suwi, suli, linja, palisa
li, pali, lipu, jelo, lape, poki, kule, kalama
wan, tawa, waso, wile, walo, lawa, kiwen, ake...
ni la mi o seme a https://djemynai.bandcamp.com/track/nasa
okay i just got around to looking at this
youve figured out how to force null to be explicit and that was pretty much my only gripe with ruamel
feel free to merge the formatting commit, and do the schema swap next
a suwi
ijo o pona e ilo pi ante ijo · ijo li ken mi li ken soko
(nasin pi poki sona li ante la mi ken ante e ona lon ijo mi kin · ni li ike ala tawa mi)
I still don't understand what to do here.
explaining briefly:
we store texts in markdown files with yaml frontmatter
the job of data collection is to find the raw text, format nicely if needed, then find all the important metadata about it
who wrote it, whats the title, what's the source link, when it was published
then you send that on github, open a pr, i read through the files, if everythings good, we merge it into the poki
because this risks two different people doing the same work twice, we like to split responsibility by claiming particular kinds of works (lipu tenpo is done by one person, utala.pona.la by another, personal blogs by another...)
i can elaborate on any of these if you need more guidance
oh and as of this moment we're also doing technical stuff to make our metadata nicer, so its better to wait a bit for that to be finished
Does that mean all of this will be put on a website? Not just text in a text file.
the key part is that when texts are stored as markdown + yaml frontmatter, other apps can take them and display them nicely
we have one right now, https://kulupu-lapo.github.io/open
but something better might come along eventually
I wouldn't know how to claim and archive works.
claiming is just saying "im gonna work on everything from <place>"!
by archive i mean the very process i just described. putting text and metadata into a markdown file
I don't wanna work on everything from a large collection.
thats fair and we're done with several of those!
la o pali pi kulupu lili
lipu monsuta li ken pona https://lipumonsuta.neocities.org/
,
I tried to write this in Toki Pona, but I can't.
I don't feel like claiming anything right now.
ale li pona
If I get bored and stop archiving it, then it will remain claimed and not archived.
I wish to humbly ask the lipu monsuta crew to publish the source code, that would be nice
bad news! theres no crew anymore
wait what
project leader stepped down
ah
💥
Juwan!
mi wawa · o utala ala e mi · https://github.com/kulupu-lapo/poki/pull/35
judging by the validation errors youre targeting the new schema
mi toki a e ni
please open a pr with that cause i expect thatll be easier to review than only having your code to look at
soko li ni ala la mi lon suno kama
[ilo penpo o lukin ala]
-# but i still think the simplest way to do it is just string-level find-and-replace
a ante li pona tawa sina la pona
the only possible ike is that comments get weird, but so far i haven't seen that, or that null gets implicit, but you found a way to avoid that
we'll need this kind of schema upgrader into the future so its good that we found something that works
@vague coral shame
:3
should we require the date field on entries that have unknown dates
also, is it ok add some sensible labels for gh issues?
something like problem, source tracker, question
just scan all of these urls and you'll have quite a good understanding
yes
put our best guess, in date-specificity put the lowest specificity
best guess: 21st century
i have PreciseToDay, PreciseToMonth, PreciseToYear and PreciseToNone
on PreciseToNone it returns true to all comparisond
sure
we dont really need to be doing any comparisons. the frontend can just assume the date given is good
yea exactly i was just talking about the library i made for poki
so is it an error or not if we have PreciseToNone and no date
no date is not allowed
all files have dates and all files have them down to the day
otherwise the frontend will suck
the precision field just tells the frontend how much of that date is a lie
for (a) us editors to be aware, (b) for display purposes maybe
nasin ni li ike tawa mi lon open · taso mi kama wile pana e poki lon ilo alasa la ilo li wile e tenpo pi wawa ale la mi kama wile sama :p
ni a
@nova gale o · sina ala sina anu seme li pali ala pali anu seme e ni ala ni anu seme
toki la sina pali lukin
miii sona ala e nasin :anpa:
a mi kama sona !
