#bookbrainz

G0maa I don't think so

IMO, if it's a community project then members should get to decide if something is "wrong", i.e. have a report button or something.

monkey could we perhaps take help of other orgs like bb who keep a record of books ?

You could potentially have semi-automated validation, meaning checking for the user and helping them identify data that is different in other websites/databases, but I do'nt see it working reliably for fully automated validation

Mainly for the same issue: what if I'm adding a book that isn't on amazon or openlibrary or goodreads?

Yup the disparity remains

Does Wikipedia also works on the same stature , like is the info their too left for users to check and validate for ?

Yes

hmm , alright last one '=D (sorry to keep nudging you with these ideas ) what about the publication , we could validate that ??

No worries, I'm here to answer questions :) What do you mean by publication?

I am talking about the publication company that published that book

Ah, the publisher. What would you want to validate?

if that publisher really exists or not , I am now trying to get that we could not totally automate the process there needs to be some level of human curation, so instead of having a perfect validation , we could aim for validation that removes those books from the db that are totally bogus , I mean like if a user spams non existing books , just to

tamper with the db

That we could tell by having these lines of simple checks , or is it even necessary at all?

I don't think it is. Bogus users will be reported, and then we can investigate their revisions and undo the damage

Alrightttt...

From my point of view there are many interesting things to do with the help of AI, but I think this type of validation is best left in the capable hands of a human community, and efforts should focus on making their job easier

Yep, I myself have been addicted to chatgpt , got all my college proposals ( for our college tech fest) getting corrected and formatted from it '=D

:') :')

I agree that AI can potentially help humans make decisions. I have doubts it can be used reliably in an automated context. I can get ChatGPT to make statements that are totally untrue. My sense of it is, a conversational AI may not make the best curation decisions about bibligraphic content. Other types of AI, trained on our and other data, could

potentially be used to detect anomalies or inconsistencies with other reliable data sets, which editors could then inspect and make decisions about correcting.

yup true

so on a side note, how long do you guys code per day, and after how many hours does that feeling of not being able to understand even a simple function begin to hit'=D

The amount of coding depends, I have to spend a fair amount of time reviewing PRs and generally managing the project, not necessarily coding.

But after 8 hours of full-on coding my brain starts to lose grasp on reality XD

In a typical day, I can go about 8 hours. But if I'm on a roll, that can extend to 12. In my younger days, even more. It's not constant.

Don't expect continuous, constant output.

That being said, these days 8 hours of solid coding would be spent on heavy complicated refactoring, so that puts the limit pretty low.

But it's somehow MORE tiring to move between reviewing 8 different PRs

Context switching takes a toll

A lot of time goes into thinking about how to implement (i.e. design), so as to be applicable to the case at hand and have flexibility or extensibility. Design goes a long way to avoid invasive refactoring later. But it's a balancing act too. You need to get the job done, while not incurring too much tech debt.

And, yeah, monkey, ding. Context swiching. Meetings. Working on different tasks. Ouch.

PR reviews! Talk about context switching. Not conducive to heads-down coding.

> You did mention some ORM work for example which would make for a quite consistent self-contained project.

To which of my three approaches was that referring, monkey? 1) refactoring to use Bookshelf's class syntax, 2) migrating to a TS ORM, 3) creating stand-alone type definitions as needed?

Well, I meant the general idea of improving the ORM, as in those three ideas

Although I think 2) would be ideal, considering the faint pulse of bookshelf

Not saying I want to propose one of these rather boring tasks as a project, but I'm happy to help getting this done outside of GSoC (once we've agreed which approach is the best for BB).

I think that somehow goes hand in hand with the "self-contained" aspect: No interactions with other parts of the website, just straight up code refactoring. This type of refactoring is up some people's alley, but yep, I also agree it's quite boring :D

refactor to use django to get more developers abroad :see-no-evil:

But as you say, the first part would be identifying which approach would strike the right one

Hah, G0maa you would be surprised at how many participants get interested in BookBrainz over other MetaBrainz projects *specifically* becuase it is JS based

Next I would try to convert one model by hand before I'll write a script that does the most repetitive parts for all other models :)

Maybe even a script that generates models directly from bookbrainz.sql and the human only has to copy in the additional parts from the old Bookshelf models.

while being on the same topic monkey when bb was created did you ever  think of using a no SQL db or was it decided before to keep the db consistent between various metabrainz projects

The latter. I wasn't there at the beginning of the project, but SQL just makes sense for these types of projects

Very structured data, a focus on reliability of transactions

hmmm

monkey: :O, I mean that is true, because I mainly searched the 172 orgs for a NodeJS backend focused project idea... the ones that were kind of up my ally were ~3, BB is one of them :D So to reply to an old message of yours I might do as you said and propose a project to BB anyway... depending on the circumstances.

G0maa did you look into the importing the db project , the one other than the admin panel listed on site

MrigeshThakur: I'll be taking a look at it next!

What's the purpose of referencing not only the same table but also the same column?

Now that is interesting

From bookbrainz.sql, btw

Just looking at the file now. No idea why this was done

Beats me. Maybe Leftmost can shed some light on this commit they authored?

monkey: Do you have more background that I can give on BB-717?

BB-717: Add Edition to Edition Group creates superfluous edition relationships https://tickets.metabrainz.org/browse/BB-717

I'd have to dig into the data model to try to describe what is superfluous about it. I suspect you know the issue.

Yes, I can answer on the ticket

Thanks.

[15:24] <monkey> You could potentially have semi-automated validation, meaning checking for the user and helping them identify data that is different in other websites/databases, but I do'nt see it working reliably for fully automated validation

sounds like userscript potential

[15:42] <monkey> From my point of view there are many interesting things to do with the help of AI, but I think this type of validation is best left in the capable hands of a human community, and efforts should focus on making their job easier

+100%

oh lol, G0maa should be interested in knowing how many languages BB went through before we settled on Js

for example HASKELL

Interesting, I didn't know that :O

:D

can tall to benochmore(lordsputnik) and Leftmost about it. also that guy who's name escapes me that als ostarted with "b"

s/tall/talk/

he did tht mb-user sruvey osme years ago (that I *still* have not see nthe results from *grumble*)

monkey regarding BB-718. as we only want to check if the user is logged into the critique brainz acc only when he hits the add review button , so can I add an event listener to the button call and call the getaccessToken function only then?

BB-718: CritiqueBrainz auth endpoint is called server-side https://tickets.metabrainz.org/browse/BB-718

we want to do it when she or they do it though tbh :P

as well*

I didn't put too much effort into BB yet, but I saw ORM being mentioned above, I can help in that using Sequelize.

MrigeshThakur : I'm worried that could make for a bad user experience if they wrote a review but can't post it straight away. Could even potentially lose their review text if we navigate away or something goes wrong, which I could see being frustraring

frustrating*

I've experimented with Objection models a bit more and am quite happy with the (untested) results so far: https://github.com/kellnerd/bookbrainz-data-js/... (new) vs https://github.com/kellnerd/bookbrainz-data-js/... (current)

I'll be back for feedback/questions (and with some additional explanations) after lunch ;)

*dinner

monkey i'll try to figure out a new approach tomorrow, till then good nights guys ,signing off....

!recall gsoc

I'm sorry, I don't remember "gsoc", are you sure I should know about it?

!remember gsoc=https://wiki.musicbrainz.org/Development/S...

I will remember "gsoc" for you rdswift.

!recall gsoc