Old topic - needed something similar today and only found this topic after.
This is the prototype I got to:
https://gist.github.com/BigRoy/60883eff23f73a34c4671395b32d858d
Note that unlike the code samples above mine doesn't go over individual uvs
but uses the bounding boxes of UV shells to compute the UDIM tiles.
Side effect of that is that it will detect UDIM tiles if it's encompassed
in a UV faces and all points are outside of it (very rare case in
practice!).
However, my code will fail to detect 'empty tiles' if a single UV shell
goes over tile boundaries like circling around it.
It is however, faster for much higher vertex/uv counts.
Also, the code snippets above in this thread can be faster even with Maya
Python API 2.0. I'll add that as a comment on my gist too.
On Saturday, October 14, 2017 at 8:58:30 PM UTC+2 kelvi...@gmail.com wrote:
> On Monday, May 4, 2015 at 2:18:24 AM UTC-4, Anthony Tan wrote:
> > Don't thank me til it works :D I'm doing all of this in such a
> non-production setup so I'm almost certainly working in a best-case
> scenario, but yes, let me know how you go and if you want feel free to
> bounce me a scene or your existing code and I'll have a go with it here to
> see if anything jumps out at me - it'll stop me from wasting time with my
> day off playing with a useless box and setting electronics on fire via poor
> soldering technique..
> >
> >
> >
> >
> >
> >
> > On Sun, May 3, 2015, at 04:47 PM, Chad Fox wrote:
> >
> >
> >
> > Thanks Anthony. I'll review and test out your code drop on Monday. As a
> side note, 500k uvs was unfortunately too sluggish for us, perhaps it's
> something to do with our environments or machines, but it wasn't resonable.
> God forbid someone runs my script with the UV editor open. Might as well
> pack up for the day :P
> >
> >
> >
> >
> > Anyways. Thanks again!
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > -Chad
> >
> >
> >
> >
> >
> >
> >
> > On Fri, May 1, 2015 at 7:24 PM, Anthony Tan <antho...@greenworm.net>
> wrote:
> >
> >
> >
> >
> >
> >
> > Thought I'd just target the uv query to see if that can be sped up any,
> or where a bottleneck would be (can't imagine it's converting UV to UDIM
> coordinates, that's almost a one-liner).
> >
> >
> >
> >
> > So yes, taking three approaches, PyMEL/getUVs feels manageable out to
> say, 2-300k UVs, but does feel sluggish at 4-500k. Maya.cmds is just
> diabolical - that's also because I'm not very good with it and I'm 100%
> sure i'm doing it a horrible way, but the API is fine through to a million
> at which point I figured it was Fast Enough.
> >
> >
> >
> >
> > One thing I'd point out with my API approach - I've just done the bare
> minimum to get a result - I'd trust PyMEL over my implementation of UV
> grabbing until I did some more testing, wouldn't be surprised if there are
> issues I'm glossing over in this test scenario that PyMEL handles
> correctly, and I don't handle at all.
> >
> >
> >
> >
> > Just using a polysphere with 1000x200 subdivs to give me ~200k UVs, my
> numbers were roughly 4s for maya.cmds, 0.4s for pymel, and 0.001 for the
> API. The API approach scales decently as well, while I could've sworn pymel
> didn't have a linear growth pattern. (my local machine, Maya 2013x64/SP2)
> >
> >
> >
> >
> > checking pSphereShape1 for 1 iterations
> >
> >
> > maya.cmds : 3.95910561149
> >
> >
> > PyMel : 0.344888150405
> >
> >
> > OpenMaya/API : 0.00120376245754
> >
> >
> > ---------------
> >
> >
> > (for 201199 uvs):
> >
> >
> > ---------------
> >
> >
> > DONE
> >
> >
> >
> >
> > The figures above are just for grabbing the UVs and not processing them,
> but depending on what cases you want to trap, the UV/UDIM thing didn't feel
> expensive at the ranges you were quoting (out to 500k) using my naive
> approach of testing every coordinate and building up a set, so I didn't
> investigate that too much further. If you wanted to do things by-shell i
> have a feeling API is going to be where you want to head.