a mi kama lukin e ike
https://github.com/kulupu-lapo/poki/blob/ffb8ad6a29da7dba05a73a96bab542de1a1e46e9/plaintext/2011/07/o-pana-wawa-e-kiwen.md
jan pi ante toki li lon · taso jan pi toki mama en nimi pi toki mama li lon ala · ni la sona ni li weka → ni li tan toki ante
@carmine quarry o · mi o seme · mi o pana e original-authors: [unknown] anu seme
mi pini aa
@heady vigil o · pali li ken open sin lon tenpo poka
-# o pana e lipu mi,,
ante la · nasin wan pi lipu kalama li wile ala wile · ma [Genius] la nasin li lon li toki e ni → "o nimi e kipisi toki lon poki [] lon ala poki () lon ala poki <> · toki li sike la o pana e sike ale ona o toki ala e «×2» anu seme" ← e ijo ante mute · mi wile ala wile e nasin sama
we can standardise ye. ive not given it much thought. the easiest way to do it is to give all songs to the same person to fill out
mi pali pi jan [Misali] la mi kepeken nasin pi mama musi ¯_(ツ)_/¯
mi ken alasa wan e nasin lon poki pi mi taso
-# taso open la mi wile wan e ijo ante,,
ilo li ni
authors: I
translators: U
↓
original-authors: I
authors: U
↓
authors: U
original-authors: I
I en U li sama la nasa ni pi lukin taso li kama
oh right i forgot how to read diffs. pakala pali. taso pakala pi pali sina ala
@vague coral
- can you also edit the validation schema in this pr, so we can merge with a ✅?
- would you like to do the
original:grouping as a separate pr?
- was there any month-level or year-level specificity?
- mi ken ken e sina
- ken !
- wan taso · taso jan li pana lon poki pi jan taso la ilo li moku e ni
ping me when you edit the schema
mi pali ante li ken ala · sina o ken
Allow edits and access to secrets by maintainers
anu seme · mi sona wawa ala
ike la sina awen pali lon poki sina la mi awen ken ala pana e ijo tawa poki sina
??
mi sona ala poki pi mi ala lon ilo [GitHub Desktop] · ken la mi o ilo ante
okay so for now lets merge what you got, ill add the updated schema, next time you do work in kulupu-lapo for simplicity
git clone [email protected]:kulupu-lapo/poki.git works wonders
in general
i use github desktop too i just clone manually
@vague coral you okay with me merging rn?
pon
@vague coral lets try to merge the medleys as well now. start by git pull; git rebase maining it
-a mi pakala lili lon wan li pana ala e mama · o awen lili
its a good time for us to decide whether we'd like to null every field and to require them at all times
sources: null
archives:
- 'https://www.youtube.com/watch?v=QlV2RM5AO-8'
this is incorrect imo
aaa lon · mi lukin lon open la mi sona ala e ni → jan pana li jan pali
[ilo penpo o lukin ala]
apologies if this ends up rude but are you certain you can't commit to the branch? everything seems to point to "yes you can, thanks to the checkbox"
awen la mi pana lon poki mama lon ala poki mi lon tenpo kama
yep, i just tried. ilo li nasa.
ale li pona
@vague coral would you be willing to add an explicit null for every field in the next pr
like, poki-wide
there are two different ways to handle optionality, by dropping the field or by setting the field to null
its kinda bad that we mix the two methods without a clear purpose
poki sin li kama la sina wile pana e null sin lon aaale anu seme
yes and we can do that by having a template
nnn lon · taso ni li ken pona · wikinanpa la poki "no value" li lon · kon ona li ante li ni → "mi lukin la ala li lon" ← li ni ala → "mi lukin ala"
taso · ona li ken ala ken ante tawa ilo moku ale
we've kind of already made explicit special cases for that, like when authors are "folk origin", or the date is "date-specificity: none"
a
ni la mi la poki pi insa ala o lon ala anu seme · ni li lili e pali pana e pali lukin e pali pi poki sin
mi sona wawa ala
soko li ken toko
a ni li weka tan sona mi https://github.com/kulupu-lapo/poki/blob/a4536252cfc2bb678d04f70997a363a23497fce3/plaintext/2023/08/akesi-apeja-li-mi.md
ni li tan jan Usawi · taso ni open li tan tan Lija · lipu o wan anu tu · @carmine quarry o
it should be listed under jan Lija, for the date jan Lija posted it, and jan Usawi shouldn't be listed at all, only a link to the performance
a pona
compare kulupu jan tenpo
kalama pi jan Usawi la toki li ante lili la ¯_(ツ)_/¯
ante seme li toki pona
toki lili li weka · toki "(sina toki e nimi akesi? tawa mi?)" li kama · sona mi la ni taso
i want to cite poki lapo for a paper i'm writing- should i list these authors?
