GitLab

The second part of the test now verify that previous sequences are still
in the filtered BioReader even though a new ex-aequo has been found.

For more informations see #3284

This reverts commit 53b44e2c.

In the second part of the test, the program verify that previous filtered
sequences are still in the BioReader when there is one more ex-aequo found.
For more informations about it, see #3279.

Cyprien Borée authored Jun 14, 2018 and Mikaël Salson committed Jun 15, 2018

For more informations see issue #3279

Cyprien Borée authored Jun 14, 2018 and Mikaël Salson committed Jun 15, 2018

Since the last way to compare produced segfault (due to undefined behaviour
of the advance method), to bypass this problem we just look at the previous
of occurences to see if it's the same as the last -1.
For more informations see #3279.

Cyprien Borée authored Jun 14, 2018 and Mikaël Salson committed Jun 15, 2018

In order to be sure that the filter function returns the accurates sequences,
every sequences is checked and also the size of them.

This reverts commit 82c5ff15.

Those files are not used anymore and were there for tests purposes. This
commit should not have been merged.

This is not the optimal way.
See #3285.

optimize FineSegmenter, filtering BioReader with an Aho-Corasick automaton

Closes #920 and #3271

See merge request !199

Mathieu Giraud authored Jun 13, 2018 and Mikaël Salson committed Jun 13, 2018

Opens #3279.

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

Since kmer_threshold parameter became a signed int, there's the possibility
that this parameter is a negative number. If it's the case the entire original
BioReader is returned.

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

For more informations see #3268

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

Since a new parameter has been added to filterBioReader (representing the
filter specifity) it has to be transmitted when calling the function. Otherwise
the filter will do the filter on "NO_LIMIT_VALUE".

For more informations see #3268

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

Since a new parameter has been added, even the germline constructor call
in the test need to use it to use the automaton.

For more informations see #3268.

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

A boolean parameter has been added in the Germline constructors in order to
check if the automaton needs to be constructed or not.
For example with -Z all, the automaton is not constructed.

For more informations see #3268

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

The FineSegmenter constructor handles the filtering of the germline. By default
the value is "NO_LIMIT_VALUE", which means the filtering isn't active.

For more informations about it #3268.

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

In #3268 we want to add a specific command-line option for the filtering.
The -Z has only been declared in the getopt. The remaining task is to
check if -Z still has NO_LIMIT_VALUE as an argument. If so the automaton
is not built, otherwise the automaton is built and the argument is used to
filter the BioReader.

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

If the program is call with -Z 0, the filter function use every K-mer with
at least one match. 0 is defined by the constand ALL_KMERS_VALUE in tools.h
For more informations see #3276

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

This constant is representing -Z 0 parameter. It's used when the filter function
is called on every k-mer with at least one match.

For more informations see #3276

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

Since in the future there might be another automaton (automaton_3), the
pair_automaton variable has been changed to automaton_5 in many files.

see #3267 and #3267 (comment 96152)

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

Since the kmer_threshold parameter became a signed integer, some comparison
were producing sign warnings due to -Wsign-compare g++ compiler.

For more informations see #3270

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

Since the threshold parameter can be -1 the threshold parameter has been switched
to a signed integer (the default int in C/C++).
Also the use of NO_LIMIT_VALUE is used insted of -1.

For more informations see #3270

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

Since the new option -Z will be added, the case where all K-mers will be selected
to filter will be represented by -Z -1.
For more informations about it, see #3268 and #3264.

Cyprien Borée authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

For the purpose of adding the -Z option (described in #3264), the -H option
will show its description.

Mathieu Giraud authored Jun 12, 2018 and Mikaël Salson committed Jun 13, 2018

Incidentally close #3271.

Cyprien Borée authored Jun 08, 2018 and Mikaël Salson committed Jun 13, 2018

Sequences are equal before 110bp so the alternative is valid.

For more informations see #3262

Cyprien Borée authored Jun 07, 2018 and Mikaël Salson committed Jun 13, 2018

Since some sequences are identical, we can add them as a result for
the test.

For more informations about it, see #3223 and #3262.

Cyprien Borée authored Jun 07, 2018 and Mikaël Salson committed Jun 13, 2018

Since some sequences are identical, we can add them as a result for
the test.

For more informations about it, see #3223 and #3262.

Cyprien Borée authored Jun 05, 2018 and Mikaël Salson committed Jun 13, 2018

When there is no sequences found in the filtered BioReader it throws a
segfault.

For more informations about it, see #3223