run issue about deepconsensus

[pichuan]

Thank you for reporting this.
To be honest we have not tested on CentOS 7. I just tested and also wasn't able to get it to work, but the error I'm getting is different:

Traceback (most recent call last):
  File "/usr/lib64/python3.6/runpy.py", line 183, in _run_module_as_main
    mod_name, mod_spec, code = _get_module_details(mod_name, _Error)
  File "/usr/lib64/python3.6/runpy.py", line 109, in _get_module_details
    __import__(pkg_name)
  File "/home/pichuan/.local/lib/python3.6/site-packages/deepconsensus/__init__.py", line 35, in <module>
    from nucleus.io import bed
  File "/home/pichuan/.local/lib/python3.6/site-packages/nucleus/io/bed.py", line 62, in <module>
    from nucleus.io import genomics_reader
  File "/home/pichuan/.local/lib/python3.6/site-packages/nucleus/io/genomics_reader.py", line 68, in <module>
    from nucleus.io.python import tfrecord_reader
ImportError: libbz2.so.1.0: cannot open shared object file: No such file or directory

I also just try testing nucleus 0.5.8, which I suspect also didn't work on CentOS7.

We'll have to think about what to do with this. I'll keep this open for now.

[MartinPippel]

Thanks for your reply. I also got your problem. And after installing bzip2 in my conda environment this error was gone.

[pichuan]

Hi @MartinPippel ,
I can confirm that I was able to get around the libbz2 issue, by doing something a bit hacky like:

sudo ln -sf /usr/lib64/libbz2.so.1.0.6 /usr/lib64/libbz2.so.1.0

After that, I'm also seeing: "`GLIBC_2.23' not found".

Actually, this reminded me that this is probably part of Nucleus - I tested Nucleus 0.5.8 (which DeepConsensus builds on), and we have the same issue. CentOS7 has an older GLIBC, and that's why we're having the issue.

I think it is possible for us to build Nucleus and DeepConsensus in a way that is manylinux2010 compliant, and hopefully resolve this issue. I will add this to our list of tasks, but I don't know for sure when we can prioritize this.

Another approach we could try to take is to use Docker (more like what we did for DeepVariant) instead of pip wheels.

Either approach will take some time for us. @MartinPippel , given that I've filed an internal issue to track this, I'll close this for now. But please feel free to add more here.

[maximilianpress]

I am experiencing the same issue on CentOS7, Python 3.6.9, conda etc. All of the notable details appear to be the same (including bzip2 sub-issue).

[hzhang-2021]

I think this is because the google-nucleus binary by pip was built under /lib64/libm.so.6: version 'GLIBC_2.23'.
the default centos7 only have /lib64/libm.so.6: version 'GLIBC_2.17'.
you have to recompile all that package from source.... that is very difficult job.