sona
nasin seme pi kipisi linja poka li pona tawa sinjale
mu\\nmumu\nmumu \\nmu
mi kin li ni nanpa tu · ona li pona tawa lukin mi · taso ken la ona li ike tawa jan pali pi ilo mi ala
(lipu pi mi taso la mi nanpa ala · mu \n mu · :p · taso ni li ken ala lon poki anu seme)
in correct markdown, this will become obne line, yes
a good ilo should handle that gracefully
A renderer may also provide an option to render soft line breaks as hard line breaks.
https://spec.commonmark.org/0.31.2/#soft-line-breaks
:3
(taso mi sona e ni → jan kepeken li ken wile ala ante e ilo li ken ken ala ante e ona)
ilo pali en ilo lukin ala · ilo mi la kon li kama · pi walo lili taso
yea we are def keeping that off
https://github.com/kulupu-lapo/poki/issues/11
mi la utala pi musi mu o lon · ale ala · taso mute
if / when people transcribe it, sure
la mi ni :3
reminder:
we are currently on pause
for new files
until we finish fiddling with the schema
- would you like to do the original: grouping as a separate pr?
- ken !
this remains one of the tasks, as well as obligatorilynulling every field if you agree with my logic for doing so
mi sona ala e pali ante wile la mi ni nanpa wan · sina la poki original o lon seme
-a kala li kalape
kala li kalalape
choose for yourself idk
sona
the pali ante is this
とりぽちゅん 「telo sewi」 / tori-pochun "telo sewi (The Rain)"
日本語版はこちら! → https://youtu.be/atR0cDv3sNU
kalama musi uta:
jan Maki, jan Lika, jan Kalin, jan Sipuju
kalama musi:
jan Toliposun
kalama:
jan Kajo, jan Kana, jan Najo, jan Utalo, jan Ke, jan Ju, jan Lijota, jan Toliposun
toki:
jan Osinikomi, jan Toli...
Alright kala_asi, I'll remind you in 16 hours about:
pr
alright kala asi
how do we feel about author: null for anonymous or unknown author
for previous texts i remember doing author: (anonymous) or similar
i prefer explicit
i decided on author: (anonymous) to match e pali ante mi
fuck im mismanaging lapo again
@jagged burrow we have an open pr from kita that touches everything, so ill have to merge it before working with your new stuf
@jagged burrow please continue your work on the newly created vivi/more-stuff branch
@jagged burrow https://github.com/kulupu-lapo/poki/pull/39 you can keep track of your progress here
@vague coral if you rebase your medley pr on main, we can see if its mergeable
@vague coral im trying out your reschema rn and i can't get this to work:
if re.search(r"\n[a-z-]+:\n - ", y):
yaml.indent(mapping=2, sequence=4, offset=2)
else:
yaml.indent(mapping=2, sequence=2, offset=0)
else runs 100% of the time any idea why that wasn't the case for you
actually fuck it ill just format to one option
@jagged burrow i have rebased your work
you are good to keep working
n · o pana e lipu ni → ona o ni ala
@jagged burrow youve received slightly improved error messages!
poki Lapo: Toki Pona library
changed a hyphen to a colon because it was bothering me
should i still use the vivi/more-stuff branch and that pull request?
-# also the mishap is just as much on me for not reading this channel properly
keep working on that branch, and now you can also see if the validation fails before merging
could we please add a particle the name of this post ("poki Lapo: the Toki Pona library" or "poki Lapo: a Toki Pona library")
we can change it but i prefer short names for threads
instructions unclear, created a particle nimisin meaning 'to interact with the poki Lapo library'
sona mi la mi pona ale e ijo Misali
this is what I was thinking lol
@heady vigil o · sina pana e lipu kalama pi mute seme
mi wile pali ni li lipu e jan kalama https://github.com/kulupu-lapo/poki/issues/37
sona ala
I tagged them all tho
either music or song can't remember
that's cool :O
what is the research paper
@vague coral merged your jan Misali pr
wawa
mi sona ala e tomo ni.
plan for today: bumping lapo/open
open shares the same-ish codebase as wasona so i can port some improvements from there
open now has previous/next buttons working
we now display the metadata, although truth be told it looks very wacky
but it works, is the only thing that matters to me rn
ill start investigating pages for tags and stuff
a good start
i love component reuse
we now have a https://kulupu-lapo.github.io/open/random link
imo poki Lapo needs a new system of 'revisions', to capture when someone modifies a preexisting toki pona text. eg 2003/05/ma-tomo-pape vs 2005/07/ma-tomo-pape, and jan Usawi's version of 2020/02/akesi-apeja-li-mi. This'd also help for adding wiki pages.
