1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
|
Bowtie 2 tests
==============
Many of the tests use the debug versions of the binaries, so first do `make allall` in the root.
#### Simple test cases
A collection of small tests with hand-crafted input and output checks are collected in `scripts/test/simple_tests.pl`, with `scripts/test/simple_tests.sh` as a driver script. Uses Perl module `DNA.pm`.
From root:
sh scripts/test/simple_tests.sh
#### Regression tests
They use the Python3 infrastructure in the `scripts/test` subdirectory, including `bt2face.py`, `dataface.py` and `btdata.py`.
From root:
python3 scripts/test/regressions.py --verbose
Val Antonescu originally set these up.
#### Big index test
Builds an index consisting of both human and mouse genomes, pushing the genome size above the 2^32 limit, and necessitating a "big" 64-bit index. This takes a lot of time and RAM.
From root:
python3 scripts/test/large_idx.py --verbose
#### Randomized tests
`scripts/sim` contains Perl infrastructure for running randomized tests. These scripts generate random reference genomes and reads, and Bowtie 2 programs are run in various combinations with cross-checks to ensure, for example, that release-binary output matches debug-binary output and that single-threaded output matches multithreaded output. The non-core `Math::Random` Perl module is needed. You can use CPAN to install it.
As of 10/1/2016, the unit tests in `AlignmentCheck.pm` don't all succeed. Also, random tests fail for reasons that seem related to the test infrastructure, e.g., not specifying the command-line arguments properly.
To run the unit tests for the infrastructure, from `scripts/sim`:
sh unit.sh
To run the actual suite of random tests in parallel on 8 cores, from the root:
sh scripts/sim/run.sh 8
|