> >
> >
> >
> >
> > -Anthony
> >
> >
> >
> >
> > (Here's my code dump, scuse the mess but you should be able to execute
> the thing as it stands to replicate my results once you create a polySphere
> or similar to play with)
> >
> >
> >
> >
> >
> >
> >
> >
> > import timeit
> >
> >
> >
> >
> > mesh_string = 'pSphereShape1'
> >
> >
> > timer_iterations = 1
> >
> >
> >
> >
> > # pymel version
> >
> >
> > import pymel.core as pm
> >
> >
> > def pymel_getuvs(mesh_string):
> >
> >
> > """return a list of two lists, idx 0 is U, idx 1 is V"""
> >
> >
> > mesh = pm.PyNode(mesh_string)
> >
> >
> > uvs = mesh.getUVs()
> >
> >
> > return uvs
> >
> >
> >
> >
> > # mc version
> >
> >
> > import maya.cmds as mc
> >
> >
> > def mc_getuvs(mesh_string):
> >
> >
> > """return a list of (u,v) tuples"""
> >
> >
> > x = mc.getAttr(mesh_string+'.uvpt', multiIndices=True) # determine
> UV idxs
> >
> >
> > uvs = []
> >
> >
> > for i in x:
> >
> >
> > # probably quite a naive way to iterate through an object's UVs.
> I don't
> >
> >
> > # use maya.cmds much
> >
> >
> > uvs += mc.getAttr('{mesh_string}.uvpt[{idx}]'.format(idx=i,
> mesh_string=mesh_string))
> >
> >
> > return uvs
> >
> >
> >
> >
> >
> >
> > # OpenMaya version
> >
> >
> > import maya.OpenMaya as om
> >
> >
> > def om_getuvs(mesh_string):
> >
> >
> > """return a list of two lists, idx 0 is U, idx 1 is V"""
> >
> >
> > # Y'know it's really weird to think of malloc-ing in python..
> >
> >
> > selection_list = om.MSelectionList()
> >
> >
> > mObject_holder = om.MObject()
> >
> >
> > u = om.MFloatArray()
> >
> >
> > v = om.MFloatArray()
> >
> >
> > function_set = om.MFnMesh()
> >
> >
> >
> >
> > # see
> https://groups.google.com/forum/#!topic/python_inside_maya/usFLgzJBrpM/discussion
> >
> >
> > # for a note on why this, instead of a flat selection_list.add.
> >
> >
> > om.MGlobal.getSelectionListByName(mesh_string, selection_list)
> >
> >
> > iterator = om.MItSelectionList(selection_list)
> >
> >
> > iterator.getDependNode(mObject_holder)
> >
> >
> > function_set.setObject(mObject_holder)
> >
> >
> > function_set.getUVs(u,v)
> >
> >
> > return [u,v]
> >
> >
> >
> >
> >
> >
> > def uv_to_udim(u,v):
> >
> >
> > '''return UDIM tile corresponding to UV coord
> >
> >
> >
> >
> > NOTE:very poorly defined response on edges..
> >
> >
> > '''
> >
> >
> > import math
> >
> >
> > return int( 1000+(math.floor(u)+1)+(math.floor(v)*10))
> >
> >
> >
> >
> >
> >
> > # the zip function itself is a bit slow
> >
> >
> > def equivalence(pymel_result, mc_result, om_result):
> >
> >
> > pm_uv = zip(pymel_result[0], pymel_result[1])
> >
> >
> > om_uv = zip(om_result[0], om_result[1])
> >
> >
> > for i,x in enumerate(pm_uv):
> >
> >
> > if pm_uv[i] != om_uv[i]:
> >
> >
> > raise ValueError('pm != om')
> >
> >
> > elif pm_uv[i] != mc_result[i]:
> >
> >
> > raise ValueError('pm != mc')
> >
> >
> > print "OK"
> >
> >
> >
> >
> >
> >
> > print "checking {mesh} for {it} iterations".format(mesh = mesh_string,
> it = timer_iterations)
> >
> >
> >
> >
> > print "maya.cmds :",
> >
> >
> > t = timeit.Timer(stmt=lambda:mc_getuvs(mesh_string))
> >
> >
> > print t.timeit(timer_iterations)
> >
> >
> >
> >
> > print "PyMel :",
> >
> >
> > t = timeit.Timer(stmt=lambda:pymel_getuvs(mesh_string))
> >
> >
> > print t.timeit(timer_iterations)
> >
> >
> >
> >
> > print "OpenMaya/API :",
> >
> >
> > t = timeit.Timer(stmt=lambda:om_getuvs(mesh_string))
> >
> >
> > print t.timeit(timer_iterations)
> >
> >
> >
> >
> > print "---------------"
> >
> >
> > print "(for {n} uvs)".format(n=len(om_getuvs(mesh_string)[0]))
> >
> >
> > print "---------------"
> >
> >
> > print "DONE"
> >
> >
> >
> >
> >
> >
> > uvs = om_getuvs(mesh_string)
> >
> >
> > udim_list = set()
> >
> >
> > for x in xrange(0,len(uvs[0])):
> >
> >
> > udim_list.add( uv_to_udim(uvs[0][x],uvs[1][x]))
> >
> >
> > print udim_list
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > On Sat, May 2, 2015, at 06:27 AM, Chad Vernon wrote:
> >
> >
> >
> > If speed is your main issue, you could write it in C++.