I think it makes the most sense to do this similarly to the collections with a separate file tracking all the iterations, but there's some argument for indicating this info directly in the file, either with a new field (metadata overload!) or by coopting the original field (but with revisions of translations this gets confusing). & then encourage all consumers to provide a mechanism to switch between revisions of the same content (for user-facing front-ends)
i think its a more specific instance of wanting links between articles
which we should look into
it's all versioned with git so that already exists
you can view all revisions with git log -p -- path/to/file
no like versions that are equally relevant for a user to see
in that case i think it's ridiculous to stay using git as we would implement our own versioning. also, someone who edits wikis, for example, can navigate a couple of smaller edits. it would also drive contributors to keep the tree clean
[ilo penpo o lukin ala]
methinks it's orthogonal to git versioning
in particular i don't think jan Usawi's cover of apeja li mi should be taken as the newest and thus canonical version of the original
even when both versions are by the same author there can be value to keeping both (e.g. my website has two translations of burger king foot lettuce because one of them is from 2 weeks after i started learning and the other is actually comprehensible)
oh, looks like i read venous's text wrong a a :D also wouldn't collections cover this?
or tags
collections now work
do they work exactly right? probably not
but for a first pass ill take it
big props to this person for having a slash in their name
we now have author pages
wawa mute
100% just going for pure functionality right now, we can pretty it up later
@heady vigil look you exist https://kulupu-lapo.github.io/open/a/jan_Juwan/
@nova gale
so i started the frontend mainly for markdown test purposes. but now that ive figured out how to make pages that are lists of works by user / tag / collection, its rapidly getting to the point of usability. what do you think of joining me to work towards improving this stack instead of building a fullstack app?
https://kulupu-lapo.github.io/open/collections/
https://kulupu-lapo.github.io/open/tags/
https://kulupu-lapo.github.io/open/authors/
its not written well but it works and loads immediately
ilo sina la mi kama lukin e pakala mi li pona e ona :3
weird bug: the root page loads a google font just fine; all other pages fail to do so. only in prod, dev works
@heady vigil @vague coral @jagged burrow @nova gale i invite you to check out progress on https://kulupu-lapo.github.io/open/, where we now have a short nav list and a dark mode
i have a theory as to whats happening, but its so incredibly stupid that i have to assume its not true
upd: i was right, its very stupid
something i think would be epic is a prev/next selector that allows for any of the lists. would be difficult to pull off with just astro but i gotta try
I think this is a good idea as it contains almost all behaviors we need (also they have shown to be quite hard to implement). We could also add some backend behaviors, as i do not think Astro has the capability to do everything we need/want (analyzing Git blames, possibly interfacing with Linku things etc.). Does Astro provide some interface with custom backends?
this, for example, would be very easy to implement in a simple backend
Does Astro provide some interface with custom backends?
astro allows components to be provided by other frameworks, e.g. react or svelte, which could in turn request something and then render it
ouch, that sounds like a bad time
that said ive actually already managed to tell individual pages what their prevnexts are, though making a good selection widget out of it is more difficult
i recommend you look around the library and think of what we're missing
of course search, but we probably shouldn't support every complicated metadata search under the sun; an average user probably only needs something simple
Sure! something like a thin HTMX client would make that easy. I also have most of the code working for the search and the current problems are mainly just
i want to make only one of them appear at a time, depending on how the user navigated to this page
could you set a cookie?
i was thinking more url queries
possible
its gonna take me a while to figure out but its progress nonetheless
anyway with the site confirmed official we need to discuss a permanent link
@nova gale @vague coral @heady vigil @jagged burrow
select all options you find acceptable:
🐿️ kulupu-lapo.github.io
🍍 lapo.pona.la
🍄 lipu.pona.la
🥷 lipu-ale.pona.la
(i can ninja edit new options if there are any in the next 10 minutes)
tenpo la ilo li wawa ni kepeken pali sina ala · (ilo [Tantivy] li ni · taso ona li tawa ilo mama li tawa ala ilo lukin · mi alasa)
I think it's funny tho
also btw im entirely open to renaming to avoid RAS
whatever yall like
I've grown attached to the name Lapo 🥺
-# poki lape · lape.pona.la · kalama suwi tawa lape pona
oh about the search thing: i am open to host a meilisearch instance as an api and loading poki there
ai..............
