Add BDD tests

[rsajulga-nmdp]

Adding BDD (Behavior-driven development) tests for

• existing unit tests

  • No MAC
  • MAC
  • GLStrings

• new development

  • "lgx" and "lg" should only return 2-fields #40  #46
  • G and P alleles  #47

[rsajulga-nmdp]

Any idea @pbashyal-nmdp, @mmaiers-nmdp, etc. why ard.redux_gl("HLA-A*01:AB", 'G') does not produce any results when I run it through behave (it works manually/locally)? For comparison, ard.redux("HLA-A*01:01:01", 'G') actually does produce the right result 'HLA-A*01:01:01G'.

[pbashyal-nmdp]

Maybe related to HLA- prefix bug that I fixed with PR #49

[mmaiers-nmdp]

There is definitely some odd behavior with regard to leading HLA-A

ard.redux_gl: A*01:AB            , G = A*01:01:01G/A*01:02
ard.redux:    A*01:01:01         , G = A*01:01:01G
ard.redux_gl: A*01:01:01         , G = A*01:01:01G
ard.redux:    HLA-A*01:AB        , G = HLA-
ard.redux_gl: HLA-A*01:AB        , G = 
ard.redux:    HLA-A*01:01:01     , G = HLA-A*01:01:01G
ard.redux_gl: HLA-A*01:01:01     , G = 
ard.redux:    DRB1*14:06:01      , G = DRB1*14:06:01
ard.redux_gl: DRB1*14:06:01      , G = DRB1*14:06:01
ard.redux:    DRB1*14:06:02      , G = DRB1*14:06:02
ard.redux_gl: DRB1*14:06:02      , G = DRB1*14:06:02
ard.redux:    A*01:AB            , lgx = A*01:AB
ard.redux_gl: A*01:AB            , lgx = A*01:01/A*01:02
ard.redux:    A*01:01:01         , lgx = A*01:01
ard.redux_gl: A*01:01:01         , lgx = A*01:01
ard.redux:    HLA-A*01:AB        , lgx = HLA-A*01:AB
ard.redux_gl: HLA-A*01:AB        , lgx = 
ard.redux:    HLA-A*01:01:01     , lgx = HLA-A*01:01
ard.redux_gl: HLA-A*01:01:01     , lgx = 
ard.redux:    DRB1*14:06:01      , lgx = DRB1*14:06
ard.redux_gl: DRB1*14:06:01      , lgx = DRB1*14:06
ard.redux:    DRB1*14:06:02      , lgx = DRB1*14:06
ard.redux_gl: DRB1*14:06:02      , lgx = DRB1*14:06

[mmaiers-nmdp]

Its better with 0.0.21

ard.redux:    A*01:AB            , G = 
ard.redux_gl: A*01:AB            , G = A*01:01:01G/A*01:02
ard.redux:    A*01:01:01         , G = A*01:01:01G
ard.redux_gl: A*01:01:01         , G = A*01:01:01G
ard.redux:    HLA-A*01:AB        , G = HLA-
ard.redux_gl: HLA-A*01:AB        , G = HLA-A*01:01:01G/HLA-A*01:02
ard.redux:    HLA-A*01:01:01     , G = HLA-A*01:01:01G
ard.redux_gl: HLA-A*01:01:01     , G = HLA-A*01:01:01G
ard.redux:    DRB1*14:06:01      , G = DRB1*14:06:01
ard.redux_gl: DRB1*14:06:01      , G = DRB1*14:06:01
ard.redux:    DRB1*14:06:02      , G = DRB1*14:06:02
ard.redux_gl: DRB1*14:06:02      , G = DRB1*14:06:02
ard.redux:    A*01:AB            , lgx = A*01:AB
ard.redux_gl: A*01:AB            , lgx = A*01:01/A*01:02
ard.redux:    A*01:01:01         , lgx = A*01:01
ard.redux_gl: A*01:01:01         , lgx = A*01:01
ard.redux:    HLA-A*01:AB        , lgx = HLA-A*01:AB
ard.redux_gl: HLA-A*01:AB        , lgx = HLA-A*01:01/HLA-A*01:02
ard.redux:    HLA-A*01:01:01     , lgx = HLA-A*01:01
ard.redux_gl: HLA-A*01:01:01     , lgx = HLA-A*01:01
ard.redux:    DRB1*14:06:01      , lgx = DRB1*14:06
ard.redux_gl: DRB1*14:06:01      , lgx = DRB1*14:06
ard.redux:    DRB1*14:06:02      , lgx = DRB1*14:06
ard.redux_gl: DRB1*14:06:02      , lgx = DRB1*14:06

[pbashyal-nmdp]

ard.redux only works on alleles, not on allele codes. Since A*01:AB is not a valid allele, it returns ''
But you can pass in remove_invalid=False to return the original.

>>> ard = ARD('3290', data_dir='/tmp/3290', remove_invalid=False)
>>> ard.redux("HLA-A*01:AB", "G")
'HLA-A*01:AB'

Yeah, we definitely need to work on the spec of what it should do :)

[pbashyal-nmdp]

Created #51 to track the inconsistencies.

[mmaiers-nmdp]

after installing 0.0.22

ard.redux:    A*01:AB            , G = 
ard.redux_gl: A*01:AB            , G = A*01:01:01G/A*01:02
ard.redux:    A*01:01:01         , G = A*01:01:01G
ard.redux_gl: A*01:01:01         , G = A*01:01:01G
ard.redux:    HLA-A*01:AB        , G = 
ard.redux_gl: HLA-A*01:AB        , G = HLA-A*01:01:01G/HLA-A*01:02
ard.redux:    HLA-A*01:01:01     , G = HLA-A*01:01:01G
ard.redux_gl: HLA-A*01:01:01     , G = HLA-A*01:01:01G
ard.redux:    DRB1*14:06:01      , G = DRB1*14:06:01
ard.redux_gl: DRB1*14:06:01      , G = DRB1*14:06:01
ard.redux:    DRB1*14:06:02      , G = DRB1*14:06:02
ard.redux_gl: DRB1*14:06:02      , G = DRB1*14:06:02
ard.redux:    A*01:AB            , lgx = A*01:AB
ard.redux_gl: A*01:AB            , lgx = A*01:01/A*01:02
ard.redux:    A*01:01:01         , lgx = A*01:01
ard.redux_gl: A*01:01:01         , lgx = A*01:01
ard.redux:    HLA-A*01:AB        , lgx = HLA-A*01:AB
ard.redux_gl: HLA-A*01:AB        , lgx = HLA-A*01:01/HLA-A*01:02
ard.redux:    HLA-A*01:01:01     , lgx = HLA-A*01:01
ard.redux_gl: HLA-A*01:01:01     , lgx = HLA-A*01:01
ard.redux:    DRB1*14:06:01      , lgx = DRB1*14:06
ard.redux_gl: DRB1*14:06:01      , lgx = DRB1*14:06
ard.redux:    DRB1*14:06:02      , lgx = DRB1*14:06
ard.redux_gl: DRB1*14:06:02      , lgx = DRB1*14:06```

[rsajulga-nmdp]

Ah. Thank you @pbashyal-nmdp and @mmaiers-nmdp ! I forgot to set an upstream repository (I was pulling from my outdated fork). It works now!

[rsajulga-nmdp]

Should we be using a virtual environment for BDD testing? And should we put testing documentation in README.rst?

[pbashyal-nmdp]

Yes.

Yes. Under a Testing heading.

[rsajulga-nmdp]

Okay, documentation for creating a virtual environment and running behave are added.