Ah, you have to name the assets correctly
#tooldev-general
mortal bone
obtuse citrus
Actually it might not work with poeskilltree, I've copied the assests from old tree + add any new entries
but I'd need to generate the asset files myself if I want to do it properly
obtuse citrus
@mortal bone hm yeah now I fixed the assets it does work
obtuse citrus
yeah probably, I t hought either tool would display a placeholder but instead they just throw the errors
mortal bone
Yeah, I don't think either of them can get to the drawing stage without assets
so no placeholders
velvet fog
a black placeholder is better
compact isle
we'll be releasing beta tree data when it launches
obtuse citrus
Good to know novynn
mortal bone
@timid hemlock prepar yourself lol
compact isle
oh geez now I'm famous on reddit
mortal bone
Oh, shit. I need to do a build test...hopefully visual studio 2017 didn't mess with anything
obtuse citrus
heh well people have been asking for the beta tree pretty often, but I'd asked if you mind this being posted on reddit at the very least
mortal bone
if anything, I can remove it
compact isle
nah it's fine
I've been informed by Rory that the tree won't have all the balance changes that they want at release (no time), so expect to have it change again pretty soon
just as a warning ๐
compact isle
two Rory's high-fiving?! I don't know if the universe can take that
mortal bone
maybe an elbow touch?
mortal bone
Time Elapsed 00:03:40.05 man this build takes forever
obtuse citrus
I take it you have never used gentoo
mortal bone
not gentoo specifically, but I have built the linux kernel
had a class where we were adding modules
with 16 threads it was a 8min build time
torpid mesa
now imagine every single useful piece of software taking that long
mortal bone
ffs
obtuse citrus
Now imagine doing that 10 years ago with a single core cpu ๐
simple ravine
@mortal bone Build time including unit/integration tests or just pure build time?
oh, linux. got it
mortal bone
@simple ravine yes, build time for all of that plus packing
That is for the planner not Linux
simple ravine
ouch, that is some serious build time
mortal bone
Yeah, checking release number, building, tests (take some time), zipping, and installer takes a while
simple ravine
integration checks with external things, or just local unit tests?
chrome topaz
when are we going to have beta trading api? @compact isle
desert needle
I thought I saw a comment saying it might be a couple days into the beta, lemme see if I can find that.
Yes, we're working to get it up. It may take a few days after the Beta starts. Then community trade tools and sites can consume that API if they want to display items for trade in the Beta. ```not sure helpful but at least it sounds like it won't be right away
mortal bone
@simple ravine I also forget the downloading of game assets (2d item artwork). The tests are just local
simple ravine
Right now, I am working on domain logic around items, mods and stats that will be appliccable to use in different scenarios, i.e. Stash indexing, augmentive tools (ctrl+c popups etc) and other "macros", crafting simulation etc. Not ready for public collaboration right now, but I plan on open sourcing at least the central domain logic parts of it, and then use the code to develop other products, such as a poe.trade alternative (no offense to poetrade et al), and other tools.
I am contemplating taking the opportunity to learn more about functional programming and perhaps move some of this logic over from c# (imperative / oop) to f# (functional).
torpid mesa
what will this domain logic include? code that understands item affixes and stat scaling etc?
simple ravine
yeah
torpid mesa
cewlz
simple ravine
my idea is also to do what Openarl and Emmit does with their offline planners, and create a library for that as well, down the road
But that's a bit bigger project, however if people are interested to collaborate around this, that'd be fun. Using the word "standardized" might be to take it a bit too far, but it would benefit the community to help out with this.
Unfortunately we do however use different languages and toolsets, so I am contemplating if this could be expressed in a pseudo-language, and interpreted and still be efficient in calculations, which is important in some cases (simulations, stash ingestion)
The problem is I don't have a very strong poe experience, but very strong development background, so some collaboration and I think we could do something interesting in the future.
torpid mesa
Doing it in a .net language would give you the most direct usage. Then c++ for easy calling from most language since there are usually bindings
simple ravine
Unfortunately never had the opportunity to dive into C++ logic accessing .NET libraries, but I'm sure there are ways. Not sure how efficient those bindings are though? Any insights?
For Node, there are bindings, and I would guess there are python and perhaps even lua bindings, if that would be of any interest to people using those languages. Java has 'JNI' etc
wind garden
If you're calling . net from c++ you're gonna have a bad time
Doing the inverse is much easier
Assuming unmanaged c++
mortal bone
c++ is a good ground layer tbh
a lot of people use c++ to be the background work horse (see the new league of legends client)
simple ravine
To be fair, I think that C++ is not warranted (for me), considering the type of functionality that is being provided. Developing C++ vs .NET (unmanaged vs managed) is for most people quite a big difference in productivity, and I think that .NET/C# is a good common ground.
For an instance, this could potentially be an option: http://www.xinterop.com/index.php/unmanaged-exports/
wind garden
Been using c# for years, it's fantastic
Both for ease of use and performance
And with recent changes to . net standard you get portability too
velvet fog
anyone downloaded the beta client and updated?
mortal bone
currently connecting
compact isle
treelink time?
mortal bone
sure
gray osprey
@mortal bone TAKE MY ENERGY เผผ ใค โ_โ เผฝใค
mortal bone
lol your new naming scheme is a pain in my ass
compact isle
;D
unique night
we dialup now
obtuse citrus
use steam if you can lol
elder hearth
I'm thinking about linking my account on steam or the first time
obtuse citrus
yea
ripe fractal
Steam finished in 5 min for me
simple ravine
yeah, I normally get 54Mbytes/s on Steam, but opted for standalone for PoE
and these downloads normally are quite fast for PoE standalone, but I suppose the CDN POPs havent saturated yet
I want a Kitava Supporter dedicated CDN endpoint ๐
simple ravine
All support is good, don't feel bad.
obtuse citrus
also ziz probably posted everything interesting already hehe
unique night
u can do monthly plan
elder hearth
damn speed jez
obtuse citrus
time to find out if I need to fix anything for PyPoE
mortal bone
allocating space lol
obtuse citrus
yea
elder hearth
i will download via steam...
obtuse citrus
Dont necessarily have to you can just download steam client and copy the ggpk I guess
simple ravine
true
elder hearth
do i need the code to download?
obtuse citrus
On steam maybe I dont know I automatically got my key
compact isle
no not to download
elder hearth
It is downloading regular version first then
My purchase takes around 6~9h to finish
๐ฆ
velvet fog
anyone done the download?
velvet fog
steam block my region
elder hearth
๐จ๐ณ ?
velvet fog
nope, taiwan
simple ravine
looks like I just had to threaten the standalone client with a steam install and it picked up speed
elder hearth
here too
@obtuse citrus do you run some tool to compare previous/current ggpk? (at least files included/excluded)
obtuse citrus
no
I run tests for the .dat files
I need to update 14 of them so it will be a bit
I renember writing some basic diff tool based on the files added/changed but it's not in PyPoE
elder hearth
so GGG added logout macro by pressing "M"?
