Revision as of 22:31, 4 April 2018 by MarkMiller (talk | contribs) (Databases)


Name Simple Rebaseline Notes
cinema-a.py Yes view/lighting
cinema-c.py (skipped)
cleanzonesonly.py No minor pixel diffs for plots, legend shifted slightly upwards. Should legends be removed from this test?
clonecopy.py (skipped)
clonefirstref.py Yes minor pixel diffs
conditional.py No some view/lighting, but also Interpolation:

conditional_04 conditional_05 conditional_06

conn_cmfe.py Yes view/lighting
coord_expr.py Yes view/lighting
curve_anim.py Yes view
curve_expressions.py No possible view clip issue with curve_exp_bin_02
cylindrical_radius.py Yes view/lighting
ddf.py No Interpolation ddf
defvar_anim.py Yes view/lighting
expr2engine.py Yes lighting
expr_cmfe.py Yes view/lighting
expressions.py Yes view/lighting
field_operators.py No Mostly view/lighting

Missing points


Inerpolation diffs

field_op_14 field_op_16 field_op_17 field_op_18

image_proc.py No Interpolation hybrid_image_proc
keyframe.py No view/lighting, but we should take a look at the strange highlighting of the endcaps on globe keyframe_43
lambda2.py Yes view/lighting
lineout.py (skipped)
locktime.py Yes minor pixel diffs
locus.py No could be just large quads split differently. look at locus_01
math_expr.py No quads split differently? see math_expr_02
matsel.py Yes view/lighting
matvf.py Yes view/lighting
merge_tree.py No Interpolation merge_tree
mesh_quality.py Yes view/lighting
mir.py Yes view/lighting
mir_cache.py Yes view/lighting
missingdata.py No Interpolation ? missingdata_0_01
movie.py (skipped)
multivar.py Yes view/lighting
pos_cmfe.py No Mostly view/lighting, except an Interpolation case pos_cmfe_01
py_exprs.py Yes minor pixel diffs
qcrit.py Yes view/lighting
replace.py No the surface is no longer smooth (curv3d.silo) see Replace3
selections.py (skipped)
selections_pp.py (skipped)
sil.py Yes view/lighting
simplify_mixed.py Yes view
specmf.py Yes minor pixel diffs
symm.py Yes view/lighting
time_cmfe.py Yes view/lighting
time_deriv_cmfe.py Yes view/lighting
time_iteration.py Yes view/lighting
timelock.py Yes view/lighting
val4mat.py Yes view


Name Simple Rebaseline Notes
surface.py Yes very small rendering changes, slight view change, line style not implemented




The images for any <database>.py test python file not explicitly listed here is ok to entirely rebaseline all images from that file.

In a lot of cases in 2D plots especially, I noticed we dropped lines at the extreme edges of the plot. I think that is ok in some cases and NOT ok in other cases. I think if our orig. baselines had a different set of extreme edge lines missing and that situation just changed, that is probably ok. I think if the orig. baselines had all extreme edge lines correct and the new images are dropping some, that is a problem.

Name Notes
lata.py lata_03 seeing through mesh plot to zone face or face lighting is different, lata_07 missing >half of vector glyph tail, lata_12 odd shading
patran.py patran4_0 lighting or normals changed, other cases here too
pdbdatabase.py pdb_multi_12 lighting or normals, pdb_multi_15 font size change in label plot
vtk.py vtk_01 good example of minor line change, vtk_16 view change, vtk_23 LQDS (large quad diagonal swap), vtk_37 shading or normals
silo.py silo_20 shading artifacts at domain boundaries or back faces showing through, silo_23 major change in y axis label position (other cases too), silo_91 funky shading issue along edges of large elements
samrai.py samrai_copper_matvf_02 great example of LQDS and bad choice, samrai_24 domain boundary line color thrashing, samrai_34 really bad and no ideas why
pixie.py pixie_07 really bad! distance between objects?
bov.py label plot font size change
chgar.py chgar_02 LQDS other cases volume plots, label font size
shapefile.py is there some line density control in VTK-8 that we're defaulting to
h5part.py h5part_03 shading appears better, h5part_04 I can't tell which is better
blueprint.py blueprint_2d_json_uniform_braid LQDS, blueprint_2d_hdf5_points_braid point glyph size change, blueprint_3d_hdf5_points_braid point glyph size change , blueprint_mfem_periodic_cube_json_sol LQDS
vista.py do we still need this test
reopen.py reopen_3_04 better (older version had air gap between domain boundary and domain pc plot)
overlink.py overlink_03 & overlink_04 (probably other cases) shading or surface nomals issue
sami.py sami_05 shading or surface normal issue
cgns.py CGNS_0_04 shading issue in single bright pixel
netcdf.py netcdf_1_00 and other cases lost smooth shading of facets to flat shading
openfoam.py OpenFOAM_lagrangian_01 and other cases change in shading or light direction or diffuse vs. specular, OpenFOAM_Decomposed_01 improved and fixed a bad artifact
PLOT3D think ok but something funky happened with test result html generation
exodus.py exodus_22 WTF? No ideas
xmdf.py Xmdv_02 and other cases too label plot font size change
metadata.py metadata_1_expr_grad_tensor_diagonal label font size change
mfem.py input_mesh_disc-nurbs_mres_0000 improved, input_mesh_escher_mres_0000 single point artifacts, input_mesh_klein-donut_mres_0000 artifacts also missing a lot of results here
exportdb.py export_db_3_00 shading artifacts at domain boundaries
global_node_ids global_node_ids10 shading change
protiendb.py proteindb_0_01 label plot font size change
tecplot.py tecplot_16 LQDS



