js/app.js: Load extra scripts depending on the administrator

Closes #3929

See merge request !481

Fix #3929

V-QUEST API change

Closes #3913

See merge request !472

Otherwise the ntseq is not given in the CSV file
See #3913

IMGT/V-QUEST now requires to have the common name of the species, not the latin one.
Thus we need a function to do the conversion.

See #3913

Those parameters are not useful anymore. They were used before #3913.

Some parameters were slightly different from what was expected.

See #3913

See #3913

See #3913

Check permissions before retrieving results

Closes vdj#880

See merge request !476

Fix vdj#880

Merge branch 'feature-c/3488-decalage-de-la-sequence-du-cdr3-et-de-la-jonction-dans-la-fenetre-info' into 'dev'

Resolve "Décalage de la séquence du CDR3 et de la jonction dans la fenêtre info"

Closes #3488

See merge request !471

Resolve "Avoir des germline de poulet"

Closes #3910

See merge request !470

link to #3910

link to #3910

Link to #3910

link to #3910

doc/vidjil-algo.md: productivity discordance + server

See merge request !475

See #2526.

link to #3488

Release 2019.05

See merge request !473

see #3914

see #3914

core/representative.h, vidjil.cpp: RandomScore by default

See merge request !464

It will depend on the order in which reads are coming when computing
the representative but this order is random. What is important is that
we find a representative of the correct length with the correct sequence
whatever the read name and read positions are.

It is hard to predict the result is not totally deterministic
(as is the case with the RandomScorer, which is now by default,
whose results depend on the pseudo-random generator that may
differ depending on the OSes)

See #3866

see #3866

see #3866

link to #3488

link to #3488

link to #3488

link to #3488

Feature a/3206 rename and merge should

Closes #3762

See merge request !467

Optimizations, including  e-value computations, max12

See merge request !468

Computing them for each read is time consuming.
But many of the are common (there are few different
index loads and the read length is small)

This prevents from recomputing it each time