velvet fog
[PoEDB] Closed Beta site is up
http://cb.poedb.tw/us/
unrevealed gems:
- Ignite Proliferation Support http://cb.poedb.tw/us/gem.php?n=Ignite+Proliferation+Support
- Chance to Bleed Support
- Lesser Poison Support
- Deadly Ailments Support
- Tutorial Support
elder hearth
damn, that right click on tree feature is ๐ฅ
simple ravine
the right click in skill tree doesn't seem to do anything other than highlight the node
mortal bone
huh?
simple ravine
right clicking in the skill tree in PoE highlights the node
mortal bone
yeah?
simple ravine
but nothing else
simple ravine
hehe allright. i guess i don't see the point in it
elder hearth
So you don't get lost
mortal bone
Nodes > 'Tagged Nodes'
simple ravine
no, I mean in the actual Path of Exile skill tree in game
mortal bone
oh yeah why would it do anything else lol
simple ravine
perhaps educate the player on the nodes effeciciency, just like your app and pob
keen nest
Does the highlighting persist through logout?
simple ravine
will have to check that in a bit
keen nest
Would be useful to me if it did, can highlight the entire build right off the bat, and I don't have to faff about tabbing in and out every time I get a passive point.
obtuse citrus
When I try to figure out the foreignkeys and I get "Sacred Cow" randomly..
elder hearth
wtf
velvet fog
Deal with the Bandits & Passives: http://cb.poedb.tw/us/quest.php?c=bandit
obtuse citrus
chuan y our life regen value is not correct
velvet fog
I see, thanks
mortal bone
oh wtf
obtuse citrus
game uses "per minute" values for regen
and then converts them to seconds on display
mortal bone
oh god
velvet fog
got some problem in separating my databases
obtuse citrus
the funny thing about this any item that has a range for regen would have an absurd amount of rolls because of that ๐
mortal bone
is that only 20 skill points?
obtuse citrus
don't forget it's missing a9 and a10
mortal bone
ah true
obtuse citrus
not sure if some of the A8 quests are still in there
torpid mesa
20 passives total it looks like?
mortal bone
if anything we are missing one
velvet fog
torpid mesa
also the new bandit rewards 
obtuse citrus
From always oak to never oak
torpid mesa
aliras is quite viable
but attack speed and movement speed
also NO CHARGES?!?!
rips
obtuse citrus
I hope we get charges as reward in A9-10 or so
simple ravine
60% of Life Regenerated per second <- does this look correct?
obtuse citrus
Pushed an update for the specification
timid hemlock
โค
obtuse citrus
I think was able to do most of the stuff during alpha
also updated the docs on my website... I should really see if I can get gibhub to generate those ๐
simple ravine
Ah, I see. I don't have that kind of relationship with GGG.. yet, at least.
obtuse citrus
It was due to my wiki work
simple ravine
Well deserved ๐
timid hemlock
And for being an amazing human being ๐
obtuse citrus
I'm humbled, but I'd not call myself that haha ๐
simple ravine
It would be handy, and mutually beneficial if they'd gather some APIs for the things that are currently being reverse-engineered to power popular tools and services, augmenting the PoE experience.
Like, trading tools, skill tree planners etc
obtuse citrus
I agree I've suggested this to ggg about 1.5 years ago
wind garden
am i the only one who isn't able to select any nodes in the tree?
there's no pathing
left clicking does nothing
simple ravine
Was it a firm no, or rather "we're resource constrained" type of response?
obtuse citrus
No, the views on differ on this
wind garden
actually tooltips aren't working at all
obtuse citrus
some people at ggg don't really want to give out information (and dislike the data extraction, in particular when it spoils unreleased stuff or spreads wrong info) or want players to figure out, but some people also see the use
simple ravine
That's why I think official APIs would be a good thing
obtuse citrus
In particular I've heard they'll also want to remove all spawn chances from data, but that hasn't happened yet
simple ravine
I think few people are actually that interested in unreleased things, but rather want a convenient way of understanding the game stuff
spawn chances can be gathered with some tools anyway, if you really wanted to, the same way Uncandango gathered the chaos/alt chances a few days ago.
obtuse citrus
yeah but it's not viable at all
simple ravine
To a certain level of course
mortal bone
I get the emails lol
mortal bone
The fix is in the wiki
wind garden
ah
mortal bone
Maybe...
elder hearth
also Siveran (Vorici Calc Dev) is planning on rerun some tests and get some new info about socket colors
and update with jewels trick
wind garden
that did it
mortal bone
Ok
wind garden
oh weird
i think it's showing the 2.6.1 tree now
or 2.6 at least
the first time I launched there were nodes behind CI were gone
now they are back
copy and pasting the contents of the .zip release to the Program Files directory seems to have resolved it
really strange though
mortal bone
Idk
There is something wrong with the appdata in a few users
And I can't reproduce it
wind garden
in this instance it appears there was a difference between the exe installed files and the zip package
mortal bone
Not possible they are produced from the same source
wind garden
hmm maybe bad appdata caused the app to mess up its install directory on launch and reinstalling fixed that
mortal bone
probably
mortal bone
https://github.com/EmmittJ/PoESkillTree/releases/tag/2.7.1 fixes an issue people are having with tooltips and such not working correctly
obtuse citrus
geez all the spam on the reddit, most of all which is easily looked up in skill tree or poedb
elder hearth
? at your PM?
obtuse citrus
No I mean on the main reddit
timid hemlock
@obtuse citrus I can't get .dat export from the beta GGPK working; I keep getting row size mismatches, even though as far I can tell it should be using the beta specification
obtuse citrus
Is it the CLI exporter?
timid hemlock
Yes
timid hemlock
:/
obtuse citrus
gimme a sec
timid hemlock
I did set the version in the config, but it doesn't seem to do anything
obtuse citrus
I'll fix it
timid hemlock
Ok thanks
obtuse citrus
fixed
timid hemlock
And you say you're not amazing โค
Working beautifully so far
No giant walls of red text ๐
timid hemlock
Evidently ๐
obtuse citrus
39.1%
It's a bit of a pain since a lot depends on having an UI and the ggpk around
timid hemlock
Now I get to find out how badly my export scripts broke ๐
obtuse citrus
I don't think I've changed many headers, just "MonsterLevel" to "AreaLevel" in WorldAreas.dat
besides new/old stuff
timid hemlock
Yeah I didn't see anything too drastic in the spec diff
Well, except for whatever the hell it is they did to GrantedEffectsPerLevel
obtuse citrus
I believe it has to do with the effectiveness stuff
Something I actually wanted to test out but forgot about
timid hemlock
Yeah, well whatever it is I'm going to have to sort it out :/
obtuse citrus
the new values relate to the stats invidually more or less
I don't think GGG communitcates this on the skills themselves anyway
I've seen dmg on spells being adjusted 0.8 for minimum and 1.2 for max whatever that means
And StatData at the end had values depending whether it actually does anything... 1 was used as far as I can tell for no effect
"3" when there were values like above
That's what I vaguely renember from memory about this stuff when I first looked at it
timid hemlock
Hmm ok
Now I get to figure out why all the minions skills don't have multiple levels
Aw... I think they scale to level now, and god knows what the formula is
timid hemlock
No, I mean skills used by minions ๐
obtuse citrus
Oh
timid hemlock
They used to have multiple levels, with the minion using the highest possible level
Now nearly all of them have only 1 level
obtuse citrus
Did it really? usually minion skills go by the level of minion, which would only then use the highest level if it wasn't restricted to being lower IIRC
but if they scale of the minion base damage it's easier
then they dont need multiple levels I guess
timid hemlock
No, minion/monster skills use the level with the highest level requirement less than or equal to the monster/minion level
I check it with the Lightning Golem's wrath aura
obtuse citrus
that's what I meant
timid hemlock
Oh yeah
Oh, I compeltely misread you haha
But yeah
God only knows how they scale now
Actually, since the wrath aura scales this way, it should be testable
I wonder if those float values in GE/GEPL have anything to do with it
chrome topaz
@compact isle is there not going to be forum-based trading in beta? other than text-only at real forums (assuming you cannot link beta items there)
compact isle
yeah no forum trading for the beta
echo summit
Are the "null" account names coming from the public stash API event from events or something?
