search

SEELab / enaR

enaR = Ecological Network Analysis in R
22 stars 13 forks source link

Pack Issue #162

Closed deh9951 closed 9 years ago

deh9951 commented 9 years ago

When using the pack function, the 'living' input uses a lower case 'l', but the 'Living' piece of the network object uses a capital 'L'. This is not a major issue, but caused come confusion with coding and may cause problems for some users.

MKLau commented 9 years ago

Hey Dave, thanks for finding this. Would you send me a few lines that show the issue you were having?

pawan1992 commented 9 years ago

@deh9951 This got buried? I played a bit with the pack and unpack functions and checked the network objects but couldn't find the issue! except for the slight inconsistency between the pack and unpack functions. _--- If this issue is resolved, stop reading! :) ---_

Find the code and outputs that I came across, the details of the package are also provided in the end of the document.

Lets get some random data:

F1 <- matrix(rnorm(100), nrow = 10)
inp <- runif(10)
exp <- rexp(10)
out <- abs(rt(n = 10, df = 3))
sto <- runif(n = 10, min = 45, max = 4500)
liv <- sample(c(TRUE, FALSE), size = 10, replace = TRUE)

On this generated data; I construct a Network object using the 'pack' function.

y1 <- pack(flow = F1, input = inp, export = exp, output = out, storage = sto, living = liv)
## [1] "respiration"

## Warning in pack(flow = F1, input = inp, export = exp, output = out, storage
## = sto, : Missing model components: respiration

Now, notice that the network object contains the vertex attribute "living" just as the pack function specifies; this is consistent with the existing Network models as can be seen later.

get.vertex.attribute(y1, attrname = "living")
##  [1]  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE
liv
##  [1]  TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE  TRUE

Whereas when I try it with "Living"; I get NAs.

get.vertex.attribute(y1, attrname = "Living")
##  [1] NA NA NA NA NA NA NA NA NA NA

However, I realized that the unpack function returns "Living" component in the list when a Network object is 'unpacked'. But this shouldn't be a scope for confusion for users. Should it?

y2 <- unpack(y1)
class(y2)
## [1] "list"
str(y2)
## List of 7
##  $ F     : num [1:10, 1:10] 0.4809 0.0345 0.7979 0.2371 0.7659 ...
##   ..- attr(*, "dimnames")=List of 2
##   .. ..$ : chr [1:10] "1" "2" "3" "4" ...
##   .. ..$ : chr [1:10] "1" "2" "3" "4" ...
##  $ z     : num [1:10] 0.955 0.847 0.578 0.569 0.413 ...
##  $ r     : num [1:10] 0 0 0 0 0 0 0 0 0 0
##  $ e     : num [1:10] 1.71 0.307 0.8 0.172 0.728 ...
##  $ y     : num [1:10] 0.462 0.674 0.383 0.794 0.343 ...
##  $ X     : num [1:10] 4480 2178 2397 2553 2677 ...
##  $ Living: logi [1:10] TRUE FALSE FALSE FALSE FALSE FALSE ...

Here, though the "Living" component is spelled out with an "L", the help file does mention it clearly.

Also, the same goes for the saved Ecosystem models in 'enaR'.

data("troModels"); data("bgcModels")
get.vertex.attribute(troModels[[16]], "living") #For troModels
##  [1]  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE  TRUE
## [12]  TRUE FALSE FALSE FALSE
get.vertex.attribute(bgcModels[[16]], "living") #For bgcModels
## [1]  TRUE  TRUE  TRUE FALSE FALSE FALSE FALSE

Hence, at least in Version 2.8.1, the 'pack' function is behaving fine.

packageDescription('enaR')
## Package: enaR
## Type: Package
## Title: Tools for Ecological Network Analysis (ena)
## Version: 2.8.1
## Date: 2015-01-14
## Author: M.K. Lau, S.R. Borrett, D.E. Hines, P. Singh
## Maintainer: Matthew K. Lau <enaR.maintainer@gmail.com>
## Description: Functions for the analysis of ecological networks
## Depends: R (>= 2.10), MASS, stringr, sna, network, gdata
## Suggests: codetools, igraph, R.rsp
## VignetteBuilder: R.rsp
## License: GPL-3
## URL: https://github.com/SEELab/enaR
## Packaged: 2015-01-14 15:58:08 UTC; Aeolus
## NeedsCompilation: no
## Repository: CRAN
## Date/Publication: 2015-01-14 19:32:22
## Built: R 3.2.1; ; 2015-06-22 20:15:00 UTC; unix
## 
## -- File: /usr/local/lib/R/site-library/enaR/Meta/package.rds
MKLau commented 9 years ago

Hey guys, I get the issue now. Having the unpack function name the living vector “Living” is a bit confusing. I went ahead and changed it on the beta branch so that unpack calls the living vector “living” so that it’s in-line with the network attribute.

I’ll add this change to the next update to the master branch and CRAN, but for now to install from beta:

library(devtools) install_github('SEELab/enaR',ref='beta')

Thanks for working on this!

