Ready for blog post release which will be at
http://blog.dhimmel.com/citation-styles/

Refs http://www.imagemagick.org/Usage/layers/

Redownloaded PMC OA Subset on 2018-03-05. Now downloads
oa_file_list.csv.

Update utils to strip some whitespace in certain XML fields.

Rerun pipeline.

Refs #1
Misnamed .nxml files listed in data/misnamed-nxml-files.tsv

Use Git LFS to track *.gz files

Refs #1

[ci skip]

See the error from
https://gitlab.com/dhimmel/pmc-citation-styles/-/jobs/55240672
CommandNotFoundError: Your shell has not been properly configured to use 'conda activate'.

Refs https://github.com/conda/conda/issues/6811
See https://gitlab.com/dhimmel/pmc-citation-styles/-/jobs/55219405

See https://beenje.github.io/blog/posts/gitlab-ci-and-conda/

Currently failing

Silence bs4 warnings

Extract citation text from <xref> children. Previousely, the
following would return None, when cite.text was called:

<xref rid="bib9" ref-type="bibr"><sup>9</sup></xref>

Although, previousely the following would get parsed due to bs4:

<xref rid="bib9" ref-type="bibr">[<sup>9</sup>]</xref>

Now both should work.

Also processes individual lxml.etree articles

Download was initiated on 2018-02-26 but the FTP modified
timestamps show the .xml.tar.gz files are from 2018-02-24.

Create pmc-articles-xml.zip with SHA-256 checksum of
28bf1f8d2bffe1d1bd9371716a40df45dd51c6cb9644a46da24af720f3f47061