list_test.js : new values after norm, due to new size of normalised clone (TRG smaller, half size than before)

There is a button for normalizing external data and a button for normalizing clones.
An ID is attributed to both of them. This fixes the browser functional testing

Mathieu Giraud authored Jul 02, 2015 and Vidjil Team committed Jul 09, 2015

We have several "'NoneType' object has no attribute 'status'" tickets on the server.
We should investigate why info.html has sometimes no run to display.

Mathieu Giraud authored Jul 02, 2015 and Vidjil Team committed Jul 09, 2015

Log a specific error message when
 - 'db.sequence_file[request.vars["id"]]' is None
 - 'data_file' is not stored

We have several errors on the server for these two points, we should investigate.

Fix e9864cd4. The previous affectation was not cloning the system_selected array,
giving strange things at the next change of system_selected.
This bug surfaced only recently: a52634a2 uses toggle_system(true).

Mikaël Salson authored Jul 08, 2015 and Vidjil Team committed Jul 09, 2015

The function could not be properly tested because unit test cannot (in an easy way?)
work on files stored by the DB.
The get_analysis function has been split in several functions.

Mikaël Salson authored Jul 08, 2015 and Vidjil Team committed Jul 09, 2015

Fix b069b161

We check that the score of the V/J regions is at least the one of 10 matches.
This range of commits further prepares the code for an actual p-value computation.

As the FineSegmenter does not accept too short 5/3 regions, the 'trd-dd2-dd3.fa' test
becomes a '.should-locus' test: It is not tested for the FineSegmenter.
It is replaced by this new test, that includes Dd2 upstream and Dd3 downstream regions.

multi-tiny.fa is a shortened version of multi-short.fa,
with only 10 bp at the end of V and at the start of J.

This test checks the representative and should not be changed by the output of the FineSegmenter.
We thus set '-z 0'.

Waiting for a true p-value computation, we just require here that
the score is at least the one of 10 matches.

As the FineSegmenter include a (now fake) e-value computation, these e-values are on stdout.
Two tests need to be updated.