nvm found my answer
deft jolt
everything looks a lot better with a nice gui https://gyazo.com/1efe6f5e9bd8ddb5ec67bca2eb65a693
simple ravine
@deft jolt heh, what is that funky thing?
elder hearth
@simple ravine looks like an API sniper based on poe.ninja prices
simple ravine
yeah, funky console ui
deft jolt
aye, i wanted to learn python so i made one of those api consumer things. console printing is done with asciimatics
deft jolt
i've not got much idea how it actually works but from i can gather it creates a screen on the console and outputs there
it will be a GUI soon, ill add some interaction later
elder hearth
so those "COPY" are place holder?
deft jolt
yes
elder hearth
haha almost tricked me
desert needle
https://www.reddit.com/r/pathofexile/comments/6g30it/xyz_on_news_about_poetrade_in_beta_you_should_ask/din5uyo/?context=1 fyi if you missed it
sturdy sparrow
@timid hemlock are the minion changes done?
prisma prism
Hmm how do you guys datamine minion's skills?
mortal bone
pretty sure they released a spreadsheet
sturdy sparrow
that was just got skeletons zombies and srs
mortal bone
not sure they are datamine-able
prisma prism
I saw a lot of minions for Spectres in PoB
There are also a lot of information on the poedb.tw that seems to be taken from somewhere
mortal bone
DefaultMonsterStats I think
I haven't messed around with the files much, but what I have gathered from this chat that should be close to what you want
prisma prism
thanks!
obtuse citrus
MonsterVarieties.dat is a good place to start
it has links to the mods a monster has, its skills, its type etc
DefaultMonsterStats for the base stats of the specific level
mortal bone
Ah, thanks for the clarification
timid hemlock
@sturdy sparrow Not yet, but minions are next on my TODO list
There's a few wrinkles there though, since GGG seem to have changed the way monster skills scale
chrome topaz
wow did the poeapp guy get banned from reddit?
velvet fog
wind garden
what did he get banned for?
desert roost
We don't know. That's a ban from actual reddit admins, not PoE subreddit moderators.
velvet fog
@chrome topaz you don't receive a beta access? as I know, many streamers got alpha test
topaz whale
so, where are they releasing this beta api link whne thye do?
i feel like parsing json
velvet fog
They will setup a beta domain to provide beta stash api
I don't have beta access too, it's too short, no much thing can be tested now
elder hearth
RIP eventloop
topaz whale
oh hey, youve been the topic of discussion it looks like
mortal bone
@polar island yeah, who did you piss off lol
elder hearth
maybe he posted some anti-trump speech
prisma prism
Hey guys, i'm still discovering monster data for spectres. I found some nice info in the GrantedEffectPerLevel.dat, but values exsits only for lvl 1 and 68, i assume they are intepolated for any other level linearly?
simple ravine
@polar island what he heck is a shadowban?
The entire reddit community seems so so strange to be fair
chrome topaz
it's probably the self promotion rule
since that account was only posting links to their website
simple ravine
so if i would tell the community about some funky tool i just made for poe
i shouldn't tell reddit about it?
chrome topaz
of course you should. but if you post about it every week and don't post anything else you might get into trouble
simple ravine
ah, gotcha
i read a few rants about reddit audience hating the people who made reddit, and the thing that was a hype before reddit, that 'if you dare to add advertisements, we're just gonna hate on ya'll'
quite toxic/hostile environment, which is kind of strange mob behavior-state-of-mind that people enter when typing reddit.com in their browser
like, they have completely different expectations on reddit and its content than other corners of the interwebs
</rant>
torpid mesa
reddit started with a 'community' feel and went away from that
so it prob impacts those rants
it functioned a bit different too
it had a recommendation engine
simple ravine
reddit.com got acquired by someone right?
torpid mesa
yea
simple ravine
i'm sure they're regretting that acquisition heh
unless they're using it completely for datamining behavior
simple ravine
Has anyone contemplated or tried to use a (or multiple) machine learning algorithms to determine pricing based on properties and their values on items?
mortal bone
Tons of people
simple ravine
Do you know of anyone who had any success doing it?
mortal bone
As far as I know nothing has been too good
chrome topaz
[6:15 PM] stungravy: update on my attempt to predict jewel prices: the dummy estimator (just guesses the mean price) has mean squared error of 17 chaos. my best attempt with linear regression has 14.7 chaos mean error. so not great.
mortal bone
You can't tell when something is sold or removed, so it is hard to get a read
Something removed doesn't necessarily mean sold
simple ravine
Right, you can only know when it's not for sale anymore. Whether it was sold or just removed from the...
right
simple ravine
But with enough data and some anomaly detection, you could get a decent hint
It requires a lot more work though
mortal bone
Yeah, it would require a ton of work, but no one wants to put in the effort
Not much gain tbh
simple ravine
If we can assume that most of the items that are removed are due to it being sold, if a fraction of those are not sold, it's still good enough.
mortal bone
I don't know that it is
woeful sphinx
Isn't there some "group of data science devs" that is trying to do it currently with a more systematic approach?
simple ravine
I think they were newly graduates right, someone wrote a paper and then attempting this now. They have some poe-pricing webpage
It's not great though, heh
mortal bone
Ye
simple ravine
In my opinion, you need to factor in a few other things as well, including
- Popularity of modifier combinations (tri-res, flat-es+es% etc) by perhaps learning from ladder data, and so on, clustering/k-means etc
Then also factor in the difficulty or crafting this item (cost of crafting currency)
mortal bone
Ladder is good for meta builds
As well
I think you also need to take into account steamers and their builds
And popular build guides
simple ravine
Those will be mimicked and be noticed in the ladder.
The "mathil-effect" is a bit overestimated
It also depends on what the question you're looking to answer.
a) What is the "true" value of this item
b) What can I get for this item
they're both interesting
desert needle
One of the biggest problems as well is the sheer volume of noise in pricing.
You can pull a list of 100 rares with stats within a few percentage points of each other and they will be priced everywhere from 1c to 10ex
It even happens with unique items which is absolutely amazing since they are generally so easy to price these days.
desert needle
Right, my point is that the idea of a "normal distribution" does not effectively exist for rare items in POE.
