abjur / abjutils Goto Github PK

View Code? Open in Web Editor NEW

50.0 19.0 12.0 69.05 MB

Useful functions for the Brazilian Jurimetrics Association

Home Page: https://abjur.github.io/abjutils

License: Other

R 100.00%

r jurimetrics toolkit package

abjutils's People

Contributors

Stargazers

Watchers

Forkers

ricarvalho42 catataw jimhester omattuy samuelbahia davisvaughan athospd karolinefirmino gaybro8777 luisclaudiodr moiseshealth

abjutils's Issues

rm_accent() poderia retornar um dataframe se um input for um objeto desse tipo

Bom, como ambos somos BR, não me importei em escrever uma issue em inglês =/

Voltando ao assunto: a função janitor::clean_names() recebe como input um objeto do tipo data.frame. Sei que ela e a rm_accent() tem objetivos diferentes, mas pensei em como seria legal poder usar a última dentro de um pipeline só:
df %>% clean_names() %>% rm_accent()

pvec() fails when output has zero length

Doesn't work:

f <- function(x) tibble::tibble()
abjutils::pvec(1:4, f)
#> Error: Tibble columns must have consistent lengths, only values of length one are recycled:
#> * Length 0: Columns `return`, `output`
#> * Length 4: Column `id`

^{Created on 2019-02-21 by the reprex package (v0.2.1)}

Curiously, this works:

f <- function(x) tibble::tibble(a = "a")[-1,]
abjutils::pvec(1:4, f)
#> # A tibble: 4 x 3
#>      id return output          
#>   <int> <chr>  <list>          
#> 1     1 result <tibble [0 × 1]>
#> 2     2 result <tibble [0 × 1]>
#> 3     3 result <tibble [0 × 1]>
#> 4     4 result <tibble [0 × 1]>

^{Created on 2019-02-21 by the reprex package (v0.2.1)}

pvec() should return names as id when they exist

pvec() works like a purrr::map_dfr() with .id = "id".

Hence, this is the desired behavior

purrr::map_dfr(letters, tibble::enframe, .id = "id")
#> # A tibble: 26 x 3
#>    .id    name value
#>    <chr> <int> <chr>
#>  1 1         1 a    
#>  2 2         1 b    
#>  3 3         1 c    
#>  4 4         1 d    
#>  5 5         1 e    
#>  6 6         1 f    
#>  7 7         1 g    
#>  8 8         1 h    
#>  9 9         1 i    
#> 10 10        1 j    
#> # ... with 16 more rows
names(letters) <- letters
purrr::map_dfr(letters, tibble::enframe, .id = "id")
#> # A tibble: 26 x 3
#>    .id    name value
#>    <chr> <int> <chr>
#>  1 a         1 a    
#>  2 b         1 b    
#>  3 c         1 c    
#>  4 d         1 d    
#>  5 e         1 e    
#>  6 f         1 f    
#>  7 g         1 g    
#>  8 h         1 h    
#>  9 i         1 i    
#> 10 j         1 j    
#> # ... with 16 more rows

function to remove path and extension

something like

f <- function(x) basename(tools::file_path_sans_ext(x))

what name should we use?

Usage of furrr

Hi there, I am working on the next version of furrr and your package has come up in my revdeps as having an issue.

The warning that I get is:

'::' or ':::' import not declared from: ‘lifecycle’

I see that you used to use furrr in this pvec() function, which has now been deprecated
https://github.com/abjur/abjutils/blob/master/R/pvec.R

Since it has been deprecated for awhile, I think the easiest solution would be for you to mark it as .Defunct(), and then remove all usage of furrr from the package, or to remove the function altogether.

If you are not ready for that, then there are two things that I will need your help with:

This global override of furrr::future_options() is what causes the warning.

abjutils/R/pvec.R

Line 91 in cc3ec0c

future_options <- furrr::future_options

That function has now been soft deprecated in favor of furrr::furrr_options() (to free up that name for future to use if it wants to). Internally, my future_options() now calls a function from the lifecycle package. Since you have assigned my function directly into one of yours, R CMD Check found that you haven't Import-ed the lifecycle package, and it returned the above warning.

Since you only Suggest furrr, the easiest thing to do here is probably to just set your .options argument to default to NULL, and inside pvec() you would call:
```
if (is.null(.options)) {
  .options <- furrr::future_options()
}
```
Then you'd remove the override of furrr:future_options() on line 91.
This second point is optional, but would be good practice. Since you now only Suggest furrr, it is unsafe to have any function in your package that calls a furrr:: function without first checking if furrr is installed. The easiest way to do that is to put this at the top of pvec()
```
if (!rlang::is_installed("furrr")) stop("The furrr package must be installed to use this function")
```

