GitLab

Cyprien Borée authored Jun 15, 2018 and Mikaël Salson committed Jun 26, 2018

Since the Germline now contains a Filter object instead of an automaton
and respectively Filter contains the automaton. Both of them have to
free the memory allocated correctly.

For more informations, see #3283.

Cyprien Borée authored Jun 15, 2018 and Mikaël Salson committed Jun 26, 2018

Since the function testBehaviourWhenHugeBioReader in testFilter was
using pair and not Filter object, it now use the filter object getter for
the pair.

For more informations, see #3283.

Cyprien Borée authored Jun 15, 2018 and Mikaël Salson committed Jun 26, 2018

Since the function testExAequoKmersWhenSignificantParameter in testFilter was
using pair and not Filter object, it now use the filter object getter for
the pair.

For more informations, see #3283.

Since the function testGetNSignificativeKmers in testFilter was
using pair and not Filter object, it now use the filter object getter for
the pair.

For more informations, see #3283.

Since the function testFilterBioReaderWithACAutomaton in testFilter was
using pair and not Filter object, it now use the filter object getter for
the pair.

For more informations see #3283.

Since the function testAutomatonBuilderFilteringBioReader in testFilter was
using pair and not Filter object, it now use the filter object getter for
the pair.

For more informations, see #3283.

Since the unit-tests were using the pair instead of a Filter object, a getter
has been created for the pair.

For more informations, see #3283.

Since the filter methods are now in the Filter object stored in the Germline,
the segment.cpp file needs to use the Germline's Filter object in order to
filter its BioReader.

For more informations see #3283.

The filter attribute has now a getter function. This way the filter attribute
can be used outside of the germline class.

For more informations see #3283.

The Germline class now contains a FilterWithACAutomaton pointer which is built
if the Germline constructor has TRUE for the boolean build_automaton.

For more informations see #3283.

Since the pair attribute is stored in the Filter class, there is no need
to put a pair object in the Filter methods. The methods were changed to use
the object attribute, instead of the parameter.

For more informations see #3283.

For more informations see #3283.

This constructor will directly call the buildACAutomaton method. Then
this method will construct the pair<Vector, Automaton> stored in the
FilterWithACAutomaton class.

For more informations see #3283.

For more informations, see #3283.

The methods filter and buildAutomaton were moved in the newly created
class FilterWithACAutomaton. The signature in filter.cpp were changed
accordingly.

For more informations see #3283.

In order to refactor the filter file into a class, the useless constants and
include were removed from the file.

For more informations about the refacotr see #3283.

With g++-8 we had a “warning: catching polymorphic type * by value”.
This is solved by using a reference

Touching it updated its last modification date which triggered the recompilation of vidjil.cpp.
We don't touch it anymore. Therefore the file may not exist and the diff may fail. We just ignore
error messages. In such a case the mv will be executed and the file will then exist.

Fixes #3309

Fix testGetMultiResults when there is only unknown K-mer

See merge request !222

Remove compiler warnings

See merge request !221

Since the map only contain one unknown k-mer it is useless to iterate over it
using a loop.
Also the tap_test method has been changed to the tap_test_equal because of the
equality comparison.
For more informations, see #3303.

see #3303 for more informations

The previous test was only looking for a single unknown K-mer in the map even
though the map shouldn't contain anything else from unknown K-mers.
Indeed when there is no K-mers in the automaton that match the input sequence,
then the map contains only unknown K-mers.
For more informations, see #3303.

The relation between _SYMBOL and _CHAR was not clear and could be
ambiguous. This should clarify a little.

See #3250.

Feature a/3250 shift kmer ascii value for ambiguous and unknown

Closes #3250

See merge request !220

For more informations see #3250.

For more informations see #3250.

Since these two K-mers need to be printed with a different char, this char
is defined as a new macro used exclusively to represent the K-mer in a string.

For more informations see #3250.

For more informations see #3250.

Since the methods returns the same thing, the getLabel instructions can simply
return the toStringValues function result.

For more informations see #3250.

the toStringValues for those two K-mers must be '_' and '?' respectively.
Even though their char macro is 1 and 0 (non printable characters).

For more informations see #3250.

To fit issue #3250 the macros values were changed.
Char n°0 for ambiguous and char n°1 for unknown K-mer.

The only are .should-locus tests.

Close #3294.

Close #3287.

We now have a public Gitlab CI status.
See #3287.

Loading HTTP URL from the file menu

See merge request !210

Fix #3015

Use the macro UNUSED for mandatory parameters that are not
used within functions.

See #1233