integrating it into lapo would be like one component with couple dozen lines of vanilla js
oh they also have a vanilla js lib
import { MeiliSearch } from "meilisearch";
const client = new MeiliSearch({
host: "http://127.0.0.1:7700",
apiKey: "masterKey",
});
also if the search goes down it just fails cleanly and doesn't bring the whole frontend down
oh wait it looks like they also have an itegration for external LLMS to generate those little summary thingies
ilo ni li pona tawa mi li ken pona tawa sina · taso ona li wile e pali namako · https://github.com/quickwit-oss/tantivy
Tantivy is a full-text search engine library inspired by Apache Lucene and written in Rust - quickwit-oss/tantivy
lon, ilo [MeiliSearch] li open e nasin ilo tawa linluwi kepeken namako ala
kin la, ken mi pi toki [Rust] li pona mute ala :D
what im trying to do with prev/next right now is almost certainly achievable in raw astro or svelte but i suck so im going with react for now
oh also just to set expectations straight: im not tackling search any time soon; if yall want it you're gonn ahave to try on your own
only one im familiar with right now
someone can look at my component later and rewrite it in svelte or whatever else
oh shid
it won't make sense to use because our library bumps will always be more than one new post
unless we automate bumping the library which i guess we could
:DD
we could dos (notice no d in front) by just curling that page on repeat
omg
nice
im sure microsoft will be extremely saddened by our use of their resources
too bad they have so many to spare
idk i think they have some kind of bandwidth limit on theese?
wooooo i can finally read lapo from Thunderbird
hmmmm
my biggest annoyance with react in astro was postponed loading and right now i just figured out how not to postpone loading
so uh my bad lmao
do you have the current version up on some branch
yes the main branch :p
upd: static astro is too static to know litearlly anything about search params
i could still get them in a vanilla js script but at that point... yea
bump this for @jagged burrow
oh im stupid, i can code client side scripts in ts
that makes things much less annoying
lol
mi pi kulupu pali li sona ala e wile pona. mi pana kin e alasa sona ni tawa kulupu suli
toki a
pali sina li pona mute a
mi wile e ni
sina poki e toki suli mi lon lipu sina.
nimi ona li 'pona len'
tan jan Asilu
mi alasa e pakala ni → nasin Markdown li wan e linja tu poka · pakala ni li lon lipu tu ale mute mute mute luka luka tu :anpa:
mi kepeken ilo ni
from mistune import create_markdown
from pathlib import Path
md = create_markdown(renderer=None)
def recursive_check(obj):
if isinstance(obj, dict):
if obj.get("type", None) == "softbreak":
return True
else:
return any(recursive_check(x) for x in obj.values())
elif isinstance(obj, list):
return any(recursive_check(x) for x in obj)
else:
return False
for i in Path("plaintext").glob("**/*.md"):
text = open(i).read().split("\n---", 1)[1]
if recursive_check(md(text)):
print(i)
(@carmine quarry) I think we should be looking into explicitly choosing a single client side JS lib early on to reduce package size, make the codebase not spiral out of hand and enabling us to start to build stuff semantically and methodically correct to that lib. my proposition is Alpine JS, because its tiny, has all the features i can think of open needing and has a quite relaxed learning curve.
(i raised this because i started playing around with the search component and related services i dedcribed earlier and want to make it right from the start.)
you are allowed to try it out and make prs using it, but im not closing the door on other people trying something different
so poki Lapo, much like sona Linku, explicitly permits the possibility of competing frontends. Linku was first delivered with vanilla js because that was what i knew at the time. Then smarter people came along and rewrote it in Svelte. We currently have a frontend in Astro (which currently doesnt, but could, use components in other frameworks) because thats what i now know and was able to get working well enough. Other people might know ways to improve upon it that might require throwing the whole foundation away, and I will accept if it comes to that
terrible news from the poll, more people like lapo.pona.la but more people i trust personally like lipu.pona.la
lipu.pona.la la nimi "Lapo" li nasa lili
nasa li seme
nasa li ni: nimi lipu li Lapo taso nimi lipu ilo li lipu.pona
n ni anu seme: mi lukin e nasin "lipu.pona.la" la mi kama pilin e ni: tan seme la nimi lapo li lon
nimi lipu ilo en nimi lipu lon li ante
sona
nimi Lapo li ken nimi pi kulupu taso li ken weka tan lipu
lipu pi sona pona la nimi li "sona pona"
aa sona
i am like 70% of the way towards adding a kulupu-lapo.github.io/add page
seme?