I'm not talking about extreme edge cases, but rather than the datapoints are often evenly spread throughout an extreme range of values.
In theory actually knowing what was sold would mitigate that, which is why that always comes up.
simple ravine
I think it's doable.
However, you need to do some trade-offs and use a bit pragmatism
desert needle
For sure take a crack at it, either way you'll learn a lot. I wasn't saying not to do it, just pointing out some background from some of us who have spent significant time trying to do it and some of the hurdles that made various people back off.
simple ravine
Absolutely, much appreciated for sure
desert needle
the tl;dr is that when you look at the data available in depth, it is a lot more complex than many other similar datasets to the point where it almost seems random
simple ravine
I'm just providing my perpsective on it, because it's such noisy and untrusted data source
desert needle
I had some fair level of success in predicting in general which combination of mods would be more likely to be valuable.
That part of the problem is somewhat easy at least and is a good starting point, instead of going straight to a price.
simple ravine
I agree and I hinted at that approach a bit earlier
desert needle
A great example we saw awhile ago was that when Lightning Coil was the meta, rings with life and lightning resist were generally turned over faster than rings with life and other resists
simple ravine
Understanding the domain can be beneficial, but sometimes it's also a drawback - because you get too caught up into the problems
desert needle
yeah totally
simple ravine
One of the businesses I'm a stakeholder in, does around 2 million transactions a day from customers in many different countries with various demographic
I think I can take some lessons learned from the ongoing BI project there... perhaps
It's somewhat similar in the way that it's very random on the surface
desert needle
nice. I didn't have much luck with financial models, the closest I got to anything usable was using an RNN designed to find patterns with an outcome, such as one that processes data to predict likelihood of certain diseases from combinations of genes
simple ravine
I think I'll start with an azure ml model to quickly try some things out
desert needle
also do note that you will need to do a fair bit of pre-processing
simple ravine
it's also esay to stack algorithms next to eachother, and then predict on their values
oh for sure
desert needle
the API data is in a JSON format, yes, but it is NOT in a consumable format unfortunately
simple ravine
80% of data science work is cleaning ๐
desert needle
it always bugged me that they didn't turn it into a problem machine consumable format
simple ravine
yeah, I have been experimenting with the data quite a bit on and off for the past few months
I completely agree!
desert needle
Increases fire resistance by 50% in awkward situations is the stupidest thing ever
simple ravine
Why on earth using human readable modifiers?
desert needle
it was ported from a system that used the JSON data to directly display items for the overlays on the web site
that's why it's in a bunch of arrays and stuff
it was never intended to actually be parsed
they just kept that old display format
simple ravine
I mean no offense when I say this - but game developers rarely are good system architects
And it's not the first evidence I've seen supporting that statement
desert needle
and, to be fair, there seems to be a bit of a desire to actually not make that stuff too easy
simple ravine
There seems to be a bit of an internal divide in GGG around this
desert needle
in this case not cleaning it up and updating it is one of the ways they show a lack of support for a trade system they don't necessarily believe in
simple ravine
According to a couple of interview I've listened to
So yes, I think that shines through quite well
I also read some funny comments on how they rather trust their own opinion and gut feeling, rather than looking at the data.
That will be the end of GGG at one point, mark my words.
desert needle
it definitely makes it harder to support the game by developing third party tools
simple ravine
Some people being a little bit too proud. These people are not necessarily the same people that will take GGG to the 'next level'. With 'next level' I don't mean become this bloated undersirerable company that people frowns upon when thinking about, but rather making sure that it statys relevant and interesting to the audience.
Of course, they're not bad people in anyway...
But it's the same mentality as saying 'Why would we want technology A when we have technology B'
Blockbuster vs Netflix, Altavista vs Google
Oh, and that jewel problem reminds me of something, actually @desert needle...
The data used needs to be temporal. An item and its price a week ago isn't relevant anymore.
sorry, mistag, @chrome topaz
desert needle
yeah, though there tends to be some stability about two weeks into a league for a couple of months
barring meta changes
simple ravine
Jewels change in price as more and more people add them to their builds, which normally happens when they reach 80+
chrome topaz
?
simple ravine
Your quote on the guy attempting to price jewels a few minutes ago. It might be because he used too wide range of sample data, and not limiting it to the past N days of pricing data.
chrome topaz
alright, well I myself don't know much about any of that
simple ravine
fair enough
velvet fog
chris_wilson said: We're working on getting the trade API hooked up for beta. The tricky part is that there's no public website for beta, but one is needed to host the API. Should be up within a few days I hope (though aiming for today of course).
elder hearth
guys, i ran a crafting simulation for chaos spam vs alt-regal with Vaal Regalia, anyone knows any crafting method or item that should be interesting to run? (I can simulate any item)
some results with regalias: https://docs.google.com/spreadsheets/d/1WcEgdcwuH9Jw_786q3kGrKZN1Su_83NGcizSzzi3-NI/edit?usp=sharing
simple ravine
so what is your conclusion - what is the best method to gain a useful item in general?
elder hearth
for 2.6 regalias?
simple ravine
yeah, or in general
elder hearth
general is too complicated, because you can't chaos spam weapons (odds are TOO low, so alt+regal is always the best)
but simulating would be good too know an average how much would cost to craft one
simple ravine
indeed
elder hearth
i think i will try some weapon
this league i crafted some good DPS 2h sword and claw
simple ravine
so with regalias, what was the best method?
I'll finish my version soon as well, and we can compare methods ๐
elder hearth
yes, that would be nice
i think current best method for SC is Chaos Spam but you must know that you may not have an open suffix (to metacraft Prefixes cannot be changed)
so if you want good regalias, chaos spam. (remember, SC)
if you want mirror worth ones, alt+regal (so you are 100% sure to have an open suffix)
simple ravine
SC?
elder hearth
Softcore
simple ravine
oh, as in lower threshold on the ES value?
elder hearth
no, currency prices makes the difference, but you can read as the following:
Chaos Spam:
0,19 ex```
avg per try to hit good regalias
806 ES (52,81%)
806+ ES (47,19%)Alt-Regal:
Alts: 2573,86
Augs: 185,62
Regal: 83,68
Scour: 83,68
Exalteds: 1,71```
avg per try to hit:
806 ES (82,01%)
806+ ES (17,99%)so you need to know your league prices, to check what is worth to try
simple ravine
yeah
would be interesting to see what the price would be if you set the threhshold higher, say ~950 es
elder hearth
lol want to know to hit a perfect 3T1 only? 23079 chaos + 11,3 ex
simple ravine
that is... expensive
elder hearth
yeah, but you hit a lot of 900+ in the process
simple ravine
true
elder hearth
how are you going to make some advanced crafting simulation with your tool?
chaos spam is "easy", but weapon crafting and alt+regal requires some conditions
simple ravine
I think I will attempt to create a brute-force algorithm to find the best order first, and then go from there
I have some vague ideas, but haven't tested them yet
to make it easier to simulate multiple different items without knowing the perfect order etc
but by giving it a goal, with must-have, should-have, nice-to-have, acceptable etc
elder hearth
what do you have in mind to test the odds to generate "X" mods when chaos spamming?
