Vidjil Team authored Mar 08, 2019 and Mathieu Giraud committed Mar 08, 2019

Following passionate discussions (!245) :
 - add many long option names, using updated terms
 - deprecate 11 short options, including many rare options coming form ancient times
 - reorder options and bikeshed some help texts

See #3295.

Ensure tab after name

See merge request !430

On each line we should have a + or - when segmented or an exclamation mark
when unsegmented.
A tab should always exist at least between the sequence name and the VDJ designation.

This should ease parsing the output.

-c designations

Closes #3772

See merge request !429

Better CLI11 error messages, link to documentation

See merge request !428

see also should#31

Add display the error message improved in the previous commit.

see #3775

close #3772.

define and use failure_message_doc(), taking inspiration from FailureMessage::simple()

see #3775

See #3772.

See #3772.

Feature cs/3722 browsers from env or file

See merge request !423

This cannot be properly tested in test_segmenter_page_errors.rb as we don't have a menu there

Mathieu Giraud authored Feb 28, 2019 and Mikaël Salson committed Mar 05, 2019

See #3722 and #3701.

Mathieu Giraud authored Feb 28, 2019 and Mikaël Salson committed Mar 05, 2019

tests/launch_functional_tests: select browsers from environment variable or from 'tested-browser' file

Partially revert 02ca0a24.
We want both to agnostically use 'launch_functional_tests', but also to be able to follow in a versioned file
the supported browsers.

Use a fixed version of web2py

Closes #3605

See merge request !419

Update e-value on SEG_METHOD_ONE with Aho

Closes #3773

See merge request !426

was done twice in 40aef950 and 556d8b0b, but git merge
did not understood what we meant there ;-)

see #3773

Random read sample for representative

Closes #3764

See merge request !422

The dataset was generated to have a minority of sequences with a TTT insertion
which would make it the sequence of choice with the default ReadScore (based
on length and quality).

On the contrary with a random sampler we should not be impacted by this
insertion as it happens in the minority of sequences.

Here is the command that generated the dataset:
for i in $(seq 1 3500); do
  if [ $((RANDOM%3)) -ne 0 ]; then
    echo ">seq$i";
    echo ctacctactactgtgccttgtgggaggtgatagtagtgattggatcaag;
  else
    echo ">seq$i";
    echo cTTTtacctactactgtgccttgtgggaggtgatagtagtgattggatcaag;
  fi;
done > test-random-consensus.fa

A reference cannot be redefined in C++.

Mathieu Giraud authored Mar 05, 2019 and Mikaël Salson committed Mar 05, 2019

see #3764.

This was put in two different places and could lead to inconsistencies. Now the
reads are necessarily selected in the same way.

Random score on a read makes possible to have a random sample of reads to compute the representative.
Thus the representative will be built on a really representative sample.

Our attempts to have a good quality sample have failed so far. Thus we attempt this solution.

Fix #3764

Aho by default

Closes #3419 and #2556

See merge request !78

Exit → return

See merge request !425

Executing valgrind can take more time, thus we need to increase the default timeout to make
sure the program has time to properly execute.

See 1a7bb7f8 for reason why.

Fix (in part?) #3771

Feature a/3729 clean memory

Closes #3729

See merge request !412