wawa
this will get sent to a hono backend on cloudflare pages, which opens a pr
which mostly works right now
adding github oauth could work too
i think the main problems right now are
- making fields exceedingly clear to users,
- testing requests against a schema,
- rate limiting valid requests
which in fairness is extremely normal stuff to do but it too takes effort
limiting submissions to only github users? thatd go against the point of having a public add page imo
democracy failed, lipu reigns
for this, may I ask if we could have a little tag/note/whatever to indicate that something was added by /add
a note where
inside the metadata
why
it goes through the pr review process and afterwards it counts just as much as anything else you've done
ooh that does make me wonder
should it be one branch accumulating submissions
or a bunch of separate prs
i currently have it be a bunch of separate prs but hmmm
@heady vigil @vague coral @jagged burrow i invite you to test out https://kulupu-lapo.github.io/open/add/
poki pi lipu mama li lon ala :<
the most glaring issue right now is the lack of rate limiting, so please don't spam click
i only included the most basic functionality for now
why is it yaml?
cause you overlook things when coding for too long
I will when I get home
mark mandatory fields
ui polish is needed, thats true; id love someone to give me a hand with that, as my ui thinking energy is exhausted
i can try
how into the open Lapo logo and outlook are we? i am thinking about redesigning it a bit for fitting in the search bar.
open stands for its status as a beta test; rebranding is very much in the cards
@heady vigil all fields must now be present
we changed that to make it easier for future editors to fill out the nulls
also you got two original fields
damn
I like the cursive for right now, it's fun :)
fix my fix pls
done
@nova gale kin
you should look into Astro Islands
in astro, they're a formalized way to deliver content from an arbitrary backend
theyre just components from various other intergrated frameworks right
i think we've discussed them already
that is one capability of them, but that's already a capability of astro without this system
astro islands let you designate
- that a segment of the page should be bundled for the client to execute (client:load)
- that a segment of the page will be bundled by the server on request (server:defer)
astro's default behavior is to bundle the page for the client to simply receive, not execute
so the former makes interactivity easier, and the latter lets you pull content from a server with arbitrary JS; it is client framework independent
How much do you like the open Lapo look?
3
6
2
I like it and want to advance with the same feel
i have set up the frontend to automatically get bumped any time we merge something in poki
so the user / maintainer interaction is reduced to one form and one pr merge
@heady vigil is tawa pona, jan pi lanpan kasi ready, do i merge
also moving forward do you want me to merge any valid files you submit
Hey is þere like an escape character like \ in Discord which would prevent þis from turning into like a code block?
I've updated þe OG text and I wanna update it in lapo too so I þought I should also deal wiþ þis issue too
What seperates þe chapters are a long line of ~~ bþw
huh, i never knew tildes can open and close code blocks, neat
ill replace them with a *** which creates a <hr /> if you don't mind
\ a
Þis also does happen (it's ~pona musi wawa~)
I think it could probs just be musi pona wawa~
Hopefull þat does uhhh noþing
I do gotta fuckin update þe text too
-# ilo li ~ li ~~ ala tan seme smh
It isn't supposed to be strookout
The GitHub Flavored Markdown only mentions that you can use double tilde, but you can use everything.
https://webapps.stackexchange.com/a/103691
anpaaa
i don't know, but Astro uses two different markdown renderers (one for most of the site, one to render markdown dynamically at /add) and both agree on this!
Could I not just update my yaml and þen it'd update everyþing?
