Some warning message / check suggestions for add_FA

[fredtal]

I got the fatty acid data loaded again today, but I had some trouble and it took me a while to figure out what was up. I eventually tracked it down to these places and thought some warning messages / checks might help.

1. DataIO: line 294: "stopifnot(dim(mean_c)[1]==n.preys & dim(mean_c)[2]==n.fats)": add something like, "number of fatty acids in Conv.Coeff.mean does not match number of fatty acids in the predators file" and/or "number of prey in Conv.Coeff.mean does not match number of preys in preys file".
2. DataIO: line 294: "stopifnot(dim(var_c)[1]==n.preys & dim(var_c)[2]==n.fats)": "number of fatty acids in Conv.Coeff.var does not match number of fatty acids in the predators file" and/or "number of prey in Conv.Coeff.var does not match number of preys in preys file".
3. DataIO: line 371: " R[,,i]=cov(alr(preys[preys.ix==unique(preys.ix)[i],]))*ni[i]". I got an error here, I think because there was one prey category for which there was only one prey item, so it couldn't do a covariance matrix? I removed that prey item and it went okay. If that is indeed the problem, maybe check to make sure there is more than one prey item in each category and throwing an error before doing the cov(alr(...)) call?

[fredtal]

Also (1) and (2) in the same place in add_SI (but I knew what the problem was this time! :)).

[Philipp-Neubauer]

IF you feel confident enough to do that, you could clone the repo from
github, add the error codes and submit a pull request to merge it in. A
good first start to contributing :)

On Wed, Jul 1, 2015 at 10:24 AM, fredtal [email protected] wrote:

Also (1) and (2) in the same place in add_SI (but I knew what the problem
was this time! :)).

—
Reply to this email directly or view it on GitHub
#23 (comment)
.

Phil

[fredtal]

HA. Okay I'll think about it. :)

[fredtal]

I made some changes and tried to push them from R Studio but then got an error. :( I'm sorry; I'm not very good with github.

error: unable to read askpass response from 'rpostback-askpass'
fatal: could not read Username for 'https://github.com': Device not configured

[Philipp-Neubauer]

ok - can you check that your git is configured? I.e., you should have a git
username and password set up for your device. It looks like it can't find
it.

On Thu, Jul 2, 2015 at 6:36 AM, fredtal [email protected] wrote:

I made some changes and tried to push them from R Studio but then got an
error. :( I'm sorry; I'm not very good with github.

error: unable to read askpass response from 'rpostback-askpass'
fatal: could not read Username for 'https://github.com': Device not
configured

—
Reply to this email directly or view it on GitHub
#23 (comment)
.

Phil

[Philipp-Neubauer]

PS you also won't have permission to push directly to the master. sorry if i confused you, but the right way (provided the user and email are correctly set up for you machine) is to
*fork the repository to your github, so it's on your personal github
*change the remote url for your local fastiR clone to your repo OR clone again from your repo and copy paste your changes there

• push to your own repo
• in github, submit a pull request to the master on my github repo.

Let me know if I explained that ok:)

Phil

On Jul 2, 2015, at 6:36 AM, fredtal [email protected] wrote:

I made some changes and tried to push them from R Studio but then got an error. :( I'm sorry; I'm not very good with github.

error: unable to read askpass response from 'rpostback-askpass'
fatal: could not read Username for 'https://github.com': Device not configured

—
Reply to this email directly or view it on GitHub.

[fredtal]

This is very helpful! And makes sense. It seemed strange that I might be allowed to just push to the master! I forked it but have to wait to do the rest of it because I'm still using github through rstudio (I need to move to command line but haven't), and my rstudio is busy running the model!! Whoop!

[fredtal]

ALL RIGHT! The problem with the credentials is one I had run into before but forgotten about: RStudio no longer supports http connections to github. You have to set it to ssh. I fixed that and I submitted my first batch of changes. But I've since made more so probably will have to submit more later. Sorry I got overenthusiastic and didn't do them all at once. :)

Also, the model ran last night! I'm going to assemble some plots and will send them to you.