Matt

On Aug 25, 2015, at 2:28 AM, Pawan notifications@github.com<mailto:notifications@github.com> wrote:

@deh9951https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_deh9951&d=BQMCaQ&c=WO-RGvefibhHBZq3fL85hQ&r=yYM9bdhQa7fj0pqtCBPxMvjOe6gKXC-k_OhnGcaB0ZI&m=wbY8mbGsUaiR2WIZcLJsi_cPIBFNc7gnL80qijyRnXQ&s=PjRPhwwYVJ5sNqmQZ3ywiLxubpwbMQlM-uKYGw3dBlw&e= This got buried? I played a bit with the pack and unpack functions and checked the network objects but couldn't find the issue! --- If this issue is resolved, stop reading! :) ---

Find the code and outputs that I came across, the details of the package are also provided in the end of the document.

Lets get some random data:

F1 <- matrix(rnorm(100), nrow = 10) inp <- runif(10) exp <- rexp(10) out <- abs(rt(n = 10, df = 3)) sto <- runif(n = 10, min = 45, max = 4500) liv <- sample(c(TRUE, FALSE), size = 10, replace = TRUE)

On this generated data; I construct a Network object using the 'pack' function.

y1 <- pack(flow = F1, input = inp, export = exp, output = out, storage = sto, living = liv)

[1] "respiration"

Warning in pack(flow = F1, input = inp, export = exp, output = out, storage

= sto, : Missing model components: respiration

Now, notice that the network object contains the vertex attribute "living" just as the pack function specifies; this is consistent with the existing Network models as can be seen later.

get.vertex.attribute(y1, attrname = "living")

[1] TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE TRUE

liv

[1] TRUE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE TRUE

Whereas when I try it with "Living"; I get NAs.

get.vertex.attribute(y1, attrname = "Living")

[1] NA NA NA NA NA NA NA NA NA NA

However, I realized that the unpack function returns "Living" component in the list when a Network object is 'unpacked'. But this shouldn't be a scope for confusion for users. Should it?

y2 <- unpack(y1) class(y2)

[1] "list"

str(y2)

List of 7

$ F : num [1:10, 1:10] 0.4809 0.0345 0.7979 0.2371 0.7659 ...

..- attr(*, "dimnames")=List of 2

.. ..$ : chr [1:10] "1" "2" "3" "4" ...

.. ..$ : chr [1:10] "1" "2" "3" "4" ...

$ z : num [1:10] 0.955 0.847 0.578 0.569 0.413 ...

$ r : num [1:10] 0 0 0 0 0 0 0 0 0 0

$ e : num [1:10] 1.71 0.307 0.8 0.172 0.728 ...

$ y : num [1:10] 0.462 0.674 0.383 0.794 0.343 ...

$ X : num [1:10] 4480 2178 2397 2553 2677 ...

$ Living: logi [1:10] TRUE FALSE FALSE FALSE FALSE FALSE ...

Here, though the "Living" component is spelled out with an "L", the help file does mention it clearly.

Also, the same goes for the saved Ecosystem models in 'enaR'.

data("troModels"); data("bgcModels") get.vertex.attribute(troModels[[16]], "living") #For troModels

[1] TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE TRUE

[12] TRUE FALSE FALSE FALSE

get.vertex.attribute(bgcModels[[16]], "living") #For bgcModels

[1] TRUE TRUE TRUE FALSE FALSE FALSE FALSE

Hence, at least in Version 2.8.1, the 'pack' function is behaving fine.

packageDescription('enaR')

Package: enaR

Type: Package

Title: Tools for Ecological Network Analysis (ena)

Version: 2.8.1

Date: 2015-01-14

Author: M.K. Lau, S.R. Borrett, D.E. Hines, P. Singh

Maintainer: Matthew K. Lau enaR.maintainer@gmail.com<mailto:enaR.maintainer@gmail.com>

Description: Functions for the analysis of ecological networks

Depends: R (>= 2.10), MASS, stringr, sna, network, gdata

Suggests: codetools, igraph, R.rsp

VignetteBuilder: R.rsp

License: GPL-3

URL: https://github.com/SEELab/enaR

Packaged: 2015-01-14 15:58:08 UTC; Aeolus

NeedsCompilation: no

Repository: CRAN

Date/Publication: 2015-01-14 19:32:22

Built: R 3.2.1; ; 2015-06-22 20:15:00 UTC; unix

-- File: /usr/local/lib/R/site-library/enaR/Meta/package.rds

— Reply to this email directly or view it on GitHubhttps://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_SEELab_enaR_issues_162-23issuecomment-2D134494513&d=BQMCaQ&c=WO-RGvefibhHBZq3fL85hQ&r=yYM9bdhQa7fj0pqtCBPxMvjOe6gKXC-k_OhnGcaB0ZI&m=wbY8mbGsUaiR2WIZcLJsi_cPIBFNc7gnL80qijyRnXQ&s=pdbBLbk-iwhFXBrPgCnX-ED4J9AH_sGH1NBNcbGhXtE&e=.

pawan1992 commented 9 years ago

@MKLau Thanks for addressing this promptly Matt. _Great that you've also changed the help files!_ @deh9951 was it the entire issue? Did we miss something?

deh9951 commented 9 years ago

Hi All,

Thanks for looking into this, it sounds like you've fixed it- the "living" vs "Living" problem was the entire issue I think.

-Dave

On Wed, Aug 26, 2015 at 2:17 AM, Pawan notifications@github.com wrote:

@MKLau https://github.com/MKLau Thanks for addressing this promptly Matt. Great that you've also changed the help files! @deh9951 https://github.com/deh9951 was it the entire issue? Did we miss something?

— Reply to this email directly or view it on GitHub https://github.com/SEELab/enaR/issues/162#issuecomment-134853072.