After I resubmit it to github
if you refresh the page youll see what you've reported has been fixed
oh I mean md not yaml
pretty much how it works, ye
yes
,
the image was me testing what you said
ye
yee
@heady vigil https://kulupu-lapo.github.io/open/2020/05/ave-maria/ https://kulupu-lapo.github.io/open/2020/05/jan-meli-o/
i think you submitted both of these, what was your logic behind that
your problem is that you've started editing on yan-Mayeka/main
but your yan-Mayeka/main is way out of date
Oh
ideally, when you start making a pr, you create a branch off of poki/main
insert pokimane joke here (i don't know any pokimane jokes)
okay uh
I press þe button which sincs þe fork but idk if þat's right
idk what a branch is
okay imagine your undo/redo history as a chain of nodes, where each node is a state you can jump to by using undo/redo
now imagine instead of a chain, its a tree
edit history can branch out
each branch is a, well, branch, and can have a name
okay
I've updated by branch to be up-to-date
I'm pretty sure þat's correct based on it working before
wait is poki also pulled automatically?
it is now!
i thought submodules couldn't do that
it took some github actions and ssh shenanigans
only as many as there are merges to poki main
which is still a lot but like idk its fine
sure
@swift imp the pr checks your work and lets you know if something still needs changes
Yeah I'm trying to figure þat out rn
hopefully the messages are human-readable. lemme know if you encounter anything that's impossible to understand
this particular one, for example, is you misspelling plaintext
oh
plaint text :D
also one dx/contributor fix would be to show those errors all at once and in a nice message by some bot
@vestal herald awesome to see you try the form! any idea why sources starts with >- ? is this something you typed?
hm, no
i fixed it for this pr but im gonna have to see if that ever occurs again
since the file name gets generated from the title, putting the title first in the form might be a good idea

theyre carrying lipu tenpo and by extension a significant part of the library
opinions? (i also changed the colors around a bit to make them consistent and matching)
though this does not take into account work length; jan Kepe would be much higher for instance
oh and i also made the main element wider do that it doesn't wrap around weird
word count ranking when
not impossible
Good idea
compile time metadata injection anyone?
word count comes with a lot more assumption than just article count
how much of an article is by a given author; which parts are text and which are html/css/comments; do non toki pona words count; you get the point
or just measure the size of every file after stripping the metadata
omg it automagically works on mobile
never happened before
the CSS gods blessed me
(and i used grid :D)
oh true
Wait I should probably add like tags and CW to my stuff
mayamail is pretty fucked
@carmine quarry are we loading react and mdx for no reason? :D
I'll do þis when I upload sitelen Onkesele pi toki pona
and also shipping react to the client??
opinions to making the "all entries" section of the front page a scrollable box and moving the normal view to a separate path as i have done here? makes the landing page way cleaner and more approachable imo.
(ignore the astro dev menu)
and here is the dark mode
yeah why not
no clue(!)
presumably just two different passes that I took
may I also ask, why are licenses spelt with a hyphen again, like: CC-BY-SA 4.0
it's just a convention. the license's name is "CC-BY-SA" and it's version is 4.0
actually the correct format would be "CC BY-SA 4.0"
(CC: @carmine quarry we should probably fix the license names)
true
ok but
that follows NEITHER scheme
it doesn't follow the correct format CC BY-SA 4.0
nor does it follow the identifier CC-BY-SA-4.0
la o ni nanpa tu e ona >:3
I do want to share text of musi Tokusu / Cave Story game, but never kept track of the hundreds of lines of dialogue. (It scattered across a hundred files, one for each area of game. And safe to say it all out of order...)
-# In hindsight, I should have kept track of it from the start, may need to replay in it's entirety...
@carmine quarry How would branching be handled? CS has 3 (+1) endings, optional side quests, and other triggers that affect dialog...
Also what about writing systems? I have translations in Lasina, sitelen pona, Hiragana. And I intentionally mix writing systems within. I have one character speaking in different writing system from the rest of the game to reflect the fact that they speak in Katakana in the original Japanese.
yes
if theyre never used they should have no effect on compilation i would assume
mdx came by default and react i attempted to use for prevnext before rewriting it in astro
a normal html thing is that you can declare headings and then later jump to those headings with # links
ah, sound's like something i might be able to automate to some extent...
like those choose your own adventure books, turn to page 12 ect...
convert to latin plaintext. if that severely undermines the experience, you could do something like:
character Name (kepeken sitelen Katakana):
dialogue line
but if you choose to keep the writing systems ig sure whatever
the biggest reason for this rule is so that we done end up with every single work duplicated for latin and sp
Yea, that make sense. Use sitelen Lasina, and caps-lock for that one character (like in English translation)
@true crag looked at your koton submission
the image links point to files that of course dont exist on Lapo
you might want to edit that to a full url
also you would have seen the images behave wrong before submitting? 🤔
nah, i dont think i can upload those cause of copyright reasons. alt text is enough... There are only 4 characters
for sources can you think of any page online that would also have this text
so i can link to it
uh, well there is romhackingdotnet: https://www.romhacking.net/games/4385/
I would submit translation there too (once i finish that...)
so basically you're posting to Lapo without posting anywhere else
uhh yea, for now...
