rsetienne / daisie Goto Github PK

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

An Rplot.pdf file is created when tests are run. Find out what is causing it and change it so that a file is not stored.

Probably a problem in a plotting test.

Rates functions must be tested and functioning. Last functioning version is in f63027a.
Not correct as of now.

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Delete?

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

NAMESPACE should be created using roxygen2.
This requires adding :: before function calls from other package, see f63027a.
I will convert to a new namespace to prevent having to juggle exported functions manually, and because the current namespace breaks pedro-ontogeny

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

If ran:

set.seed(1)
DAISIE::DAISIE_sim(
  time = 10,
  M = 1000,
  pars = c(1.0e+00, 4.0e-01, 1.5e+01, 1.0e-04, 2.0e-01),
  replicates = 100,
  divdepmodel = "IW"
)

on any branch develop or above format IW fails due to btimes being empty.

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

While working on #11, I noticed the following:
The cladogenesis calculation in get_clado_rate is different than the one in loglik_CS_time.

In the loglik_CS_time rhs function, it is computed this way:
lacvec <- pmax(rep(0, lnn), lac0 * (1 - nn / (area * K0)))

while in get_clado_rate it is computed like this:

clado_rate <- max(c( N * lac * island_area(timeval, Apars, island_ontogeny) * (1 - N / (island_area( timeval, Apars, island_ontogeny) * K)), 0), na.rm = T) where island_area evaluates to area as in above.

The question is: should the nn vector be included in the calculation (missing in the first example)? I added this vector in 9c72df4, which makes the null-ontogeny and ontogeny likelihoods (last test of DAISIE_test.R) be different, implying the change is incorrect. Is the N * term in get_clado_rate incorrect, then?
Asking @rsetienne for help, please!

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Increase coverage
• Merge

After changes of today, all branches must PR from develop. PR #9 broke pedro-ontogeny.
After fixed, will PR for others.

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

I see in e.g. Josh's Travis build that Travis produces 10k lines of output. Due to this, one cannot read the reason why a build fails.

I suggest to follow the UNIX Golden Rule of Silence: if everything works, no output should be produced. In our contexts: tests should not produce output (unless you test them to have output, which is very rare).

I will try to assign @Neves-P to do this, he's perfectly able to let another team member fix this, as this is a rather easy Issue 👍

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

The Travis CI build log states:

Undocumented code objects:
  ‘DAISIE_loglik_all_choosepar3’ ‘DAISIE_ML3’ ‘order_pars1’
  ‘translate_island_ontogeny’

Apparently, there is some undocumented code that breaks the build.

I can only suggest the authors of those functions to document these.

Related to (and to prevent) #21

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

Depends on:

• #5

Rules of how to contribute should be clearly laid out

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge

• Lint
• Document
• Coverage
• Merge