GitLab

Previously the file was overwritten for each call to the function (which was a pity with VDJ recombinations).

Few people care about alleles and we should focus (first) on genes.

This is stored in 'samples.db_key'.

This result file is now stored in db.results_file, like the usual results files.

When a window is given through the 'grep_reads', Vidjil is launched with -FaW <window>,
and fuse.py is not called.

This could be useful for other tasks as well.

This link enables to download the result file of a successful task/run.
This is now always a .vidjil file, but, who knows ?

Thanks to a question by Tim Xiaoming Hu.

The purpose of these tests is to check whether the reads with labeled windows
are kept, and not to test the VDJ segmentation on these difficult reads.

Don't remove anything: the sequences used in segmentation.fasta
are quite specific to test tricky case to see how behaves the
affectanalyser in strange cases. Such tricky cases will be hard to find
only with the last 100nt of V genes

Obviously the number of kmers is lower. But the number of ambiguous k-mer is also lower.

This is a hand-made test with the *start* of a V gene. Therefore it cannot be found
with default parameter where we only keep the last 100bp

It will be default in a next release.

This option is passed to the germline which is passed to the kmerstore (in the insert method).
It is used to only consider one end of a V/J gene

Used to be pass through to the insert() method in KmerStore (to avoid indexing all the Vs or all the Js)

With keep_only parameter we can index either the start or the end of a sequence.
Useful since we may not need to index all the V genes (but only the end).

Discussions in British pubs are valuable!

(merged by magiraud, 2015-05-07, light conflict with 07adde3c)

Fix 9aa52e68.
Thanks to the functional Watir tests.

Partially revert 59f3e19d
Tests should work even without a server on localhost (especially for unit tests run by Jenkins or Travis).

We changed several times this line to see if the clone is the same.
The purpose of the test is to check if we have the same clone in both datasets.
The purpose of the test is not to check the segmentation.

To check that, we remove the part of the lines that differ among both datasets
and only keep the segmentation information. Hence both lines should be equal and
should not appear in the diff.