RF: Delegate isotropic smoothing to nistats #118

kfinc · 2019-02-27T19:47:34Z

This PR delegates smoothing to Nistats. After making the changes we came into multiprocessing error. Any idea why this may be happening? Thanks in advance for any help!

Expand to see stack trace

	 [Node] Running "_l1_model7" ("fitlins.interfaces.nistats.FirstLevelModel")
190227-19:32:39,265 nipype.workflow INFO:
	 [MultiProc] Running 4 tasks, and 4 jobs ready. Free memory (GB): 13.19/13.99, Free processors: 0/4.
                     Currently running:
                       * _l1_model7
                       * _l1_model6
                       * _l1_model5
                       * _l1_model4
exception calling callback for <Future at 0x7f1dcf8fec18 state=finished raised BrokenProcessPool>
Traceback (most recent call last):
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/concurrent/futures/_base.py", line 324, in _invoke_callbacks
    callback(self)
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/site-packages/nipype/pipeline/plugins/multiproc.py", line 151, in _async_callback
    result = args.result()
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/concurrent/futures/_base.py", line 425, in result
    return self.__get_result()
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/concurrent/futures/_base.py", line 384, in __get_result
    raise self._exception
concurrent.futures.process.BrokenProcessPool: A process in the process pool was terminated abruptly while the future was running or pending.
exception calling callback for <Future at 0x7f1dcf052a58 state=finished raised BrokenProcessPool>
Traceback (most recent call last):
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/concurrent/futures/_base.py", line 324, in _invoke_callbacks
    callback(self)
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/site-packages/nipype/pipeline/plugins/multiproc.py", line 151, in _async_callback
    result = args.result()
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/concurrent/futures/_base.py", line 425, in result
    return self.__get_result()
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/concurrent/futures/_base.py", line 384, in __get_result
    raise self._exception
concurrent.futures.process.BrokenProcessPool: A process in the process pool was terminated abruptly while the future was running or pending.
exception calling callback for <Future at 0x7f1dcf0b0c18 state=finished raised BrokenProcessPool>
Traceback (most recent call last):
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/concurrent/futures/_base.py", line 324, in _invoke_callbacks
    callback(self)
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/site-packages/nipype/pipeline/plugins/multiproc.py", line 151, in _async_callback
    result = args.result()
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/concurrent/futures/_base.py", line 425, in result
    return self.__get_result()
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/concurrent/futures/_base.py", line 384, in __get_result
    raise self._exception
concurrent.futures.process.BrokenProcessPool: A process in the process pool was terminated abruptly while the future was running or pending.
exception calling callback for <Future at 0x7f1dcf060550 state=finished raised BrokenProcessPool>
Traceback (most recent call last):
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/concurrent/futures/_base.py", line 324, in _invoke_callbacks
    callback(self)
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/site-packages/nipype/pipeline/plugins/multiproc.py", line 151, in _async_callback
    result = args.result()
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/concurrent/futures/_base.py", line 425, in result
    return self.__get_result()
  File "/opt/miniconda-latest/envs/neuro/lib/python3.6/concurrent/futures/_base.py", line 384, in __get_result
    raise self._exception
concurrent.futures.process.BrokenProcessPool: A process in the process pool was terminated abruptly while the future was running or pending.

Sketch implementation of smoothing kernel in nistats

oesteban · 2019-02-27T19:51:49Z

You are running out of memory. Try not to parallelize 4 models at a time or increase the RAM available to the docker engine.

use 0 to denote that contrast does not include parameter

subst na in place

fix smoothing params and arg

effigies

Thanks for this. A few comments. Also, I would recommend modifying the Dockerfile by removing this section of .neurodocker.json, and then re-running neurodocker:

fitlins/.neurodocker.json

Lines 21 to 26 in bebc4c7

    
           [ 
        
             "fsl", 
        
             { 
        
               "version": "5.0.11" 
        
