Upstream Spyne crashes on POST requests to the HTTP/RPC+JSON webservice.
We patched it in a rather hacky way.

This was a regression from the old codebase, where we installed Spyne
separately from our LUMC GitHub mirror. This is now also referenced in
the requirements.txt file.

Thanks to Ken Doig for reporting the issue.

Rename this webservice method. Note the capital letter L in the old
name. Also add a short note to the documentation that data arguments
must be base64 encoded.

Closes #11

Get the current announcement from Redis on each request, instead of
reading it once at module load.

Closes #12

Last remaining relevant todo notes have been filed as issues in GitLab.

Previously, Mutalyzer would after writing any file check the cache size
and start removing files while it exceeded the maximum. However, this
caused long delays in case many files had to be removed (it would re-
calculate the total size after each removal).

Following the principle of separating concerns, this is now handled by
a separate script on our production servers, which uses the inotifywait
tool to cleanup the cache whenever files are added to it. It also
doesn't suffer from the performance problem.

Note that this removes the `MAX_CACHE_SIZE` configuration setting.

Fixes #18

Using fetchChromSizes [1] and selecting *Download the full sequence report*
from the NCBI assembly overview [2] we can generate a mapping from UCSC
chromosome names to accession numbers:

    ./fetchChromSizes hg38 > human.hg38.genome
    for contig in $(cut -f 1 human.hg38.genome | grep 'alt$'); do
        code=$(echo $contig | cut -d _ -f 2 | sed 's/v/./')
        echo -n $contig$'\t'
        grep $code GCF_000001405.26.assembly.txt | cut -f 7
    done > alt_chrom_names.mapping

Generate the JSON dictionary entries:

    >>> import json
    >>> entries = []
    >>> for line in open('alt_chrom_names.mapping'):
    ...     chr, acc = line.strip().split()
    ...     entries.append({'organelle': 'nucleus',
    ...                     'name': chr,
    ...                     'accession': acc})
    ...
    >>> print json.dumps(entries, indent=2)
    [
      {
        "organelle": "nucleus",
        "name": "chr12_KI270837v1_alt",
        "accession": "NT_187588.1"
      },
      {
        "organelle": "nucleus",
        "name": "chr13_KI270842v1_alt",
        "accession": "NT_187596.1"
      },
      ...
    ]

[1] http://hgdownload.cse.ucsc.edu/admin/exe/linux.x86_64/fetchChromSizes
[2] ftp://ftp.ncbi.nlm.nih.gov/genomes/ASSEMBLY_REPORTS/All/GCF_000001405.26.assembly.txt

See http://pytest.org/

Based on comments by Jeroen in issue #14.

See http://semver.org/

Because we don't know how to name them anyway. We also ignore genes
without any transcripts.

The LRG model allows for more than one CDS per transcript (and each CDS
can be coupled with more than one transcript). The Mutalyzer gene model
does not allow that.

Previously, multiple CDSs per transcript where treated incorrectly (I
think), so now we just ignore any CDS except the first per transcript.

Unfortunately, it is a bit more involved to give good warnings for this,
so that is not implemented.