This address is set with the new `BATCH_NOTIFICATION_EMAIL` config
setting and defaults to the value of `EMAIL`.

Caching of transcript protein links received from the NCBI Entrez
service is a typical use case for Redis. This implements this cache
in Redis and removes all use of our original database table.

An Alembic migration copies all existing links from the database to
Redis. The original `TranscriptProteinLink` database table is not
dropped. This will be done in a future migration to ensure running
processes don't error and to provide a rollback scenario.

We also remove the expiration of links (originally defaulting to 30
days), since we don't expect them to ever change. Negative links
(caching a 'not found' result from Entrez) *are* still expiring,
but with a longer default of 30 days (was 5 days).

The configuration setting for the latter was renamed, yielding the
following changes in the default configuration settings.

Removed default settings:

    # Expiration time for transcript<->protein links from the NCBI (in seconds).
    PROTEIN_LINK_EXPIRATION = 60 * 60 * 24 * 30

    # Expiration time for negative transcript<->protein links from the NCBI (in
    # seconds).
    NEGATIVE_PROTEIN_LINK_EXPIRATION = 60 * 60 * 24 * 5

Added default setting:

    # Cache expiration time for negative transcript<->protein links from the NCBI
    # (in seconds).
    NEGATIVE_LINK_CACHE_EXPIRATION = 60 * 60 * 24 * 30

Previously transcript-protein links were assumed to always be
indexed by transcript, and cached entries were allowed to have
a `null` protein (meaning caching the knowledget that there is
no link for this transcript).

Now we can cache links in both directions. Both transcript and
protein are allowed to be `null` (but not at the same time),
and the protein column has a new unique constraint.

The Werkzeug reloader is disabled by default due to a bug with
using it in combination with `python -m mutalyzer.entrypoints.website`.

https://github.com/mitsuhiko/werkzeug/issues/461#issuecomment-139369694

By default, the first request to the SOAP service will trigger a
build of the WSDL document, using the context (service location)
from that request. For example, if the first request is on
`http://localhost/` and subsequent requests are on
`https://mutalyzer.nl/services/`, they will not have a valid WSDL
document.

This is actually what we do on our production infrastructure,
where the service is tested (on localhost) after it has been
started.

The fix is to force a build of the WSDL document and specifying
the location to use.

http://spyne.io/docs/2.10/reference/server.html#spyne.server.wsgi.WsgiApplication

Issue #50 showed a problem in our file encoding detection, caused
by our cut-off for the confidence as reported by the cchardet [1]
library:

    >>> import cchardet
    >>> s = u'NM_000052.4:c.2407\u20132A>G'
    >>> b = s.encode('WINDOWS-1252')
    >>> cchardet.detect(b)
    {'confidence': 0.5, 'encoding': u'WINDOWS-1252'}

We require a confidence stictly greater than 0.5 and default to
UTF8 otherwise.

If, however, we try the same thing using the chardet [2] library,
we get a higher confidence for the same string:

    >>> import chardet
    >>> chardet.detect(b)
    {'confidence': 0.73, 'encoding': 'windows-1252'}

So the two obvious ways to solve this are:

1. Lower the confidence threshold.
2. Use chardet instead of cchardet.

We implement the second solution here, since it also removes a C
library dependency and we are not worried by performance.

Of course the detected encoding remains a guess which can still
be wrong!

[1] https://github.com/PyYoshi/cChardet
[2] https://github.com/chardet/chardet

Fixes #50

This is a work in progress as there still seem to be some bugs. For example,
some unit tests fail due to incorrect descriptions generated and others fail
due to a crash.

We use a separate `doc/requirements.txt` file for building the
documentation. Since on Read The Docs where we cannot install
packages depending on C libraries, we mock these packages in
`doc/conf.py` and omit them from the requirements.

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

See http://pytest.org/

Based on comments by Jeroen in issue #14.