Danbooru

[Prototype] User Report Ver 6.3

Posted under General

Squishy said:

That's a torso bend, not a torso twist?

And how come your line goes from the front of the neck to the sternum's ventral side, and then directly connects from there to the dorsal side of the sacrum by cutting through upper anterior of the waist toward the posterior of the pelvis, completely disregarding the suggested alignment of the waist as indicated by the waist bow?

I'm pretty sure that line does not represent the spine nor the body's vertical axis due its shift along sagittal plane. And that abrupt bend from the chest doesn't even follow the curve of her posture, not to mention the lack of any rotation going on. This does not help me understand your reasoning at all.

Since you apparently have a much more expansive of scientific knowledge of human anatomy why don't you please draw how her spine should look?

Latest Update

  • (2018-07-04)
    • Updated data for previous 30 days
    • Temporarily removed page analytics since Google Analytics is no longer used
Edit:

Just realized I forgot to compare the data versus the prior month. >_< Will probably have it up in half an hour or so.

...and done!

Updated

fossilnix said:

How is the keeper report calculated? In June I added tags to a bunch of old minimally-tagged posts, but they aren't showing on the report. Does it only count posts uploaded recently?

It currently only checks uploads during the checked period (i.e. June 1st - June 30th).

Early on I started going down the rabbit hole of wanting to account for older uploads, but then that meant looking up the post data for every single post version during that same period. That would require at least an order of magnitude more time as there are about 200K+ post versions in a month compared to only 30K+ uploads. However, it's not as simple as just getting the post data, as it must also be ascertained on whether THAT particular post version was what changed the ownership so as to not count the same post multiple times for the same user or for the wrong time period. This means getting ALL of the post versions for that post (since the keeper calculations use the tag contributions from every user along with who added a tag first), which in some cases could increase the order of magnitude by another factor or two of time to query and calculate.

TL;DR it's at least 100-1000 times slower, and it was still having issues, so I decided to just scrap it and only account for posts over the checked time period.

skylightcrystal said:

Is this going to be updated?

Sorry about that. It's mostly an automated process, but the final compilation and uploading is all manual so it only gets done when I remember without being distracted by something else.

Anyways, it's updated now.

BrokenEagle98 said:

Sorry about that, it's mostly an automated process, but the final compilation and uploading is all manual so it only gets done when I remember without being distracted by something else.

Anyways, it's updated now.

Thanks

skylightcrystal said:

no update this month?

A limit was added to forum posts last month that limits it to 200K characters. The user report is just under 400K characters. So it's pretty much officially dead. I could split it up perhaps, but that would be even more manual work, and the limit might be lowered again and again in the future.

It was suggested by someone to put it up on another site or something, but retooling the report code will be neither a simple nor easy task, and right now it's toward the bottom of my to-do list.

For what it's worth, I still uploaded the updated CSV and JSON raw data for the month of March, so you can still mostly get everything that's needed.

1 10 11 12 13 14 15