Problem in saving Pipeline

[shenlongsanti]

Dear Experts,

I am running CPAC on a Ubuntu 16.04, and when I try to save a pipeline I got the following error:

Traceback (most recent call last): File "/usr/local/lib/python2.7/dist-packages/CPAC/GUI/interface/windows/config_window.py", line 802, in submit_item defaultFile=("pipelineconfig%s" % pipelinename), wildcard="YAML files(.yaml, *.yml)|.yaml;.yml", style=wx.SAVE) File "/usr/lib/python2.7/dist-packages/wx-3.0-gtk2/wx/_windows.py", line 3139, in init windows.FileDialog_swiginit(self,windows.new_FileDialog(_args, **kwargs)) wx._core.PyAssertionError: C++ assertion "volDummy.empty() && pathDummy.empty()" failed at ../src/common/filename.cpp(568) in Assign(): the file name shouldn't contain the path

Does anybody know what is going wrong?

Thanks in advance, Santiago

[jpellman]

Hi Santiago,

Did you enter a path under the "Pipeline Name" field of the "Output Settings" tab in the pipeline configuration builder? If so, that could be producing this error- you should only enter a descriptive name here rather than specify a path to where you want to save the pipeline config YAML.

[shenlongsanti]

Thank you very much, Jpellman! That was really the issue.

Now, I configured the pipeline to run ALFF/fALFF and ReHo. Just as an observation, I am getting the following warning: "MatplotlibDeprecationWarning: The set_color_cycle attribute was deprecated in version 1.5. Use set_prop_cycle instead." I imagine that this is probably not important; I suppose this is only a aesthetic issue. Nevertheless, it may be good to use the suggestion.

The only error I got up to now is, which is concerning me a little bit is: Error: 'reho_to_standard_smooth_fisher_zstd' Should I be worried, or is it just saying that ReHo results won't be saved in z-scored values?

Thank you again, Santiago

[jpellman]

Hi Santiago,

I'm not certain what that line alone could mean without further context. Is there anything else in the error message that is more specific? Are there crash files being produced in the crash directory for this step? If so, could you forward them and a copy of your pipeline configuration file to me at john.pellman@childmind.org so that I can take a closer look at what might be going on?

Best, John Pellman

[rfcasseb]

Hi John and Santiago,

I am having the same error when trying to run ReHo: "reho_to_standard_smooth_fisher_zstd"

Could you solve it? Thanks in advance

[jpellman]

Hi @shenlongsanti and @rfcasseb ,

I spoke with one of our developers recently and she informed me that this is due to an issue that C-PAC 0.3.9.2 has with the newest version of nipype. Prior to mid-July of this year, a function in nipype output a 1D time series file. C-PAC 0.3.9.2 still expects this 1D file, but the latest version of nipype outputs a csv file instead. We have fixed this in the upcoming release of C-PAC (0.4.0), but for the current version you will need to downgrade nipype for this functionality to work properly. You can accomplish this with the following commands:

pip uninstall nipype pip install nipype==0.11

Let me know if that helps or if you continue to encounter this issue.

[rfcasseb]

Thank you very much! I will try it here.

With regards, Raphael

[rfcasseb]

It did not work... Do you have an idea of when version 0.4.0 will be released for linux?

[jpellman]

Hi @rfcasseb ,

We're hoping to have the release ready within the next couple of weeks. In the meantime, would you be willing to double-check that the Python downgrade went as expected by opening a Python terminal and typing the following commands?

import nipype print nipype.version

Also, is there any possibility that you're using the Neurodebian-based nipype, or another non-pip version that wouldn't respond to the 'pip uninstall' command?

[rfcasseb]

Dear John,

I had other things in my way, but now I am back trying to make C-PAC run smoothly,

1) So I did as you said, but there is no "version" attribute for nipype (you suggested doing nipype.version). But I reinstalled nipype and in the end of the installation it says it is version 0.11.0. And I have a pip responsive version, because I uninstalled and installed python using pip command lines.

However I still get the ReHo error at the beginning of the processing: Error: 'reho_to_standard_smooth_fisher_zstd'

2) And now it is not even finishing to run. The following are the last lines shown in my terminal: file_: qc_scan_Func_ASF_RS_wm1.motion1.csf1_CSF_0.96_GM_0.7_WM_0.96_bandpass_freqs_0.01.0.1___fwhm_6_hp_0.01_lp_0.1.html params_file: /media/nf05/D4E47912E478F7D6/Teste1_FSL/Results/pipeline_pipeline/ASF/power_params/_scan_Func_ASF_RS/_threshold_0.2/pow_params.txt ['excel-tab', 'excel'] 0.1364 27.1909

When I look the other lines, there are thousands of lines with the following message "VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future"

3) Do you suggest anything else that I could do in order to fix this? I am a little familiar with python language, so if you need me to try something else on the code, just let me know.

4) And in our previous messages that a new version would be coming out in the following weeks. Is it about to come out?

5) My last issue: I tried running CPAC on another machine which is Ubuntu 14.04 to see if CPAC would work, but I get the following error:

Error compiling Cython file:

... cdef err_cookie set_error_handler(err_cookie handler):

Note: exceptions here will be printed instead of raised.

cdef err_cookie old_handler

if H5Eget_auto(&old_handler.func, &old_handler.data) < 0:
                                 ^

---

Do you know why do I get that?

Sorry for the huge message. Thank you very much for your help, Raphael

2016-09-07 18:18 GMT-03:00 John Pellman notifications@github.com:

Hi @rfcasseb https://github.com/rfcasseb ,

We're hoping to have the release ready within the next couple of weeks. In the meantime, would you be willing to double-check that the Python downgrade went as expect by opening a Python terminal and typing the following commands?

import nipype print nipype.version

Also, is there any possibility that you're using the Neurodebian-based nipype, or another non-pip version that wouldn't respond to the 'pip uninstall' command?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/FCP-INDI/C-PAC/issues/539#issuecomment-245422641, or mute the thread https://github.com/notifications/unsubscribe-auth/ARPgg7lk5N199x6uBDxVvBezLUQtg2Qbks5qnypCgaJpZM4JFYQ7 .

Raphael F. Casseb Medical Physicist, Ph.D. Student Medical Physics Lab - University of Campinas Contact: +55 19 3521-8246

[jpellman]

Hi @rfcasseb ,

1) To clarify, there is not "version" attribute, but there should be a "version" attribute, which is where version numbers for Python packages are kept by convention. This attribute exists for other packages as well, and if you ever need to determine the version of a package in your environment, importing it and then running "print .version" will give you the info you need.

2/3) It looks like C-PAC is stopping at the QC report generation step. I'm not certain why it would do this without knowing more details. Could you send the file /media/nf05/D4E47912E478F7D6/Teste1_FSL/Results/pipeline_pipeline/ASF/power_params/_scan_Func_ASF_RS/_threshold_0.2/pow_params.txt to john.pellman@childmind.org so that I can take a look at it? Also, are there any files in your crash directory being generated at this step?

The VisibleDeprecationWarning reflects a change in numpy, but that should not affect the end results of the computation. It mostly serves to inform programmers to revise their code to reflect changes to numpy's code.

4) We recently released a new version of C-PAC. Release notes for it are available https://github.com/FCP-INDI/C-PAC/releases/tag/v1.0.0.

5) Please see the C-PAC forum post here for more information about this error.

[sgiavasis]

Closing this issue because it has been awhile. If there are any more issues or questions, please feel free to open a new one.

Best, The C-PAC Team