search

rsemeraro / PyPore

A python tool box for fast and accurate quality control, conversion and alignment of nanopore sequencing data
GNU General Public License v3.0
20 stars 5 forks source link

install error #1

Closed bitcometz closed 5 years ago

bitcometz commented 5 years ago

hello, I try to install PyPore without Root authority on our linux cluster and the following error message was encountered: [20/22] Cythonizing /tmp/easy_install-1Rxpzk/h5py-2.9.0/h5py/h5t.pyx [21/22] Cythonizing /tmp/easy_install-1Rxpzk/h5py-2.9.0/h5py/h5z.pyx [22/22] Cythonizing /tmp/easy_install-1Rxpzk/h5py-2.9.0/h5py/utils.pyx

/tmp/easy_install-1Rxpzk/h5py-2.9.0/h5py/api_compat.h:27:18: error：hdf5.h：no such file or dierectory /tmp/easy_install-1Rxpzk/h5py-2.9.0/h5py/defs.c:476:23: error：H5Cpublic.h：no such file or dierectory /tmp/easy_install-1Rxpzk/h5py-2.9.0/h5py/defs.c:477:21: error：hdf5_hl.h：no such file or dierectory ...

Thanks

Best

rsemeraro commented 5 years ago

Hello, is the hdf5 library installed on your machine? You can check this by typing h5cc -showconfig. You'll get a screen like this. Screen Shot 2019-04-19 at 09 11 07 If missing, you have to install the hdf5 library by means of your packages installer (sudo apt-get install libhdf5-dev). Alternatively, if you have a conda distribution, you can get it by typing conda install -c anaconda hdf5.

bitcometz commented 5 years ago

Thanks! Since I did not have the root authority. So I find another way to solve this problem. I used conda to install all the Dependencies. So when I run "python setup.py install" and I do not need to the do the installation again. Anyway, I run PyPore successfully on the "test_dataset" and PyPore generate the files "Test_pore_activity_map.html" and "Test_sequencing_summary.html".

However, I did not get the right results from the files generated by Guppy(Version 2.0.10) with PromethION. Usually, guppy generate the following files:

fast5_skip/ sequencing_summary/ fast5_fail/ fastq_fail/ fast5_pass/ fastq_pass/

So, I build the tmp directory "test_in" and link the original directory "fast5_fail" and "fast5_pass" to "fail" and "pass", just like those in the "test_dataset" and then I run PyPore: ... [PyPore][WARNING] - Channel 504 inactive! [PyPore][WARNING] - Channel 505 inactive! [PyPore][WARNING] - Channel 506 inactive! [PyPore][WARNING] - Channel 507 inactive! [PyPore][WARNING] - Channel 508 inactive! [PyPore][WARNING] - Channel 509 inactive! [PyPore][WARNING] - Channel 510 inactive! [PyPore][WARNING] - Channel 511 inactive! [PyPore][INFO] - Plotting results...

Traceback (most recent call last): File "dir/pypore/bin/pypore", line 11, in load_entry_point('PyPore==0.1', 'console_scripts', 'pypore')() File "dir/pypore/lib/python2.7/site-packages/PyPore-0.1-py2.7.egg/lib/pypore.py", line 61, in main run_subtool(mod, ags, parser) File "dir/envs/pypore/lib/python2.7/site-packages/PyPore-0.1-py2.7.egg/lib/pypore.py", line 16, in run_subtool sq.run(ags) File "dir/pypore/lib/python2.7/site-packages/PyPore-0.1-py2.7.egg/lib/seqstats.py", line 125, in run ArgsReader(args) File "dir/envs/pypore/lib/python2.7/site-packages/PyPore-0.1-py2.7.egg/lib/seqstats.py", line 52, in ArgsReader sq.run(ags) ...

It did generate the *_activity_map.html but all the channels are empty. Do PyPore support the data generated from Guppy? Or it required specific names format of the data?

Thanks

best

rsemeraro commented 5 years ago

I never worked with PromethION data. However, i suppose it produces a multi-read fast5 dataset. If i'm right, you can try to activate the multi-read flag via the --multi_read_fast5 yes command.