Time to make a deal with the devil

@theabbie since you love downloading the entire devRant db and writing little gimmicks, I have an idea for you.

Avatars are envcoded as URLs. Each part of an avatar is separated by an underscore. Shirt, pants, desk, whatever.

Make a bot or script or website or what-fucking-ever to query users with the same avatar as you. This would be:

- Same EXACT avatar (desk, pets, etc)
- Same body parts as you (face, skin color, hair, etc)
- Same body parts and clothing (everything that shows in the mini avatar next to comments, plus pants and shoes, I guess)

The doppelganger finder. Honestly I think it would be neat.

Would be even cooler if you could filter by active users (last post/comment within past 3 months)

sounds like a large project, and also cool one, I accept the challenge.

@theabbie From what I know off the top of my head (if you didn't already), there's two avatar variants that devRant uses:

avatar and avatar_sm.

avatar_sm is used with the mini-circle things in comments, whereas avatar is used when you visit profile (or rant on web platform)

They're returned as distinct items in the API, but I don't actually know if they're different at all. Never had a reason to check, maybe you could report back on that.

But yeah the idea would be matching a few conditions. Matching based on avatar and avatar_sm would be the most important (and not just the data returned by the API, but what's actually visually included in the images we see on the app/ website)

Then you can use your imagination to come up with other cool comparisons. Stuff like people wearing my same shirt, people wearing my glasses, same hair style, etc.

@theabbie Finally, not telling you how to do your job or anything, but if I were to implement it, for each "category" to search by, I would have a list of indices that I could check against the encoded avatar string.

Split the encoded string by _ and remove any unnecessary items, and you're left with an ordered list of cosmetics

Then, you can write filters like

const SAME_CLOTHES = [4, 8, 10];

4, 8, and 10 would be the indices of shirt, pants, and shoes (Not the actual indices. I don't know the encoding).

That way you can filter in a very simple and declarative way. Pop them in an object:

{

"same-clothes": []

},

...and you can even return a queryable list of filters via Object.keys();

Anyways, my drunk ass will leave you to it now.

@AlgoRythm I don't see any problem in creating a USER table in SQL with fields like

ID
SHIRT INT
PANT INT
ETC.

Then just running queries like

SELECT ID FROM USERS WHERE SHIRT IN (1,2,3) AND PANT IN (1,2);

Time to start fetching user info slowly and not get IP banned.

@theabbie You'd certainly need to have a large portion of user data already downloaded in some format to avoid abuse.

That's why I suggest filters: even with smaller datasets, you may be able to find neat queries (who's wearing my shirt?)

Collect the data responsibly and store it in whatever way you feel most comfortable.

Is it just me, or is devRant slow today.... oh ok I see what's going on 😅

How will you keep it in sync all the time without being ip banned?

@Hazarth safest way is to grab recent rants and comments and use the active users to seed the data, but if you want to know every user, it's just a brute force on every ID till you grab them all.

@theabbie I’m backing you, go kick some ass! You got this!

@Hazarth it's not going to be real-time, maybe it will randomly update 100 users everyday, it will collect all data in advance.

You could at least subscribe to offset the increased costs this will incur.

@Root I've been subbed for like three years. Since this was my idea, I'll cover the costs.

@AlgoRythm I meant to tag @theabbie, but alright.

@Root just checked, it's more like 4

4*12*3=about $150 which should be more than enough for a slow download of the public data on this project plus a few more

@AlgoRythm Definitely enough.

Though: google, in their infinite greed, takes 30% iirc. I’ve been asking for a more direct means of donation for years, and have been denied every time.

@Root that's part of operational costs ¯\_(ツ)_/¯ as a consumer I've paid more than enough to use a free platform. I don't like botting and whatever, but as a cool gadget, I think this is fine as an experiment. Especially since not much has changed on devRant for a few years...

what costs are we talking about? Last time I fetched all profiles, it was like free.

if you make a deal with the devil you'll be sitting in a starbucks having one of the former victims of the people you made a deal with intimating hes going to slice you up with a knife.

don''t. lol. i'm busy as it is.

course the way they work the person in question probably went off to do something worse, immediately after playing good and pretending to wait on camera :P fucker :) HA HA HA they never behave. they just pretend to. more charges.

would be nice if i didn't have my mind polluted by the strangeness of a parent bringing their 5 year old or so kid into a starbucks to order highly caffeinated drinks.

@AvatarOfKaine Oh great, another loon

@AlgoRythm *slaps algo with a handy trout* callback to irc !

@AlgoRythm btw we all did this before :P

90+% of matches will be green dots.
Called it.

@c3r38r170 people without avatars don't have an encoded string. There's another field returned by the API just called bg that has a hex of the color they choose

Unfortunately green dots won't even make the cut :(

@AlgoRythm Yeah I imagined. Then I don't think there will be that many matches. Probably around (√users)/log(users).

@c3r38r170 I think many people will have exact matches, but not everyone.

You and me match based on body parts but not clothing or accessories

@theabbie I think they meant for serving the script and all about running it periodically. But either you can do it your own way moreover absolutely free

@Eklavya @theabbie me and @root were more talking about the bandwidth and resource toll it would cost devRant itself, which is why she suggested subscribing to ++

@AlgoRythm you've subscribed and it's your plan, so, should be fine, I will subscribe when I start earning.

@AlgoRythm Wdym? I know subscribing to the ++ plan will add on some more features. But why? If the fetching and reading are alright through API, does it provide ease in handling the database too?

Maybe make this a Collab?