#poobrains — Public Fediverse posts

Aaand, I have the basic logic for renderable objects implemented in #ooze.

I guess that's the first feature that's specific to #poobrains that's now ported over.

Still a bunch of stuff that flask did for me that I still need to implement tho – mostly cookies, sessions and form data.

Fixed some things in the new webframework and extended the test suite.

Might get to cookies tomorrow.

Then I can already start bringing over #poobrains features – starting with the theming system. Because why should I add two layers for templating and theming when theming logic is largely just choosing the right templates?

Then renderable objects, storage and after that sessions, which need cookies and storage…

I hope this is another change like when I decided to just drop pandas to write my own data analysis stuff which suddenly fixed all my serialization issues…

If I'm lucky, I can start porting over #poobrains features once I got cookie, session and form data handling as well as templating implemented.

Feels like everything else in poobrains is custom anyways – I mean I even wrote my own CLI abstraction before #flask had one and only ported over to the click integration later on…

This time it's just raw WSGI because I'm tired of putting in labor because of upstream changes in minor versions.

The current version of #poobrains uses a lot of #flask internals that aren't documented parts of the API (but weren't marked with '_' either) because of historical reasons.

@evilham The directives for example sound like a great idea, especially for the stuff I want to do in #poobrains.

But I still want the parser and its extensions to work on a tree that is officially part of the API. I'm not getting that with mistune.

Website up again, multiple problems – uwsgi going from python 3.7 to 3.8 necessitating creation of a new venv…

I also took the liberty to finally update my site to the latest version of #poobrains with the rewritten dataviz/analysis pipeline, but that necessitated manually writing a small database migration to introduce the locking mechanism without losing content…

Mostly untested, so there might still be some service degradation, but comments were broken before and that might be fixed, so…

@e8johan
I'll try to remember. This is mostly about #poobrains of which I'm currently the sole author, but this particular thing has been itching me for years now…

I wonder if I really need a CLA in that case, tho… I honestly have no idea why they ought to be needed for a multi-licensed project where all licenses share open access as base minimum – isn't that stuff only to have contributors forfeit their copyright? And how is that different from any GPL project?

@ArneBab @alcinnz @fsf @fsfe

One of the last things in the quickstart, also from 2017:

"There's a lot more to #poobrains and I'll get into these things at a later point, but currently most effort will be focused on working toward the first alpha version."

And it's still exactly the same, tho it looks like this year it might actually be true – the milestone for the first alpha just reached 60%. :)

Finished updating the quickstart and am now proofreading the thing.

And I still really like the first line I wrote some time in 2017:

"#poobrains is a webframework based on Flask, peewee and visions of the apocalypse."

I see past me was already good with slogans. :thounking:

Currently updating #poobrains quickstart document – stumbled onto this:

> PS: I'm actually considering writing a tool to manage CAs comfortably,
> since everything that exists sucks. Maybe in 2018…?

4 files changed, 184 insertions(+), 8 deletions(-)

2 files changed, 76 insertions(+), 25 deletions(-)

I think I'll actually get the documentation renderer in #poobrains done* today. :thounking:

* ᶠᵒʳ ˢᵒᵐᵉ ᵛᵃˡᵘᵉ ᵒᶠ "ᵈᵒⁿᵉ"

31 files changed, 481 insertions(+), 752 deletions(-)

Fuckin' hell…

PS: Commit number 800 for #poobrains! \o/

By the by #poobrains and #gulik now also have channels on https://libera.chat! :)

https://rnd.phryk.net/phryk-evil-mad-sciences-llc/poobrains/issues/15

Another issue for the first #poobrains alpha closed: Over 200 format string substitutions were replaced with f-strings. :)

@UberActivist

Also, you're kind of picking up on one of my motivations behind developing #poobrains: I want a decentralized leftist press that can cooperate over site borders. >:3

Geodata filtering with manually built polygons in #poobrains now works. \o/

Managed to finally move the match/filter logic into dtype abstractions.

Just realized that this also enables adding custom data types to the visualization/analysis system in #poobrains-based projects.

Can't wait to merge it back into the main branch – after I add x/y/lat/lon filtering to geometric data, add validation and clean up a bit.

Hopefully happening next week, finally bringing "Project trashpandas"[1] to fruition. 🦝

https://rnd.phryk.net/phryk-evil-mad-sciences-llc/poobrains/issues/51

Slowly but surely the picture around generalized match/join and filtering in #poobrains coalesces.

Will probably get the basic use case for numeric data down today. Might even get started on doing the same for geometric data. :3

Time for some hacking on #poobrains. :3

So, now I finally have a bit of headspace to continue hacking on #poobrains. :3

Currently implementing dtype-dependant join conditions and filters for #poobrains…

Can't wait to finally have the ability to filter on geodata.

Since the release of the alpha version has been pushed back so long, I posted a little intermezzo devlog for #poobrains:

https://phryk.net/article/poobrains-devlog-7-5/

Just put in my application to the current round of the @PrototypeFund – and with the topic this time being journalism, I hopefully have a better shot at it with #poobrains than in the last round. ^^;

My #Blueleaks seed ratio cracked 10 a while back – over 2.6T uploaded. :)

And with #poobrains finally making progress in terms of analysis again, I might finally get to actually do some preliminary analysis on it soonish – stuff like word frequency in relation to time and such.

Getting that far is going to be a red letter day for me…

Operation trashpandas is now underway.

I have begun ripping #pandas out of #poobrains' code.

Concerning futher #poobrains development, I know I haven't exactly shown a lot of activity this year, but I currently have a lot on my plate.

But I *have* pretty much settled on finally ripping out #pandas – it's already given me so goddamn much pain and all I'm even using it for for right now is simple comparisons.

And there's no guarantee it will ever become any less painful and not necessitate constant fixes for new versions.

I'll just roll my own simple analysis tooling with #numpy. >:[

@neil @rysiek Yeah, it's part of #poobrains – my nojs webframework

Results look like this: https://paste.xinu.at/rVq/

Pretty responsive without any js. :)

I ported my #poobrains-based task-management tool to python 3 and set up a local instance because I currently got so much going on project-wise that I'm in serious danger of just completely loosing track… Hopefully, this will help. :F

Well, so much for making the #poobrains alpha before the year is over, I guess. 🤷

I think now I'll put poobrains down for a bit and invest some time into learning 2D animation in blender and continuing my porting effort of the pretty easy privacy stack to #FreeBSD.

I just rediscovered #Gource [1] and used it to visualize #poobrains development.

Damn thing is a joy to watch.

https://mastodon.social/media/sMh61ett6SybO-FEurc

[1] https://github.com/acaudwell/Gource