AttributeError: 'File' object has no attribute 'createGroup' error upon running genomedata-load-seq

EricR86 commented 7 years ago

Original report (archived issue) by Celia Siu (Bitbucket: celiasiu).

My end goal is to use Genomedata to convert my data into genomedata format for use in Segway, but ...

When I run genomedata-load-seq --verbose genomedata.test hg19.chr22.fa (as a test because using all the chromosomes did not work), I get an AttributeError i.e.

Implementation unspecified. Found 1chromosomes/scaffolds, so using: dir
hg19.chr22.fa
Traceback (most recent call last):
  File "/home/csiu/anaconda2/envs/segway/bin/genomedata-load-seq", line 6, in <module>
    sys.exit(genomedata._load_seq.main())
  File "/home/csiu/anaconda2/envs/segway/lib/python2.7/site-packages/genomedata-1.3.5-py2.7.egg/genomedata/_load_seq.py", line 319, in main
    mode=options.mode, seqfile_type=options.seqfile_type)
  File "/home/csiu/anaconda2/envs/segway/lib/python2.7/site-packages/genomedata-1.3.5-py2.7.egg/genomedata/_load_seq.py", line 261, in load_seq
    read_seq(chromosome, seq)
  File "/home/csiu/anaconda2/envs/segway/lib/python2.7/site-packages/genomedata-1.3.5-py2.7.egg/genomedata/_load_seq.py", line 133, in read_seq
    seq_unambig, *span)
  File "/home/csiu/anaconda2/envs/segway/lib/python2.7/site-packages/genomedata-1.3.5-py2.7.egg/genomedata/_load_seq.py", line 76, in create_supercontig
    supercontig = h5file.createGroup(where, name)
AttributeError: 'File' object has no attribute 'createGroup'

and I am unable to figure out why.

I am using a Linux/Centos6 x86_64 system

genomedata was installed by using conda install segway. A relevant list of conda packages & versions include:


python                    2.7.12                        1  
hdf5                      1.8.17                        6    conda-forge
numpy                     1.11.2                   py27_0  
genomedata                1.3.5                    py27_0    bioconda

gmtk 1.4.4 0 bioconda segway 1.4.1



- hg19.chr22.fa is produced by
    1. Downloading hg19.2bit from http://hgdownload.soe.ucsc.edu/goldenPath/hg19/bigZips/
    2. And running `./twoBitToFa -seq=chr22 hg19.2bit hg19.chr22.fa` to produce the fasta. (twoBitToFa was download from [UCSC utilities](http://hgdownload.soe.ucsc.edu/admin/exe/))
- I then ran `genomedata-load-seq --verbose genomedata.test hg19.chr22.fa` and got the above error
- I tried removing the "chr" prefix from the sequence name, but the same error persists

EricR86 commented 7 years ago

Original comment by Celia Siu (Bitbucket: celiasiu).

Edited issue description

EricR86 commented 7 years ago

Original comment by Eric Roberts (Bitbucket: ericr86, GitHub: ericr86).

Hi @celiasiu. This is currently fixed in the latest development version. A dependency, PyTables, unfortunately when updating to a new point release, removed the createGroup function.

A release is planned for Genomedata shortly. So you can either use the development version by executing python setup.py install from downloading the source from here. Or downgrade so a slightly older PyTables version.

Sorry for the brief inconvenience!

EricR86 commented 7 years ago

Original comment by Eric Roberts (Bitbucket: ericr86, GitHub: ericr86).

changed state from "new" to "resolved"

Already fixed

EricR86 commented 7 years ago

Original comment by Eric Roberts (Bitbucket: ericr86, GitHub: ericr86).

@celiasiu a newly released version of Genomedata just went up on pip. You can simply pip install --upgrade genomedata and it should fix your issues!

hoffmangroup / genomedata

AttributeError: 'File' object has no attribute 'createGroup' error upon running genomedata-load-seq #27