> >
> >
> >
> >
> >
> >
> >
> > On Friday, May 1, 2015 at 9:43:18 AM UTC-7, thirstydevil wrote:
> >
> >
> >
> > I do as Janos suggested, worked quite well and pretty fast. We had
> quite heavy props from speed tree and it wasn't that bad
> >
> >
> >
> > def getMeshUVBounds(mapName, mesh):
> > mesh = pCore.PyNode(mesh)
> > uList, vList = mesh.getUVs(mapName)
> > if not uList:
> > uList = [0]
> > if not vList:
> > vList = [0]
> > return [[min(uList), min(vList)], [max(uList), max(vList)]]
> >
> >
> > def getMeshUVBoundsAsUDIM(mapName, mesh):
> > udimMap = range(1001, 1011)
> > bounds = getMeshUVBounds(mapName, mesh)
> > minU, maxV = int(math.trunc(bounds[0][0])), int(round(bounds[1][1] +
> 0.5))
> > return udimMap[minU] + (10 * (maxV - 1))
> >
> >
> > def isUDIMOverlappingBoundaries(mapName, mesh):
> > bounds = getMeshUVBounds(mapName, mesh)
> > if int(bounds[0][0]) == int(bounds[1][0]) and int(bounds[0][1]) ==
> int(bounds[1][1]):
> > return False
> > else:
> > return True
> >
> >
> >
> >
> > On Friday, 1 May 2015 07:37:06 UTC+1, Chad_Fox wrote:
> >
> >
> >
> > Hey everyone,
> >
> >
> >
> >
> > I've been brute force checking UV coords to determine what UDIMs any
> selected object(s) are using, but my simple python query for all UVs and
> checking each of their coords is painfully slow and in some cases unusable
> in scenes my team will be working with.
> >
> >
> >
> >
> > Do any of you have any snippets of python / python api code that can
> efficiently return a large list of occupied UDIMs from selected object?
> >
> >
> >
> >
> > Appreciate any help you may be able to offer, thanks!
> >
> >
> >
> >
> > Chad
> >
> >
> >
> >
> >
> >
> >
> >
> > --
> >
> >
> > You received this message because you are subscribed to the Google
> Groups "Python Programming for Autodesk Maya" group.
> >
> >
> > To unsubscribe from this group and stop receiving emails from it, send
> an email to python_inside_m...@googlegroups.com.
> >
> >
> > To view this discussion on the web visit
> https://groups.google.com/d/msgid/python_inside_maya/2eb3c040-ca0c-4148-a1c9-e2e69e2eabe6%40googlegroups.com
> .
> >
> >
> > For more options, visit https://groups.google.com/d/optout.
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > --
> >
> >
> > You received this message because you are subscribed to the Google
> Groups "Python Programming for Autodesk Maya" group.
> >
> >
> > To unsubscribe from this group and stop receiving emails from it, send
> an email to python_inside_m...@googlegroups.com.
> >
> >
> >
> >
> > To view this discussion on the web visit
> https://groups.google.com/d/msgid/python_inside_maya/1430533492.2337937.261681645.479A74A1%40webmail.messagingengine.com
> .
> >
> >
> >
> >
> >
> > For more options, visit https://groups.google.com/d/optout.
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > --
> >
> >
> > You received this message because you are subscribed to the Google
> Groups "Python Programming for Autodesk Maya" group.
> >
> >
> > To unsubscribe from this group and stop receiving emails from it, send
> an email to python_inside_m...@googlegroups.com.
> >
> >
> > To view this discussion on the web visit
> https://groups.google.com/d/msgid/python_inside_maya/CAJ4oLs7A76SUDc5V9Ppz3M6wX_8KSDPi_oCx9ZvHzbZWsH%3DaxA%40mail.gmail.com
> .
> >
> >
> > For more options, visit https://groups.google.com/d/optout.
> >
> >
> >
>
> Hi Ive been trying to write my own UDIM helper Script and came across this
> page. I was wondering If anyone is continuing the development of this ?
--
You received this message because you are subscribed to the Google Groups
"Python Programming for Autodesk Maya" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to python_inside_maya+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/python_inside_maya/b9234618-b426-413a-ae27-3528115526cbn%40googlegroups.com.