I tested in game and got some values.
Then went to poeapp.com and got a 1300 items sample and it matched.
I don't know if @polar island would help us querying his DB
also don't know if items in public stash would be "biased"
simple ravine
not for chaos spamming, as that is unpredictable as you know, but with transmute-alt-aug-regal-exalt
elder hearth
with normal > magic, its pretty much 50% 1 mod, 50% 2 mods
for chaos spam i'm not 100% sure if it really is 65/25/10 or 65/27.5/7.5
prisma prism
http://betaapi.pathofexile.com/api/public-stash-tabs?id=0-0-0-0-0
prisma prism
@compact isle Can you enlighten me a bit about new monster's skill effect power calculation? I noticed changes in GrantedEffectsPerLevel.dat, it seems like new float fields used somehow to interpolate values between levels, but i'n not sure what is the real formula there...
compact isle
I just do web stuff ๐ I have no idea how that works
prisma prism
๐ฆ Do you know anyone who can help me with this question? ๐
timid hemlock
You could just wait for me to figure it out ๐
It's next on my list of things to fix
prisma prism
@timid hemlock That's my fallback plan ๐
timid hemlock
๐
Well I think I know how to go about doing it, so with some luck it won't take long
chrome topaz
@compact isle did you see my email I sent to Paul and Thomas?
about an hour ago
or even, did they see it
chrome topaz
okay well Chris told me to email them, it's about beta online api having wrong encoding
it would be great if you could make it utf-8 instead of latin-1
compact isle
does the regular one have utf8?
chrome topaz
yeah
topaz whale
foreign language suppport prob means everything will have utf-8. Char names n such
but latin 1 is weird
prisma prism
rip non-asci char and stash names?
topaz whale
maybe? dont know about latin 1
compact isle
it's for a different API than the public stash one
looks like it's wrong all the way down, so you'll have to wait for Thomas to fix it I'm afraid @chrome topaz
chrome topaz
you mean like the db's wrong encoding?
compact isle
no the tool used to export the data
compact isle
indeed
prisma prism
@chrome topaz thanks for poe.trade for beta man
velvet fog
anything improved on beta stash api?
chrome topaz
I don't think so
mortal bone
Looks like it was just reposted on a different sub domain. No changes that I can see
obtuse citrus
@timid hemlock Not sure what your progress is, but my idea was to check old vs new skill values
There are new values in default monster stats, possibly the StatXFloat is just a multiplier to that
obtuse citrus
After looking at it it's more complciated
This is what might be of relevance
GameConstants.dat:
SkillDamageIncrementalEffectiveness 360246
SkillDamageBaseEffectiveness 3885209
DefaultMonsterStats.dat:
Unknown2 column
"Difficulty" column
Damage2 column
GrantedEffects.dat:
Multiplier1 column
Multiplier2 column
GrantedEfffectsPerLevel.dat:
StatXFloat columns
Keys0 [I'm guessing this is a link to GrantedEffectGroups, which may determine how it works exactly]
StatData
elder hearth
@obtuse citrus how can you guess those weird long values are linked to something? (is that a pointer value?)
obtuse citrus
longs are almost always pointers
(well, foreign keys to be exact. pointers into data and self are ints)
elder hearth
@obtuse citrus so long = pointer to table value + row value?
obtuse citrus
no, long are usually foreign keys containing the row number of another dat file
elder hearth
@obtuse citrus do you use any HEX editor? to view those files or something?
elder hearth
so u don't need to extract them raw and view for any particular reason?
obtuse citrus
Not really
prisma prism
I looked through the damage growth formula for the SRS and it seems like it is some sort of quadratic function.
The basic formula for it is: a * x^2 + b * x + c
Two multipliers from GrantedEffect can be used as a and b coeffs
I tried simple formula (Mult2 * (level ^ 2) + Mult1 * level) * effectValue for lvl3 GrantedEffect and the result was surprisingly close to the lvl68 GrantedEffect values (i used Flame Sentinel data)
3: 7,36 - 12,27
8: 24,36 - 40,61
13: 47,28 - 78,79
18: 76,10 - 126,83
23: 110,83 - 184,71
28: 151,47 - 252,45
33: 198,02 - 330,03
38: 250,48 - 417,47
43: 308,85 - 514,75
48: 373,13 - 621,89
53: 443,32 - 738,87
58: 519,42 - 865,71
63: 601,44 - 1002,39
68: 689,36 - 1148,93
While the lvl68 values for Flame Sentinel should be 612 - 920
I'm pretty sure here float values comes into play to tune formula even more
simple ravine
What, are there APIs that only a select have access to?
That must be very aspiring for other developers to get into supporting POE with auxialiary tools
prisma prism
I think there is a special api to check player's online for poe.trade
velvet fog
they never said an API improve at 3.0
yes, a api with access key with a online players accouont name list
for privacy issue, they don't want to provide this in public
simple ravine
i completely understand that it's not open for anyone at any time
topaz whale
which is fine for most people, if an inv is updated in the last couple mins, player is most likkely online
simple ravine
but documenting that it exists for site developers such as xyz/poe.trade in case you want to provide an alternative site, and document how it works so you can mock its functionality and plan for it
because all of the sudden, someone with the resources and the will to invest some time and effort into this game will get addicted to this game, and may want to develop something, and make a noticable impact with their efforts... cough hint
but if it's near impossible... well...
velvet fog
I hope they can provide private stash as an API, it's really hard to access large private stash
simple ravine
I hope they wake up from their indie-stubbordness and realize that growing a company to the next level isn't the same as getting it off the ground with a good idea.
mortal bone
45 requests every 60 seconds sucks
velvet fog
but this API may against how they want to sell pubic stash for earn money
mortal bone
what kind of functionality are you looking for?
simple ravine
if they provide a public stash api with literally 0 documentation (except the one u have to google and dig for on various forum posts and reddit), you'll be having gray hair before you figure it all out
topaz whale
its not that bad
mortal bone
I bet that if you were to create a sucessful site they would be more than willing to give online api access
topaz whale
its just json
simple ravine
the json itself is fine
mortal bone
pretty easy to consume
simple ravine
right, but i have 0 knowledge if that's also a river, or a lookup API or whatever it is
velvet fog
the format for public stash api is pretty bad
topaz whale
http://pathofexile.gamepedia.com/Public_stash_tab_API shows whats what
simple ravine
@mortal bone I'm well aware
simple ravine
I believe it was pretty clear
velvet fog
they just merge forum display json and left much garbage inside it
mortal bone
you don't get to know what the online api is
that is just how it goes
develop something and build their trust
simple ravine
I don't expect to get the actual data. I would love to see how it works, so I can plan for it in my architecture
velvet fog
online api a url like online-users-api, and every line is current online account name
simple ravine
i don't understand
mortal bone
You could just ask @chrome topaz if he can say anything about how the data is presented. I highly doubt that he can though
velvet fog
I got one for garena
mortal bone
it is a big list of account names
velvet fog
yes
velvet fog
a utf8 txt file, each line is an account name
topaz whale
lastCharacterName is about as close as your going to get, and can check a db to see if the user has any recent changes unless you talk to devs and see if they can open up that API for consumption for you
that or accountName
velvet fog
accountName for online api
velvet fog
you just need a well formatted documents :p
mortal bone
You wanted to know how it was laid out, so he told you lol
simple ravine
What I meant was, if GGG would have invested a couple of hours into documenting and outlining the details, people might have picked it up, and made something interesting that would impact GGG's popularity and indirectly revenues
mortal bone
I get that you want official information, but if you have an open online api, you can cause a lot of problems for players
It doesn't really
simple ravine
I... do not care... if it is open
I just wanted the documentation
so I could plan for it
topaz whale
docs on the stash tab api? or just on a player online/offline change api?
