Closed monicabritton closed 4 years ago
Hi, I cannot produce the error (which is probably not surprising...). I don't really have a good suggestion besides trying a clean install :(
--Liz
I wiped everything (conda, git repos, etc.) and re-installed from scratch. During the installation there were some warnings about "Using deprecated NumPy API". However, now the unittest fails, right after another Deprecation Warning (see below). I'm guessing I have mis-matched versions of something, but am not sure what. Any suggestions?
python setup.py test running test Searching for bx-python==0.7.3 Best match: bx-python 0.7.3 Processing bx_python-0.7.3-py2.7-linux-x86_64.egg
Using /afs/genomecenter.ucdavis.edu/home/mbritton/.conda/envs/anaCogent/lib/python2.7/site-packages/bx_python-0.7.3-py2.7-linux-x86_64.egg
running egg_info
writing requirements to Cogent.egg-info/requires.txt
writing Cogent.egg-info/PKG-INFO
writing top-level names to Cogent.egg-info/top_level.txt
writing dependency_links to Cogent.egg-info/dependency_links.txt
reading manifest file 'Cogent.egg-info/SOURCES.txt'
writing manifest file 'Cogent.egg-info/SOURCES.txt'
running build_ext
test_cogent (Cogent.test.test_cogent.TestCogent) ... 2.0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Sketching human_test.fa.0...
Writing to human_test.fa.0.s1000k30.msh...
human_test.fa.0.s1000k30.msh.human_test.fa.0.s1000k30.msh.dist completed
Output written to: /tmp/tmpII4mah/human_test.fa.s1000k30.dist
making weight graph from /tmp/tmpII4mah/human_test.fa.s1000k30.dist
graph contains 19 nodes, 79 edges
performing ncut on graph....
(subgraph) has 10 nodes. ncut down to 1 partitions in 0.174357891083 sec.
(subgraph) has 9 nodes. ncut down to 1 partitions in 0.00503993034363 sec.
0 ['PB.58.29', 'PB.58.28', 'PB.58.24', 'PB.58.34', 'PB.58.14', 'PB.58.8', 'PB.58.4', 'PB.58.5', 'PB.58.3', 'PB.58.1']
2 ['PB.246.8', 'PB.246.2', 'PB.246.41', 'PB.246.6', 'PB.246.4', 'PB.246.5', 'PB.246.15', 'PB.246.17', 'PB.246.18']
2.12-r827
/afs/genomecenter.ucdavis.edu/home/mbritton/.conda/envs/anaCogent/lib/python2.7/site-packages/Bio/Seq.py:407: BiopythonDeprecationWarning: This method is obsolete; please use str(my_seq) instead of my_seq.tostring().
BiopythonDeprecationWarning)
Traceback (most recent call last):
File "/home/mbritton/.conda/envs/anaCogent/bin/reconstruct_contig.py", line 4, in
Traceback (most recent call last): File "/share/biocore/mbritton/software/Cogent/Cogent/test/test_cogent.py", line 43, in test_cogent subprocess.check_call("reconstruct_contig.py .", shell=True) File "/home/mbritton/.conda/envs/anaCogent/lib/python2.7/subprocess.py", line 190, in check_call raise CalledProcessError(retcode, cmd) CalledProcessError: Command 'reconstruct_contig.py .' returned non-zero exit status 1
Ran 1 test in 5.825s
FAILED (errors=1)
Test failed:
Hi @monicabritton ,
What version of networkx are you using? The one I have in my conda environment is 2.1.
Can you try to downgrade/upgrade networkx to 2.1 and see if it still fails?
$ conda install networkx=2.1
>>> import networkx
>>> networkx.__version__
'2.1'
--Liz
Hi Liz:
The default networkx on our system is 2.0. But, when I first tried to install Cogent, one of the Traceback error messages included:
pkg_resources.ContextualVersionConflict: (networkx 2.0 (/afs/ genomecenter.ucdavis.edu/software/anaconda2/5.0.1/lssc0-linux/lib/python2.7/site-packages), Requirement.parse('networkx==1.10'), set(['Cogent']))
That seemed to indicate networkx 1.10 was a requirement, so I installed it, and made sure that my PYTHONPATH had the location ahead of the default one.
Does that pkg_resources.ContextualVersionConflict error mean something other than a requirement for networkx 1.10?
Thanks,
Monica
Monica Britton Sr. Bioinformatics Analyst Genome Center and Bioinformatics Core Facility University of California, Davis mtbritton@ucdavis.edu
On Wed, Dec 12, 2018 at 10:25 AM Elizabeth Tseng notifications@github.com wrote:
Hi @monicabritton https://github.com/monicabritton ,
What version of networkx are you using? The one I have in my conda environment is 2.1.
Can you try to downgrade/upgrade networkx to 2.1 and see if it still fails?
$ conda install networkx=2.1
import networkx networkx.version '2.1'
--Liz
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/Magdoll/Cogent/issues/59#issuecomment-446691832, or mute the thread https://github.com/notifications/unsubscribe-auth/AC2pMeZGudcgO-lNcCukJuK_nD7yib_7ks5u4UougaJpZM4ZJZst .
If you are using Conda environment, everything should be controlled/installed via conda. Using PYTHONPATH to point to other versions may cause issues like the one seen here.
Interestingly it looks like I have two networkx in Conda, probably because networkx 1.10 was installed as requirement when I ran Cogent's setup.py
.
Using conda list -n <my_env_name
I saw this:
networkx 1.10 <pip>
networkx 2.1 py27_0
I would recommend wiping that PYTHONPATH clean. Maybe remove all networkx instances from conda using conda uninstall
and installing networkx again.
I installed Cogent, dependencies, etc. When I run the unittest and test script, they run to completion ("OK" at the end), but I do get this warning several times:
/home/mbritton/.conda/envs/anaCogent/lib/python2.7/site-packages/Bio/Seq.py:407: BiopythonDeprecationWarning: This method is obsolete; please use str(my_seq) instead of my_seq.tostring(). BiopythonDeprecationWarning)
When I run the tutorial (test_human), the first step (run_mash.py) runs fine and produces non-empty output files, but the second step (process_kmer_to_graph.py) runs a few seconds, then ends with a Segmentation fault. There are no other error messages.
FYI, my $PYTHONPATH is /home/mbritton/.local/lib/python2.7/site-packages:/share/biocore/mbritton/software/Cogent/Complete-Striped-Smith-Waterman-Library/src:/software/anaconda2/5.0.1/lssc0-linux/lib/python2.7/site-packages
I get the segfault with my own isoseq hq.fasta too. Any suggestions?
Thanks!