Closed smason closed 2 years ago
Would it be worth registering these tests to ctest?
For the various tools, lets alias '-h' to '--help'
Text for overlap_checker help
./overlap_checker --help
Usage: overlap_checker [OPTION...] input.brep
Perform imprinting of BREP shapes.
--bbox-clearance=C Bounding-boxes closer than C[=0.5] will be checked
for overlaps
--imprint-tolerance=T Faces, edges, and verticies will be merged when
closer than T[=0.001]
-j, --jobs=N Parallelise over N[=4] threads
--max-common-volume-ratio=R
Imprinted volume with ratio <R[=0.01] is
considered acceptable
-?, --help Give this help list
--usage Give a short usage message
Mandatory or optional arguments to long options are also mandatory or optional
for any corresponding short options.
Says that its for imprinting stuff, which it isn't
overlap_checker - default seems to be debug being on, should this not be off by default?
No useful summary from overlap checker, I would expect something like a list of overlaps, or a message to say there weren't any
No useful summary from overlap checker
what would a useful summary look like? it did follow the old school unix paradigm of no output indicates success, at least it did until I broke the code that sets logging levels
it'll output a line if there are problems, are you saying it should always output a line, listing things like number of comparisons done, number of shapes touching, with small overlaps, etc?
re. the other comments, I'll see what I can do this afternoon
I guess I would expect a summary table as a default option, with some useful stats, like number of overlaps, number of overlap checks, number of coincidences and so on?
No useful summary from overlap checker, I would expect something like a list of overlaps, or a message to say there weren't any
you could always tee
the CSV output to watch progress
that said, I've added a summary line in 074b8c1ae540efce1940227fd77dee15097a5feb that outputs:
202.961 [Info] processing summary: bbox tests=525825, intersection tests=3404, touching=158, overlapping=1, bad overlaps=0, tests failed=0
@makeclean think that's all done now
please let me know if I've missed anything!
The end of run summary is great!
Im seeing a bunch of things like
204.528 [Warn] 435-129 imprint failed with (73 filler and 0 common) warnings
204.528 [Error] 435-129 failed to classify overlap
What is a filler warning?
I'm not sure if you want to branch some of the following of here into issues, or just fix em all here?
I'm noticing a tool step_to_brep - I presume I should run this on STEP files first, would it not be easier (better?) to have filename based dispactch in the overlap and imprint tools to either convert a step file to brep then run?
Hmm, I actually quite like the the one tool one job, unix pipeline maybe type dispatch would be overkill. Though some additonal verbositry form the input tool would be good, I've currently no idea what its doing, or how long it might take?
What is a filler warning?
it's coming from the deep guts of opencascade, I'm reporting it because it was easy to pull out and in the hope it'd be useful for debugging if/when something like this fails
maybe open another issue with a pointer to the file an I can have a look?
the main reason for separating things out is to help with materials later, I've added some detail as part of another merge: https://github.com/ukaea/overlap_checker/tree/shape-merge-tool#file-formats
i.e. step_to_brep
flattens the file into a list of shapes, then all other tools keep shapes the "same", when I start handling materials it'll hopefully be an easy job to match things back up afterwards
basically drop everything I can that isn't in the STL, use CMake for building, and catch2 for unit tests