mortal bone
We all have sat down and figured it out. It doesn't impact them at all. We all sit here and bitch about lack of documentation, but we get over it and figure it out
@topaz whaleI think he wants public design documents
topaz whale
mm
simple ravine
Right. But what you do not know is how many that gave up, and might have made something very impactful if they had official documentation.
mortal bone
You don't know that either
simple ravine
And I understand that you couldn't care less about that, but GGG should
mortal bone
We are making the an assumption, albeit opposite sides, but we have the same data pool lol
simple ravine
Heh, whatever bud
mortal bone
Most of the 'apis' we use are chopped up things that were meant to be used internally
We have 2 apis that are official, and they have documentation
simple ravine
You call that documentation
I mean no disrespect or offense whatsoever
But that documentation page is as barebones as it gets
I'm not sure what you're used to, but... I wouldn't call that documentation
topaz whale
it was enough for my purposes, not sure what the issue is still
mortal bone
Well, I am finishing up college, so I am pretty used to figuring shit out. I also work with a ton of legacy software with almost no documentation. That documentation page has the end point and tells you how to consume the json
Seems good enough
simple ravine
well...
mortal bone
http://pathofexile.gamepedia.com/Public_stash_tab_API community documentation is pretty good if you don't like the official stuff
simple ravine
Thanks for pointing it out (seen it). I just wish that GGG as a business would tkae it more seriously
mortal bone
I get that you want what means what, but none of this was meant to be consumed publicly
simple ravine
I'm a shareholder/stakeholder in a company roughly at the same size as GGG, slightly bigger, but probably a bit larger turnover. If we would have 'winged' it like they do, we'd be out of business by now.
mortal bone
The river is a bandage for the hack we had before
Thing is, they aren't in the business of devs
Third party software helps them, but the lack of it doesn't hurt.
simple ravine
Indirectly, it does.
topaz whale
its player QOL, but people are happy with xyz's work right now, so its not going to help thema ton to hold peoples hands through making new stuff
mortal bone
Honestly, if you don't want to do third party dev work then don't, but you will have to use the resources everyone has put together to dev or figure it out on your own.
At the end of the day it isn't changing, nor will complaining about it fix it. We have been doing it for years.
simple ravine
I'm not sure why you take such an offense because I voice my opinion
topaz whale
don think anyone is taking offense
simple ravine
I thought GGG set this thing up
mortal bone
It was a bandage on a system that was hacked together
There are tons of points where Chris says that he regrets the river, and they were very reluctant for years
Not taking offence, it just seems odd that you seem to be ignoring all of the information being given lol GGG documentation won't happen, and you are in an echo chamber here. We want more documentation and official support, but it just won't happen
simple ravine
How do you know it just won't happen?
mortal bone
Because we have been asking for years. Lol
simple ravine
You're saying that investing effort and resources into GGG third party stuff is betting on a dead horse?
mortal bone
Yep, pretty much
velvet fog
I think yes
simple ravine
If so, I'll jump on another bandwagon
mortal bone
You can definitely get value out of your investment (Poe.trade), but you will have to put some effort into it.
If you are looking to make money, you may want to look somewhat else though. I think there are only a couple of websites that are profitable, and I think that is it.
simple ravine
I'm not in it for the money, to be honest. I don't need the money. But I don't want to put effort into something that is never going to improve whatsoever.
topaz whale
poe.trade monoploy is strong. for people who want more stuff, they usually dev for themselves and keep it from others so they have a competetive market edge
mortal bone
The public stash tab is the only 'improvement' that has happened for devs
I can't think of any other improvements that have been done
simple ravine
Evolve or get extinct. Trust me. I've been in the online business game for quite a while
not 'game' as in a game to play... you get it
mortal bone
The thing is devs aren't their main business. As long as Poe.trade sticks around they don't need anyone else.
mortal bone
And if it goes down someone will pick up their place because it is profitable.
No, xyz makes too much to leave lol
http://www.alexa.com/siteinfo/poe.trade no way to you quit something like this
elder hearth
well, about trading, GGG DO NOT want to be easier as it already is.
So what we can expect about trading (or API related to it) is to stay the same way it is or get worse.
topaz whale
its really not bad how it is
elder hearth
yes, its currently "fine"
mortal bone
Eh, I am just providing counter examples to your points.
mortal bone
You're telling me if you had a site with ~6mil page views a month you would leave it?
topaz whale
eh, think ur just being difficult. if you can articulate your point a little better maybe we could see your side. right now youre making it sound like no dev would want to touch the api to make something... and that its terrible docs that you cant use to figure out how to consume the data. As a hobby dev coming from softwae QA thats messed with it i dont see that thats a valid arguement
simple ravine
Ok, so the documenttion clearly points out how the data is supposed to be used?
topaz whale
there is no intended use, thats up to the 3rd party dev
some do price stats, some do trade tools
its all just whatever u want
mortal bone
It is if you understand where the data comes from
mortal bone
It is forum display data that has evolved over the years, and there is documentation explaining what it means
simple ravine
not the official docs, no
Emmitt, you've pointed that out, and that is so totally besides the point
topaz whale
k, so you just want the wiki rewritten on official site?
simple ravine
๐คฆ I think we'll end this topic now
Last time I checked, it was unofficial
But GGG acknowledges it
velvet fog
asking GGG to provide more API or more documents is a dead end
simple ravine
because the business is operated by developers, and not by business people
and develoeprs hate documenting
elder hearth
simple ravine
devs that don't document and comment code in my businesses are quite short-lived
it's boring but extremely important
topaz whale
from what I've seen in my time as software QA at companies, it usually an effort form both QA and docs coordinating learning sessions with docs peeps so they can write it up. left alone adev isnt likely to document much except for thier own use
*with devs
simple ravine
it depends on how your organization looks like
topaz whale
obviosly if you dont employ docs people it has to happen differently
simple ravine
management is another thing that needs certain qualities
had a cto in one of our businesses, that was technically savvy, young and wasn't ready for the role - he was there when we acquired it
after some time, it was apparent we had to replace him with someone with more business experience
now, the team thrives
chrome topaz
lack of documentation is not even the worst part
topaz whale
i think its easy enough too figure out once you break it down
but I only do simple name, price , league searches, so i dont have to deal with the nightmare of figuring out rolls and junk
btw, just curios what your current db size right now is @chrome topaz it has to be huge
mortal bone
@chrome topaz notice any major changes to the beta api?
