EVE Technology Lab

Hey FoxFour, it's nice to see there is someone working hard for the API! (And probably suffering: to be sure to avoid double-posting I read ALL THE THREAD just now. I can still hear the whining...)

Anyway, I'm working on a tool for gaming groups and guilds who play more than one game and allow their members to view who is online, and on which game, in a centralized place.
Despite demand, I stopped developing the EVE module due to the 6 hour cache on corp member tracking. (Knowing if some of your buddies were online 5 hours ago is useless.)

Reducing the cache to one hour is a change which would bring back this "buddy monitoring" tool in the realm of "usable", since people generally play EVE for extended periods of time.
However, do you think it could be reduced to 30 minutes, or even better, 15? This would put the EVE information on par with the rest.

At the moment, I think a frequent train of thought among users is "Oh, I wonder if there's anyone on EVE? Let's see... To look at corp channel I have to click 5 times, wait 3 minutes, and enter a password... Meh, there's 3 people on Minecraft judging by the monitor, I'll just play that..."

Oddly enough I'm getting this too, and it seems to be a new thing, as previous Corp/IndustryJobs pulls did have the productTypeID with them.

Thank you!

Awesome! I'm going to be using that too, and I'm not even writing an Android application!

I did have some issues with the code though - I've got it running (had to compute numIterations from total running time divided by cycle time, and there were a couple issues with variable names), but it's not computing identical results to what the EVE client does. Is it doing integer math and having issues with me using floats, or? The difference is generally pretty slight, but the total ends up being off by 13% or so. I'm also using the default decayFactor (0.12) and noiseFactor (0.8), because I didn't find any overrides in dgmTypeAttributes.

(Also, I changed it to use sec=1 as opposed to sec=10 000 000, and I'm feeding time in as Unix time with installTime clamped to zero and expiryTime adjusted to match, because currentTime also starts at zero. You get the same results if you leave installTime as-is and just add it to currentTime; if you don't do either, you get nonsense results because cycleNum starts at some hilarious offset.)


Any tips?

Edit: Fixed, PHP is horrible. Ported to C#: http://pastebin.com/8tgU4E8K

Hi Dragonaire,

The facilities API point should be accessible here:

https://api.eveonline.com//corp/Facilities.xml.aspx

It should accept the usual keyID and vCode credentials, and I've been using it successfully since Crius came out.

Hopefully that helps.

I know about corp/Facilities but if you read the patch notes seems to say there's now an industry section to the API

So work on the blueprints endpoint finally progressed today so a quick update:

I hope to have it on Sisi next week and TQ the week after, no promises though.

The cache time on it will be 24 hours.

Here is a sample row:

row itemID="1000000012172" locationID="60014929" typeID="23058" typeName="Dragonfly Blueprint" flagID="4" quantity="32" timeEfficiency="0" materialEfficiency="0" runs="-1"

locationID is the actual item location, in this case a Polaris station. The flagID is then where in that item it is. So 4 is my personal hanger. If it was owned by a corp then flagID would be the hanger division. Similarly if it's located in a POS structure that will be the hanger division.

So determining a blueprints location is a combination of the ownerID (which you should know based on your request), the locationID, and the flagID.

The endpoint will require CEO or directorship as there is no easy way for me to filter the results based on your characters roles.

I do want to mention that there is the possibility we will pull this endpoint and not release it if performance is really ****, but we will see.

The maximum results returned will be 200k which is far better than the 30k max of the client industry window.

The endpoint is restricted to the asset list access mask.

Thats all for now.

Hi FoxFour,

I'm still rather concerned that blueprint PE & ME (especially ME) isn't also passed through to the new industry API. If I understand you right, even with industry roles in my corp, I (as a corp director) would need to manually share the blueprint PE & ME levels to any industrial subordinates (factory managers) for them to be able to run accurate calculations on cost of production from any corp BP's they used? There's no way I'm sharing my API key with a subordinate, and it'd be quite a hassle to update a blueprint list every day just so my industry subordinates could properly calculate cost of production.

Even with the blueprints API I still run into the problem of needing to cross-check whatever ME/TE research I've finished on a particular blueprint between the most recent blueprint API pull I have and whenever I'm looking to start (or have started) a production. Were they in the industry data all I'd need to do is pull the most recent industry job with that blueprint and extract the ME/TE data from there, making adjustments if the most recent job is a ME/TE job. That's one data pull per blueprint as opposed to what would be a variable number with the blueprint API.

Is it really so hard to bring back the blueprint ME/PE levels to the industry API? Having the blueprint ME/TE levels show there would solve the first problem, and alleviate the second.

24 hours is fail. 23 hours is win.

Not going to happen from our side. You can get that data from zKillbaord but we just don't store it in a way that allows us to query based on anything but final blow or victim really. Making all killmails public is something we have considered and will continue to consider, but nothing is going to happen with that for now.