Unable to successfully run Slingshot

[matthieubulte]

Hi!

I'm having issues running the script-to-generate-datasets.sh, in particular the pseudo-time inference part.

More precisely, when running only the slingshot part of the script, the python script seems to fail after a couple of hours.

Here are the logs

$ python $path_to_boolode/scripts/runSlingshot.py     --expr $output_dir$output_name"/ExpressionData.csv"     --pseudo $output_dir$output_name"/PseudoTime.csv"     --outPrefix $output_dir$output_name     --nClusters=$numclusters -r 200
Computing TSNE...
docker run --rm -v ~/Documents/.../Synthetic/mCAD/:/data/temp slingshot:base /bin/sh -c "Rscript data/run_slingshot.R --input=/data/temp/rd.tsv --input-type=matrix --cluster-labels=/data/temp/cl.tsv --start-clus=0 --end-clus=1"
2019-10-02 17:40:52 INFO::Checking input arguments
Using full covariance matrix
2019-10-03 04:48:51.595 python[35108:4214084] -[NSApplication _setup:]: unrecognized selector sent to instance 0x7fb4da457a70
2019-10-03 04:48:51.601 python[35108:4214084] *** Terminating app due to uncaught exception 'NSInvalidArgumentException', reason: '-[NSApplication _setup:]: unrecognized selector sent to instance 0x7fb4da457a70'
*** First throw call stack:
(
	0   CoreFoundation                      0x00007fff3d52d2fd __exceptionPreprocess + 256
	1   libobjc.A.dylib                     0x00007fff67c01a17 objc_exception_throw + 48
	2   CoreFoundation                      0x00007fff3d5a7106 -[NSObject(NSObject) __retain_OA] + 0
	3   CoreFoundation                      0x00007fff3d4cf18f ___forwarding___ + 1485
	4   CoreFoundation                      0x00007fff3d4ceb38 _CF_forwarding_prep_0 + 120
	5   libtk8.6.dylib                      0x0000001a22c6431d TkpInit + 413
	6   libtk8.6.dylib                      0x0000001a22bbc17e Initialize + 2622
	7   _tkinter.cpython-37m-darwin.so      0x0000001a229e4a0f _tkinter_create + 1183
	8   python                              0x000000010cbf6446 _PyMethodDef_RawFastCallKeywords + 230
	9   python                              0x000000010cd32ce2 call_function + 306
	10  python                              0x000000010cd2fa83 _PyEval_EvalFrameDefault + 42243
	11  python                              0x000000010cd2434a _PyEval_EvalCodeWithName + 410
	12  python                              0x000000010cbf50d7 _PyFunction_FastCallDict + 231
	13  python                              0x000000010cc773d1 slot_tp_init + 193
	14  python                              0x000000010cc814d1 type_call + 241
	15  python                              0x000000010cbf5e03 _PyObject_FastCallKeywords + 179
	16  python                              0x000000010cd32d75 call_function + 453
	17  python                              0x000000010cd30a7e _PyEval_EvalFrameDefault + 46334
	18  python                              0x000000010cbf5be5 function_code_fastcall + 117
	19  python                              0x000000010cd32c67 call_function + 183
	20  python                              0x000000010cd2fa83 _PyEval_EvalFrameDefault + 42243
	21  python                              0x000000010cd2434a _PyEval_EvalCodeWithName + 410
	22  python                              0x000000010cbf50d7 _PyFunction_FastCallDict + 231
	23  python                              0x000000010cbf9002 method_call + 130
	24  python                              0x000000010cbf6a82 PyObject_Call + 130
	25  python                              0x000000010cd30bf2 _PyEval_EvalFrameDefault + 46706
	26  python                              0x000000010cd2434a _PyEval_EvalCodeWithName + 410
	27  python                              0x000000010cbf50d7 _PyFunction_FastCallDict + 231
	28  python                              0x000000010cd30bf2 _PyEval_EvalFrameDefault + 46706
	29  python                              0x000000010cd2434a _PyEval_EvalCodeWithName + 410
	30  python                              0x000000010cbf6313 _PyFunction_FastCallKeywords + 195
	31  python                              0x000000010cd32c67 call_function + 183
	32  python                              0x000000010cd30a7e _PyEval_EvalFrameDefault + 46334
	33  python                              0x000000010cd2434a _PyEval_EvalCodeWithName + 410
	34  python                              0x000000010cbf6313 _PyFunction_FastCallKeywords + 195
	35  python                              0x000000010cd32c67 call_function + 183
	36  python                              0x000000010cd309c5 _PyEval_EvalFrameDefault + 46149
	37  python                              0x000000010cbf5be5 function_code_fastcall + 117
	38  python                              0x000000010cd32c67 call_function + 183
	39  python                              0x000000010cd309c5 _PyEval_EvalFrameDefault + 46149
	40  python                              0x000000010cd2434a _PyEval_EvalCodeWithName + 410
	41  python                              0x000000010cd87b80 PyRun_FileExFlags + 256
	42  python                              0x000000010cd86ff7 PyRun_SimpleFileExFlags + 391
	43  python                              0x000000010cdb4ccf pymain_main + 9711
	44  python                              0x000000010cbc88dd main + 125
	45  libdyld.dylib                       0x00007fff693cf3d5 start + 1
	46  ???                                 0x000000000000000b 0x0 + 11
)
libc++abi.dylib: terminating with uncaught exception of type NSException
Abort trap: 6

I'm running this on MacOS 10.14.6 with Docker 2.1.0.3.

Please let me know if there is any other information that could be useful to you.

Thanks!

[adyprat]

Hi @matthieubulte,

Thank you for bringing this to our attention. So far I've only tested the runSlingshot.py code on Linux and I have not encountered this error before. Could you confirm if you can see a curves.csv file in your output folder under 'Synthetic/mCAD/'? If so, that would mean the docker part of the code has run successfully.

As far as I can tell the NSException error is specific to Mac OS. If I understand the issue correctly, it is probably the plotting function here that's causing the issue. Can you try the solution posted here? Let me know if that fixes the issue.

Best,
Aditya

[matthieubulte]

Hi @adyprat,

Yes, the curves.csv file seems to have been generated. I will try re-running the script during the night with the modification suggested on stackoverflow.

If it works, I'll prepare a PR to add this fix and give the user the option to skip the plotting part. Would it make sense?

EDIT: thanks for your answer!

[adyprat]

Sure thing! We plotted the Slingshot output by default for all instances to visualize the computed pseudotime while testing the code. I suppose giving a user option for plotting makes sense.

Thank you,
Aditya