chrome topaz
not even minor changes
topaz whale
still not utf-8?
mortal bone
Damn, OK...
chrome topaz
I think it always was utf8, no?
topaz whale
you mentioned it being in latin 1 the other day. I havent looked at beta api yet. my tool is only really useful when your site is having issues so i can stilll trade
maybe I just misinterpretted the convo. dunno
chrome topaz
it was online api
topaz whale
ah. k
thin loom
Does anybody know what is the class type of vaal skill gems for lootfilter syntax?
obtuse citrus
Active Skill Gem
thin loom
That's showing all skill gems though
topaz whale
Class Gem
BaseType "Vaal"
thin loom
ah OK
topaz whale
jus searched it out in neversinks filter since i have it downloaded
thin loom
I'm trying to modify carcass's since it's simpler
Good idea to search stuff in Neversink's, he has pretty much everything in there
topaz whale
yeah, i figured he had something like that
btw, http://filterblade.xyz/ is pretty good if you want to graphiclly edit lootfilters. its neversinks latest project
timid hemlock
@obtuse citrus Well I'm getting nowhere trying to figure out this stat scaling
The values don't seem to scale in any reasonable way from any input I can find
topaz whale
you doing fun maths?
timid hemlock
Very
topaz whale
is it just trying to find a furmula for a sequence?
timid hemlock
Pretty much
topaz whale
you have the values?
timid hemlock
Yeah
topaz whale
timid hemlock
I doubt it would help; assuming it's a polynomial, it appears to be at least 4th degree
thin loom
I can't find a way to filter gems by their level
Like if I want to hide all but level 20 gems
topaz whale
๐ฆ
timid hemlock
It's gone now
lethal flume
gone with the rage
timid hemlock
Yes.
obtuse citrus
I messed around a bit as well and didn't get any results
I think first step to take here is to go in game and start recording damage values on skills and then see if that helps getting anywhere
or maybe ask chris about it but idk if they want to share the info on monster dmg calcs
topaz whale
you tring to do monster dmg or player skill dmg?
obtuse citrus
monsters
topaz whale
yikes
obtuse citrus
they changed stuff in 3.0.0 for spells, before you had the values
now they scale differently
topaz whale
sounds like a nightmare. ingame tests have so many variables to consider
prisma prism
What is the easiest way to guess function from dataset?
topaz whale
http://pathofexile.gamepedia.com/Monster_damage even the old explination and estimation and spreadsheet source are cringe. if youre trying to replicate its gunna be a fun time. gl.
timid hemlock
Oh, I don't think I've mention in here the one thing I am certain of: the float values in GEPL scale the stats from one base value that is calculated for each level
So most of my number crunching has been working from those base values
And the float values in GE are definitely correlated with those base values
Beyond that, NFI
prisma prism
@timid hemlock Yep, came to the same idea suddenly
timid hemlock
Also, I think I know what those 2 float values imply
Multiplier1 is the magnitude
Multiplier2 is the scaling rate
Higher values of Multiplier2 cause the base values to scale up more quickly
prisma prism
@timid hemlock Do you have example with Multiplier2 higher then 0.05?
timid hemlock
Not yet
prisma prism
Most of those i've seen were around 0.03
timid hemlock
Yeah
I'm trying to gather more data on that ATM
Bit of a pain in the ass to do though
prisma prism
@timid hemlock filtered some skills with M2 > 0.06 https://pastebin.com/p7dcuxuE
timid hemlock
Because people who play spectre builds like to know how much damage they can deal, and my program can tell them, but only if I know how to calculate said damage ๐
elder hearth
nice, I never played spectre.
Are their damage just the same as the wild ones or they scale down/up when they are a spectre?
prisma prism
yeah
this is also may be usefull to answer questions like "How many damage will deals X boss"
obtuse citrus
Still need to figure out the formula and what kind of formula it is at the end of the day
timid hemlock
Yup
prisma prism
Currently i'm looking for two samples with fixed M2 and variable M1 both to be present in old and new data
obtuse citrus
if they just didn't copy damage scaling 1:1 there is a chance that new values would deviate from old values
hence it might be good to pick a monster and test it
timid hemlock
True, but I'm guessing they'll be close enough to at least hint at the formula, given enough data
obtuse citrus
might be yea
In the python shell you can have both versions side by side pretty well
prisma prism
I'm pretty sure it will be close enough, i've seen exmples with 8-10 EffectPerLevel in new data to match old values
timid hemlock
I don't think the stat values in GEPL are even used when there's scaling in play
obtuse citrus
yeah I suspect so as well
mortal bone
What files are you guys looking into? I might be able to help a bit as well
timid hemlock
GrantedEffects and GrantedEffectsPerLevel
mortal bone
๐
mortal bone
Do we have any target values?
timid hemlock
Oh I can give you a massive text file of target values ๐
timid hemlock
Yep
And my analysis so far seems to strongly suggest that it is the case
I've made some other discoveries, BTW
Multiplier1 is the final multiplier for the base value
So, if Multiplier2 is, say, 0.03, then a Multiplier1=2 gives twice the base values of Multiplier1=1
But Keys0 seems to interfere with that somewhat
Only when the corresponding Keys0 is the same for both stats do the values line up properly
mortal bone
Hmm, interesting...
prisma prism
For example Stats0 = 8 mean that there is no approximation
timid hemlock
Yeah I got that far
prisma prism
1 - damage scaling
timid hemlock
Anyways, I gotta sleep; so good luck ๐
prisma prism
Also there are different damage scaling mods
I spotted mode 1 used for DOT damage in RepeatedPulseDegen
simple ravine
the feeling when updating wildcard ssl on your microservices, you realize a dev use the SSL cert to sign JWT tokens... ๐คฆ
simple ravine
Guys... in stats, there are (for an eaxmple) these stat objects:
"spectral_throw_deceleration_override": {
"alias": {},
"is_aliased": false,
"is_local": false
},
"spectral_throw_projectile_deceleration_+%": {
"alias": {},
"is_aliased": false,
"is_local": false
}
however, only the second has a translation
elder hearth
๐ค
torpid mesa
perhaps the first is for monsters
simple ravine
might be, indeed
I'll just index the stats tthat actually have a translation and throw the other ones away, and see what happens.
