search

SWS-Methodology / faoswsImputation

Repository for storing code and associated files for the imputation package. This package implements flexible data imputation with the use of ensembles of models.
0 stars 0 forks source link

Fix problem with vignette #2

Closed joshbrowning2358 closed 8 years ago

joshbrowning2358 commented 8 years ago

The current vignette depends on the faoswsProduction package (which contains datasets and a good example of how to use faoswsImputation). However, faoswsProduction depends on faoswsImputation and thus we have a circular dependency if we build the faoswsImputation vignette. Thus, the vignette has simply been ignored in the past, and this needs to be fixed.

joshbrowning2358 commented 8 years ago

I've fixed this issue. We can close this issue and return to building vignettes for faoswsImputation.

sebastian-c commented 8 years ago

I'm still getting the following when building the vignette:

Quitting from lines 734-738 (faoswsImputation.Rnw) 
Error: processing vignette 'faoswsImputation.Rnw' failed with diagnostics:
Type of RHS ('double') must match LHS ('integer'). To check and coerce would impact performance too much for the fastest cases. Either change the type of the target column, or coerce the RHS of := yourself (e.g. by using 1L instead of 1)
Execution halted
Error: Command failed (1)
Execution halted

It seems to be in this block:

<<fig.height=3>>=
model = ensembleModel(model = defaultMars, extrapolationRange = Inf,
                      level = "local")
imputationParameters$ensembleModels = list(model)
imputeVariable(data = exampleData, imputationParameters = imputationParameters)
@

I'm guessing it's a type problem with a data.table merge. Does it happen on your machine as well?

joshbrowning2358 commented 8 years ago

Strange... I have seen that error before with this vignette, but it seems to be running fine on my machine. Can you try it again after a git pull? I pushed a few changes shortly after closing this issue...

sebastian-c commented 8 years ago

My latest commit hash is a41c1fff725ff48bae5922c71598a24b46ab26c5, which is where the master branch is now. I'm getting some inconsistent errors, though.

I tried knitting the Rnw file myself using Rstudio, but it breaks really early:

Error in match.arg(options$results, c("verbatim", "tex", "hide")) : 
  'arg' should be one of "verbatim", "tex", "hide"
Calls: <Anonymous> -> SweaveParseOptions -> check -> match.arg
Execution halted

I tried using purl to grab out the code, but then I get stuck on line 19 (okra = fread(input = file)) because system.file doesn't find an extdata directory. Using the knit function directly on the document gives the same results.

So frankly, I can't work out how it gets as far as it does in the first place when knitting the vignette.

joshbrowning2358 commented 8 years ago

Ok, I've changed the way I'm storing the data. Now, instead of being in the inst/ directory, it's in data, and it's saved as an .RData instead of .csv. Both compile fine for me, but maybe this will work for you... Let me know!