             } 
        
           ],

cat .neurodocker.json | \
    docker run --rm -i -a stdin -a stdout -a stderr kaczmarj/neurodocker:master generate docker - > \
    Dockerfile

I can do this and push to your branch, if that's easier.

fitlins/interfaces/nistats.py

effigies · 2019-02-28T16:49:04Z

fitlins/interfaces/visualizations.py

@@ -94,6 +94,7 @@ def _visualize(self, data, out_name):
        contrast_matrix = pd.DataFrame({c['name']: c['weights'][0]
                                        for c in self.inputs.contrast_info},
                                       index=data.columns)
+        contrast_matrix.fillna(value=0, inplace=True)


Hmm. I'll need to look into this. nan should be fine for plotting matrices, and will simply show up as white, rather than the intermediate gray. But this seems fine, if nan is no longer possible.

fitlins/workflows/base.py

Co-Authored-By: kfinc <karolinafinc@gmail.com>

@effigies

update to reflect @effigies review

effigies · 2019-03-01T21:55:47Z

LGTM. Is this still a WIP?

kfinc · 2019-03-02T08:26:24Z

Everything works fine, so I think it's no longer WIP. Thanks everyone for help!

kfinc and others added 4 commits February 27, 2019 01:13

Removed FSL installation

1576cd5

Removed FSL smoothing

82b7f7b

sketch implementation of smoothing kernel in nistats

54bba83

Merge pull request #1 from rciric/nistats_smoothing

2f95cfb

Sketch implementation of smoothing kernel in nistats

rciric and others added 8 commits February 27, 2019 16:39

use 0 to denote that contrast does not include parameter

a4a847a

Merge pull request #2 from rciric/nistats_smoothing

f7f95a6

use 0 to denote that contrast does not include parameter

subst na in place

b527027

Merge pull request #3 from rciric/nistats_smoothing

1a38ebf

subst na in place

fix smoothing params and arg

caa3728

Merge pull request #4 from rciric/nistats_smoothing

ad5a91b

fix smoothing params and arg

fix smoothing params and arg

ad35584

Merge pull request #5 from rciric/nistats_smoothing

b732974

fix smoothing params and arg

effigies reviewed Feb 28, 2019

View reviewed changes

effigies and others added 7 commits February 28, 2019 22:38

Update fitlins/interfaces/nistats.py

5ed6a7d

Co-Authored-By: kfinc <karolinafinc@gmail.com>

removed FSL section

ebbdd42

new file generated with neurodocker

586e991

changed traits.Either() to traits.Float()

4cc33ac

added smoothing_fwhm attribution and check

d453b1a

update to reflect @effigies review

8fab580

Merge pull request #6 from rciric/nistats_smoothing

3ecaf63

update to reflect @effigies review

effigies changed the title ~~[WIP] Nistats smoothing~~ Delegate isotropic smoothing to nistats Mar 2, 2019

effigies changed the title ~~Delegate isotropic smoothing to nistats~~ RF: Delegate isotropic smoothing to nistats Mar 2, 2019

effigies merged commit 620f795 into poldracklab:master Mar 2, 2019

effigies mentioned this pull request Mar 18, 2019

Delegating the optional smoothing to Nistats #116

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

RF: Delegate isotropic smoothing to nistats #118

RF: Delegate isotropic smoothing to nistats #118

kfinc commented Feb 27, 2019 •

edited by rciric

oesteban commented Feb 27, 2019

effigies left a comment

effigies Feb 28, 2019

effigies commented Mar 1, 2019

kfinc commented Mar 2, 2019

RF: Delegate isotropic smoothing to nistats #118

RF: Delegate isotropic smoothing to nistats #118

Conversation

kfinc commented Feb 27, 2019 • edited by rciric

oesteban commented Feb 27, 2019

effigies left a comment

Choose a reason for hiding this comment

effigies Feb 28, 2019

Choose a reason for hiding this comment

effigies commented Mar 1, 2019

kfinc commented Mar 2, 2019

kfinc commented Feb 27, 2019 •

edited by rciric