torpid mesa
there is a monster that has a modified projectile speed in the beta
for spec throw
so its prob related
simple ravine
this is 2.6 data still though
torpid mesa
hmmm
elder hearth
there's no "StatsKey1": 2731, (2731 = spectral_throw_deceleration_override in Stats.dat)
in Mods.dat
maybe not used
simple ravine
and then, there seems to be no 'local_jewel_effect_base_radius' in stats
ugh
(reading mods, trying to find the equivalent in stats)
elder hearth
isn't local_jewel_effect_base_radius just a radius value? (don't need description, its visual)
simple ravine
probably, however I would of assumed that there was going to be an entry in stats.dat for it
elder hearth
it is there
simple ravine
hmm, ok - perhaps I need to rerun RePoE with other options
elder hearth
? what did you want in stats.dat besides
"Id": "local_jewel_effect_base_radius",
"Flag0": true,
"IsLocal": true,
"IsWeaponLocal": false,
"Unknown2": 3,
"Flag3": false,
"Text": "",
"Flag5": false,
"Flag6": false,
"MainHandAlias_StatsKey": null,
"OffHandAlias_StatsKey": null,
"Flag7": false,
"Unknown3": -349201562,
"Unknown0": []
}```nothing really, to be honest, i wrote my application assuming that every stat-id on each mod generated would have an equivalent entry in stats ... based on the json files that are generated by RePoE using PyPoE
so my application threw an exception when there was no entry in the dictionary for that
elder hearth
yeah you will need to get exceptions hehe
simple ravine
DEFAULT_APPNAME Warning: 0 : There is no stat object for movement_velocity_while_spider_+%
DEFAULT_APPNAME Warning: 0 : There is no stat object for is_intimidated
DEFAULT_APPNAME Warning: 0 : There is no stat object for local_off_hand_visual_identity_override_hash
DEFAULT_APPNAME Warning: 0 : There is no stat object for local_jewel_effect_base_radius
DEFAULT_APPNAME Warning: 0 : There is no stat object for some_passives_apply_to_minions
DEFAULT_APPNAME Warning: 0 : There is no stat object for phase_through_objects_while_spider
DEFAULT_APPNAME Warning: 0 : There is no stat object for power_charge_art_variation
DEFAULT_APPNAME Warning: 0 : There is no stat object for shock_art_variation
(and a lot more)
elder hearth
while_spider
wtf
elder hearth
what kind of stuff your app is going to work with? (i mean items like, jewels, flasks, equipments, maps, map areas, monsters, etc)
simple ravine
right now, i'm creating the underlying generic library that the application will use
linking together mods, stats, translations and base-items with domain model logic
elder hearth
hmm, maybe you can filter your stuff by domain (if you need it)
simple ravine
I made json deserializing into domain models, and indexing them - and the unit test threw an exception
so at this point, it doesn't filter by any domain or groups, it'll index whatever it's given to it
timid hemlock
Ok, made another discovery with the stat scaling
Keys0 is an index to EffectivenessCostConstants
The values in that table serve as an extra multiplier to the base value
mortal bone
Can't get pypoe to play nice with the beta data ๐ฆ
timid hemlock
Oh? I've had no trouble
Well, I did have a little trouble but Omega clonked it on the head
mortal bone
Yeah, I am having incorrect rowsize troubles
I saw that he had fixed it at some point above, but I have the latest release and have issues
timid hemlock
This in the exporter?
mortal bone
both exporter and cli
mortal bone
I'm dumb
timid hemlock
K ๐
mortal bone
I had set the path but not the version...
timid hemlock
Yerp, that'd do it ๐
mortal bone
Thanks ๐
timid hemlock
Too much data?
obtuse citrus
It's pretty
timid hemlock
Very
And it still tells me nothing :/
So, there's some function that takes level and Multiplier2 and produces that output
obtuse citrus
I'll change Keys2 since you got that one
timid hemlock
๐
So the formula for calculating a stat value that I've derived so far is:
StatXValue = F(level, Multiplier2) * Multiplier1 / EffetiveCostConstant * StatXFloat
It's "F" that's the tricky bit now
compact isle
so what column names do you guys have at the moment?
timid hemlock
For what exactly?
timid hemlock
compact isle
oh cool, I'll take a look
obtuse citrus
starting @ line #2268
timid hemlock
Or he could just search for the .dat name ๐
compact isle
^ is what I ended up doing ๐
huh, the column positions all change in the .dat files that's weird
timid hemlock
Interesting; I wonder how that happens
obtuse citrus
odd
prisma prism
@timid hemlock hmm if Keys0 is multiplier then how does the "not-interpolated" stat marked?
timid hemlock
The Id of #8 in EffectivenessCostConstants is "Override", and the multiplier is 0
But I don't think that's what marks it
StatData in GEPL is 3 for interpolated stats, and 1 for non-interpolated
I think it's a key to StatInterpolationTypes
timid hemlock
๐
prisma prism
@obtuse citrus btw i have an issue with PyPoe on my Win10... console parser doesn't work when i use python 3.4 (works fine with 3.6) and gui doesn't support anything above 3.4...
obtuse citrus
it sounds like a bug in python perhaps if it works with newer versions of python
prisma prism
yeah, probably
obtuse citrus
I can upgrade to pyside2 but it would require everyone to compile it themselves, it's still under heavy development
What is the exact problem you're getting anyway? maybe there are work arounds
prisma prism
@obtuse citrus argument parser never execute code and only shows help even if all arguments are correct
obtuse citrus
yeah that sounds like an issue with python itself since I use the argparse module
you could try to execute the files directly inestad of the entry point scripts
see if it helps
prisma prism
i wasn't able to perform proper cli configuration
anyway, i ended with two python versions. not that bad, just some tricky pathing required
vapid glade
is procurement working with beta?
compact isle
No, there isn't any forum trading in beta
south mauve
They said API will be open soon no ?
simple ravine
@south mauve If you're refering to Stash Tab API for Beta, it is open.
beta.pathofexile.com/....
simple ravine
the stat value weight within a specific mod should be equal, right?
velvet fog
Mod have weight
simple ravine
within a tier, i.e. IncreasedLife5 has a X to Y for an instance
within that tier, I am assuming that the values are equally weighted?
velvet fog
I think yes
elder hearth
does anyone knows if brownymaster (poecraft.com dev) uses this Discord?
deft jolt
Anyone know for how long the api timeouts after too many requests?
desert needle
it's just a few minutes iirc
but I think they may have used some sort of auto rate limiting that may increase if you hit the limit repeatedly
deft jolt
Hm, ill write a script to figure out the exact numbers then
mortal bone
For which api?
desert needle
(I was presuming the stash tab api, but yeah it has different rate limits than the other api's iirc)
mortal bone
Yep it does
2 per second with a minute or so timeout iirc for the stash tab api. The other apis are 45 requests then you are timed out for 1 minute
deft jolt
Ah yes, I had the sash api in mind
mortal bone
I think it is around a minute
Oh, you are asking when you get timed out?
You get a timeout if you make more than 2 requests per second
mortal bone
Making more than 2/s will time you out, and I believe that time gets increased
mortal bone
Ah ok
mortal bone
Been a while since I messed with the time outs
deft jolt
Might aswell increase the delay
simple ravine
I think it's 1 request per second for Stash API
deft jolt
I wrote some code about a month ago that calulates a delay to keep requests about 0.6 to 0.9 sec apart
Sometimes it messes up and ends up getting timed out
Was looking into rewriting it
simple ravine
What I did was making sure that the time between requests are 1s apart (including download time) plus a little padding
Because the time to get the payload takes quite a while sometimes
deft jolt
I stream the data
First chunk always contains the next change id so I can pretty much make a new request instantly
while the stream finishes downloading in another thread
simple ravine
Wouldn't that create concurrency issues with the item data?
elder hearth
I tried that too, it creates, at least on my end
deft jolt
Not really
simple ravine
Not really, as in no definitely, or?
deft jolt
Thats a good question
can you explain it in more detail?
English isnt my first language
deft jolt
ah i see