GitLab

missing comma in control file dependencies

germlines and data files are no longer packaged with the Vidjil Browser

Added automatically built VERSION files to all packages

Dissociated the browser from the server package and added browser
dependency on server package

Fixed te issue with build-algo and build-server where copied files were
not removed from the packaging folder once the build had ended

Added a script whichs clears duplicate sequence_file-config pairs from
the results_file table. The most recent entry containing said pair is
kept. Scheduler tasks associated to the deleted entries are also
cleared.

Added execution user to systemd web2py scheduler service

With the new vidjil packaging, the vidjil executable may not be situated
in the vidjil directory. Therefore it has become necessary to add a
separate DIR_GERMLINE variable to task.py which needs to be added to
defs.py

Resolved an issue with the packaging configuration due to some nasty
copy/pasting from a script which introduced some unwanted backslashes

Added a paragraph to indicate the existance of the automated package build
scripts

Added automation scripts for building the vidjil server Debian package

Added automation scripts for building the vidjil algorithm debian
package

Added scripts to build a package containing the germline directory and
the Makefile which allows the user to download the IMGT germlines

All files inside refer to curated-vdj, not should-vdj

It passes with Vidjil now. IMGT agrees.

BUG → TODO

Should we keep the two first sequences as there is one very short D (3nt) each
time?

Only the two last sequences were checked by myself.

Some preferences should not be reset when switching between patients.

It is better to keep the same germline selected between patients, as
well as the same color methods and the same top (to compare patients more
easily). Some preferences should also be kept for convenience (notation
and time format).

Due to the previous commit we know display identity rate rather than mutation rate.
The number of unit digits will likely differ. Therefore we align the numbers to the right.

We also slightly increase the padding to the right because there are more digits.

Following previous commit, we now display the identity rate rather
than the mutation rate. Therefore variables and CSS classes should
have a sound name.

This commit is just about renaming.

Biologists are more used to talk about identity rate rather than error
rate. We therefore adapt to this terminology.

Fix 6077ae

When several calls are made to IMGT, we must append the results to imgt2display
otherwise they would be overwritten.

We also remove an unused (and not so exact) value @width_nameBox

* The options can be overriden usig the option override_imgt_options.
* The results are overwritten in m.clones[].seg.imgt
* We now launch IMGT Post twice: 1. With nt-sequence 2. With summary (% id in V gene)

With malformed files, Fuse will raise an exception that must be caught here.