You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Error with matching_distance function in iPython example #4
Do you know what might have caused the following error when I try to replicate what is shown in your ipython example with the matching_distance function? I managed to get the same results on all other parts of the notebook. I am using Linux and a Python 3.7 under Conda environment. Thanks!
Error in file /tmp/tmpahk85_rb/self.txt, line number 15: cannot parse "--"
ERROR:root:Failure in invocation of Hera, input files copied to error-hera-yellow-31885-1553538069.0262306 for reference
Traceback (most recent call last):
File "/home/milad/anaconda3/envs/myenvpy37/lib/python3.7/site-packages/pyrivet/hera.py", line 86, in multi_bottleneck_distance
["bottleneck_dist", t1_name, t2_name, str(relative_error)])
File "/home/milad/anaconda3/envs/myenvpy37/lib/python3.7/subprocess.py", line 395, in check_output
**kwargs).stdout
File "/home/milad/anaconda3/envs/myenvpy37/lib/python3.7/subprocess.py", line 487, in run
output=stdout, stderr=stderr)
subprocess.CalledProcessError: Command '['bottleneck_dist', '/tmp/tmpahk85_rb/self.txt', '/tmp/tmpahk85_rb/other.txt', '1e-10']' returned non-zero exit status 1.
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home/milad/anaconda3/envs/myenvpy37/lib/python3.7/site-packages/pyrivet/matching_distance.py", line 99, in matching_distance
[bars for (_, bars) in multi_bars2]
File "/home/milad/anaconda3/envs/myenvpy37/lib/python3.7/site-packages/pyrivet/hera.py", line 86, in multi_bottleneck_distance
["bottleneck_dist", t1_name, t2_name, str(relative_error)])
File "/home/milad/anaconda3/envs/myenvpy37/lib/python3.7/subprocess.py", line 395, in check_output
**kwargs).stdout
File "/home/milad/anaconda3/envs/myenvpy37/lib/python3.7/subprocess.py", line 487, in run
output=stdout, stderr=stderr)
subprocess.CalledProcessError: Command '['bottleneck_dist', '/tmp/tmpahk85_rb/self.txt', '/tmp/tmpahk85_rb/other.txt', '1e-10']' returned non-zero exit status 1.
Error in file /tmp/tmpahk85_rb/self.txt, line number 15: cannot parse "--"
My first guess would be that you installed Hera either not from the xoltar fork, or else you did use the xoltar fork, but didn't check out the c-api branch before building.
The original Hera code only supports processing a single pair of barcodes, but for efficiency we want to compare many pairs in a single invocation. These multiple pairs are separated by the "--" separator. The c-api branch in the xoltar fork of Hera contains the code that knows how to process these.
Please let me know if this does or doesn't solve the problem.
I actually used your forked version of hera with the c_api.
My guess would be that I am doing something wrong with the cmake when I am building the bottleneck_dist. What I am doing is simply uncommenting the two commented lines in the CMakeList.txt and building the app bottlleneck_dist and the library libbottleneck.a. I thought there might be something there I am missing. I also tried to reinstall the python-rivet in case something went wrong there but that didn't help as of yet.
Activity
xoltar commentedon Mar 25, 2019
Based on this part of the error message:
My first guess would be that you installed Hera either not from the xoltar fork, or else you did use the xoltar fork, but didn't check out the
c-apibranch before building.The original Hera code only supports processing a single pair of barcodes, but for efficiency we want to compare many pairs in a single invocation. These multiple pairs are separated by the "--" separator. The
c-apibranch in the xoltar fork of Hera contains the code that knows how to process these.Please let me know if this does or doesn't solve the problem.
miladkiaee commentedon Mar 26, 2019
I actually used your forked version of hera with the c_api.
My guess would be that I am doing something wrong with the
cmakewhen I am building the bottleneck_dist. What I am doing is simply uncommenting the two commented lines in theCMakeList.txtand building the appbottlleneck_distand the librarylibbottleneck.a. I thought there might be something there I am missing. I also tried to reinstall the python-rivet in case something went wrong there but that didn't help as of yet.