I should have rest of patch file done this month though
merged
thanks!
should be up on the site in like 3-5 minutes
pa kala
btw /2018/07/ma-sike-laso-kalama-mun is empty
@carmine quarry what is the original field for btw
is it for linking to other items in poki lapo or is it for any original thing it was based on
the latter
okie
somewhat related but for the curious as to why im not using the confucianism tag for my translation of dizigui its because although its extremely popular and cited as being a confucian text (given that it was based in part on a section of the analects of confucious), its mostly just a guy giving a brief summary of his thoughts and several qing scholars (scholars at the time of the qing dynasty, not modern scholars about the qing dynasty) critisied it for not being faithful to the original confucian ideals (especially with regard to its description of filial piety)
Should stories from an anthology be added together or separately?
so your choice is between one article several headings and one collection several articles
id make the choice primarily based on whether the stories are meant to be read together or make sense in isolation, + whether they were published on the same date or gradually
@carmine quarry (lmk if you want to be notified of this in the future lol) the pr is up :) not sure if i did the metadata right so lmk if it looks good
ping me any time
checking now
@bright leaf so for linebreaks within one paragraph markdown actually has a thing, its either \ at the end of the line or two spaces
ah i see
ive always just used br HAHA
I see that there are like a gazillion tags I can add. How many of those should I include? What details are necessary?
-# (like why is "moku li lon" a tag?)
When is making a new tag justified?
-# Is "detective fiction" or "mystery" needed for example
I don't think it's that deep. If you feel like adding a tag, do it imo.
I just submitted it, but I noticed that there wasn't an "original" box for me to fill even though it is a translation. Can I not add that?
making an /add page was difficult so i wanted to do the bare minimum. ideally this should be added
Aha I see
ona li tan https://utala.pona.la/
:3
(ken la ni li pona e ala)
hi you're free to make a pull request to improve the /add page. this seems entirely parallel to the work already done
lon
selo pi lipu https://utala.pona.la li sin
ma pona pi toki pona la utala musi li lon! o utala o lukin e musi pi tenpo pini!
wawa
@carmine quarry i'd like to modify the poki bumper script so that it squashes the commits together instead of adding a new one every time. is this ok?
so like, rewriting lapo/open's git history?
not precisely, but kind of. it just wouldn't show the bumps as different commits
i personally don't like the thought of modifying the history, but this kind i would be fine with
make sure it doesn't accidentally nom any other commit in between, but otherwise ye go for it
ok 👍 ill also add a note of this happening in the commit messages
someone with spoons for technical work can:
- familiarise themselves with https://github.com/kulupu-lapo/submit/
- fix a bug where
sourcesisn't being split up properly - implement
originaland other fields in/add, i already have most of the code for that but commented out
is the squashing only happening if the last commit message starts with "!!!This is an automated commit by the poki bumper workflow!!!" enough to ensure that the workflow doesn't fuck up the history
good enough for me?
ok
pushed the workflow change 👍 hopefully doesn't break anything 🤞
i'll also document the squashing in open's README
aaaand it broke
ive moved your WIP work to the ci-improvements branch, so that main isn't hanging around with a ❌ for too long
you can merge it back when you return to it
@heady vigil @vague coral we can have a tag for all transcriptions of live monologue and dialogue
sure
@heady vigil @vague coral @jagged burrow @nova gale i moved the frontend to a permanent url:
https://lipu.pona.la
branding is up in the air
lapo.pona.la had more votes but from less prominent community members; compare wikipedia talk pages
yay
wawa
which entries need to be marked with null again?
should really be put in the README schema
something else, does VS Studio have templates? it gave me a screen to use one as well as their stupid Copilot
I am not in the author list :(
(Context: I made one story written in Toki Pona and published it on this Discord)
It doesn't matter, just a goofy litte thing.
#pali-musi message
I wonder if anything can be added to this library or only known works.
Like well known.
anytging
Oh, interesting.
Should I add this story to it?

