GitLab

heto authored Apr 04, 2017 and Mikaël Salson committed Apr 06, 2017

Typing errors, misspellings, etc.

See #2338.

If there need some space, we will do it through CSS
See #2338.

See #2338.

See #2338.

The sample_set/all page was redirecting to an inexistant page if not
logged in

The redirection after deleting a sample_set did not take the sample_type
into account.

call would append trailing question mark to a URL which did not have any
args. This was causing some issues with some redirects, namely for the
login page.

runs jshint on all javascript files in browser/js/

We need to update lines in list.js/segmenter.js.

(We still have #1763.)

This is necessary as process substitution is not available in sh.
However the command is portable (tested under Bash on FreeBSD)

In some cases we may not have access to /dev/stderr which causes an error.
This has been rewritten to redirect the output both to STDERR and STDIN. The
output which is on STDIN is still grep-ped to check the number of errors.

Deploy to $PROD_CLIENT_SERVER where the SSH config is as is:

Host app.vidjil.org
        StrictHostKeyChecking no
        UserKnownHostsFile=/dev/null
        User = bonsai-ci
        IdentityFile = ~/.ssh/bonsai-ci

See vdj#210

It also brings valuable information on nb of occurrences of the sequence

This is commented for now as it can be resource-consuming. We may want to
wait a bit before putting it in production for everyone.

SO that they can be displayed in the clone info page

Closes #2304.

See #2304.

Actually it is not a good idea that the client directly calls
the CloneDB. Indeed some tags from the CloneDB directly refers
to data on the server-side. Therefore the server can bring valuable
informations.

That's what we do know: the client calls a controller which calls
the CloneDB. The controller then adds other information useful to the client
(URL of the sample sets, name of the sample sets).

We would probably benefit from a class offering such functions.
The class would have a constructor relying only on the sample set ID
and would not need anything else.

The SampleSet class doesn't behaves in this fashion as this would not
be efficent enough (for the patient list for instance, see #2295).
However in some cases we would need a more flexible way of accessing
a sample set, even it is slower.

This would only display something unreadable for the users. So don't.

Closes #2300.

Two reasons:
1. It is cleaner
2. os.system fails when there are too many arguments

Importing vidjil-to-fasta.py is a bit tricky because of the dashes in its name.
So we use imp.load_source (Python 2.7+) which does the job.