Name Simple Rebaseline Notes
hohlraum.py Yes lines are slightly off
length.py Yes lines are slightly off
avg_value.py Yes text and lines are slightly off
scf.py Yes text and lines are slightly off
conncomp.py Yes slight difference in the pixels
curvature.py Yes slight difference in the pixels
consistencyChecks.py Yes lines are slightly off. When only lines, % difference is ~50%. Could be difference line drawing algorithm
pickarray.py No There are new lines being drawn on these plots that do not exist in the baseline. I'm not sure about this
pick.py Maybe There are numerical differences in the output. If the view changed, then a different zone/node may have been picked. There are also big differences in the psuedocolor plot
queriesOverTime.py No most are minor differences in line drawing. AllTimeQueries_02, however is very different. This is a guess, but I think the new image is correct and the old image is an error.
line_scan.py Yes these are missing portions of lines
bestfitline.py Maybe seems like some are missing baseline images. Other than that, this looks like a rebaseline
pickNamedArgs.py Yes minor line diffs
scalabe2.py Maybe minor line diffs, but lots of lines. I think its ok but would like another set of eyes.



Name Simple Rebaseline Notes
ic_pics.py (skipped)
ex_surf.py Yes diffs look like view change
transform.py No view clip issue with one curve plot ops_transform08
lcs_ic_2.py Yes diffs look like view change
lcs.py No pseudocolor interp / rendering diffs
displace.py No pseudocolor interp / rendering diffs, see ops_disp02
elevate.py Yes diffs look like lighting changes (new images look better)
clip.py No clip fast looks different (could be better, but we need to look at it) clip_globe_fast
threeslice.py Yes diffs look like view or lighting change
onionpeel.py No diffs look like view or lighting change

labels look better, but we should review bigsil_mesh_01

amr_stitch.py Yes diffs look like view or lighting change
ic_integration.py (skipped)
slice.py No most diffs look like view or lighting change, but may be some clipping changes

Some may be better:


Some may be worse:

ops_Slice37 ops_Slice39

tube.py Yes diffs look like view or lighting change
ic_pathlines.py (skipped)
multires.py no diffs
ic_geometry.py (skipped)
lcs_operator.py (skipped)
persistent_particles.py No most diffs look like view or lighting change

points are smaller, which is probably fine -- but we should review as a team persistent_particles_03

indexselect.py Yes diffs look like view or lighting change
resample.py No interp differences operators_resample
inverse_ghost_zone.py No most diffs look like view or lighting change

missing lines: ops_inverse_ghost_zone05

lcs_lc_1.py Yes diffs look like view or lighting change
sph_resample.py Yes diffs look like view or lighting change
isovolume.py Yes diffs look like view or lighting change
defer_expr.py Yes diffs look like view or lighting change
cart_proj.py Yes diffs look like view or lighting change
reflect.py No most diffs look like view or lighting change

lighting diffs, probably ok but we should review as a team ops_refl06 ops_refl07 ops_refl08 ops_refl09

vol rendering diff ops_refl14

radial_resample.py Yes diffs look like view or lighting change
cracksclipper.py Yes diffs look like view or lighting change
box.py Yes diffs look like view or lighting change
threshold.py Yes diffs look like view or lighting change
explode.py Yes diffs look like view or lighting change
dual_mesh.py Yes diffs look like view or lighting change
lic_streamlines.py (skipped)
ic_termination.py Yes diffs look like view or lighting change
moveoperators.py Yes diffs look like view or lighting change


Name Simple Rebaseline Notes
arbpoly.py Yes view/lighting
csg.py Yes view/lighting
degen_mesh.py Yes view/lighting
emptydomains.py No Seems like view/lighting, but there appears to be highlighting along back edges of non-mesh plots:

emptydomains_02 emptydomains_03 emptydomains_04 emptydomains_07 emptydomains_10

globe.py No Mostly view/lighting, except Volume Rendering globe_04
point.py Yes view/lighting
rect3d.py No need another opinion, is this just view/lighting/opacity?