don't use volume for database in tests

Closes vdj#1034

See merge request !637

docker-compose down -v removes all that without the need of doing a docker ps
(and potentially removing more containers than needed)

The volume is classically mounted on the host which is not desirable when testing.

See vdj#1034

It can be a problem while testing as the application is launched just after the docker-compose up

Otherwise the string contains the quotes…

Fix vdj#1034

Different V/D/J/ seeds

Closes #3954

See merge request !634

See #3954

Fix #3954

Resolve "Indication sur le graphe quand on affiche uniquement les clones d'un sample"

Closes #4221

See merge request !629

Missing affectations

Closes #4225

See merge request !633

Since another MR, now demo data have date and are set by default in graph
Link to #4221

Merge branch 'feature-c/4221-indication-graph-one-sample' of gitlab.inria.fr:vidjil/vidjil into feature-c/4221-indication-graph-one-sample

Link to #4221

Link to #4221 and !564

Merge branch 'feature-c/4105-ameliorer-l-affichage-de-beaucoup-de-samples-avec-le-graphe-actuel-graph-list' into 'dev'

Resolve "Améliorer l'affichage de beaucoup de samples avec le graphe actuel, graph-list"

Closes #4105

See merge request !565

In Aho-Corasick automaton we know the affect once we traversed the sequence.
However for backward-compatibility reasons with k-mer indexes we need to move
those affects so that they correspond to the start of the sequence.
However for an Unknown Affect, it is hard to tell where we should put it
as the affect is unknown. If we see that we will overwrite an existing
affect, we try another position.

This solves some cases in #4225

Resolve "Verrou sort: quand on change de point, -"

Closes #4222

See merge request !630

Utiliser fuse pour faire une sortie AIRR

Closes #4199

See merge request !610

This is an arbitrary choice but when building the final automaton we may
add information from previous states in the current state (so information from
shorter seeds). We prefer to keep the longest one.

When transferring information we may overwrite them. We need to take into
account the case where the destination state already has some information stored.

That wasn't a bug (as a null info has a 0-length). But I prefer preventing
an overwriting of an actual result with an empty result.

When finalising the automaton we can define new final states that
actually are a “copy” of another final state. In such a case we
also need to copy the Info that are in that former final state.

Added a test to check that.
The test uses a IGKJ sequence: thus all the kmers should be +k
or (potentially undetermined).

Fix #4225

Link to #4105

Link to #4222

Merge branch 'dev' of gitlab.inria.fr:vidjil/vidjil into feature-c/4105-ameliorer-l-affichage-de-beaucoup-de-samples-avec-le-graphe-actuel-graph-list

Merge branch 'feature-t/4199-utilise-fuse-pour-faire-une-sortie-airr' of https://gitlab.inria.fr/vidjil/vidjil into feature-t/fuse_export_as_airr

Link to #4199