Typo

abjutils/README.Rmd

Line 74 in 21adc74

    
           `test-fun()`    | Verifica se todos os argumentos de outro função já tem valores atribuídos

pvec() with .flatten=TRUE does not work for data frames

I think pvec() should work like purrr::map_dfr() when .flatten = TRUE

# devtools::install_github("abjur/abjutils")
fun <- function(x) {
  Sys.sleep(1)
  mtcars
}
res <- abjutils::pvec(letters, fun)
#> 
 Progress: ────────────                                                                       100%
 Progress: ─────────────────────────                                                          100%
 Progress: ─────────────────────────────────────                                              100%
 Progress: ──────────────────────────────────────────────────                                 100%
 Progress: ──────────────────────────────────────────────────                                 100%
 Progress: ───────────────────────────────────────────────────────────────                    100%
 Progress: ───────────────────────────────────────────────────────────────────────────        100%
 Progress: ───────────────────────────────────────────────────────────────────────────        100%
 Progress: ────────────────────────────────────────────────────────────────────────────────── 100%
 Progress: ────────────────────────────────────────────────────────────────────────────────── 100%
res
#> # A tibble: 26 x 3
#>       id return output                
#>    <int> <chr>  <list>                
#>  1     1 result <data.frame [32 × 11]>
#>  2     2 result <data.frame [32 × 11]>
#>  3     3 result <data.frame [32 × 11]>
#>  4     4 result <data.frame [32 × 11]>
#>  5     5 result <data.frame [32 × 11]>
#>  6     6 result <data.frame [32 × 11]>
#>  7     7 result <data.frame [32 × 11]>
#>  8     8 result <data.frame [32 × 11]>
#>  9     9 result <data.frame [32 × 11]>
#> 10    10 result <data.frame [32 × 11]>
#> # ... with 16 more rows
res <- abjutils::pvec(letters, fun, .flatten = TRUE)
#> 
 Progress: ────────────                                                                       100%
 Progress: ─────────────────────────                                                          100%
 Progress: ─────────────────────────────────────                                              100%
 Progress: ──────────────────────────────────────────────────                                 100%
 Progress: ──────────────────────────────────────────────────                                 100%
 Progress: ───────────────────────────────────────────────────────────────                    100%
 Progress: ───────────────────────────────────────────────────────────────────────────        100%
 Progress: ───────────────────────────────────────────────────────────────────────────        100%
 Progress: ────────────────────────────────────────────────────────────────────────────────── 100%
 Progress: ────────────────────────────────────────────────────────────────────────────────── 100%
#> Error in eval_tidy(enquo(var), var_env): object 'output' not found

Also, should we assume that there will be no errors when .flatten = TRUE?

Errors not getting caught

Since the latest change to pvec, errors are being reported as results and appear as <NULL>.

abjutils and purrr 0.3.1

Hello, I'm preparing the release of purrr 0.3.1 and I see this revdep failure:

https://github.com/tidyverse/purrr/blob/13ea8bfbf52696d12ba0453a241434b64e4a8066/revdep/problems.md

I think you have already fixed this 8 months ago in 7452d1b but this hasn't been pushed to CRAN yet. I'm planning to release purrr in 2 weeks.

You could do a preemptive fix release, or wait until purrr 0.3.1 is out. With that version, I think you can skip the unlist() entirely because modify("foo", function(...) "bar") now returns the same type as the input (here a character vector), as it always should have.

carefully doesn't work properly

Release abjutils 0.3.2

Prepare for release:

Submit to CRAN:

usethis::use_version('patch')
devtools::submit_cran()
Approve email

Wait for CRAN...

Accepted 🎉
usethis::use_github_release()
usethis::use_dev_version()

CPF mascarado

Oi, pessoal. Tudo bem?

Pesquisei aqui, mas não consegui descobrir. O pacote abjutils tem alguma função para pegar um CPF comum (111.222.333-67) ou (11122233367) e transformar em CPF mascarado (***.222.333-**)?

Obrigada!

Use progress::progress_bar in carefully() if cores = 1

if cores = 1 do not use parallel backend and add progress bar. Maybe also a verbose option

Add selection to escape_unicode

FEATURE REQUEST: escape_unicode function works on the whole document including comments, which as far as know doesn't need to be escaped. Rstudioapi has the primary_selection() function that allows to get just the selection. My suggestion is to allow users to escape just the body of the function by selecting it instead of the whole document.

Warning quando eu rodo abjutils::separate_cnj()

library(magrittr)
da <- tibble::tibble(
  a = "1000315-72.2016.8.26.0156"
)
da %>% 
  abjutils::separate_cnj(a)
#> Warning: Prefixing `UQ()` with the rlang namespace is deprecated as of rlang 0.3.0.
#> Please use the non-prefixed form or `!!` instead.
#> 
#>   # Bad:
#>   rlang::expr(mean(rlang::UQ(var) * 100))
#> 
#>   # Ok:
#>   rlang::expr(mean(UQ(var) * 100))
#> 
#>   # Good:
#>   rlang::expr(mean(!!var * 100))
#> 
#> This warning is displayed once per session.
#> # A tibble: 1 x 6
#>   N       D     A     J     T     O    
#>   <chr>   <chr> <chr> <chr> <chr> <chr>
#> 1 1000315 72    2016  8     26    0156

^{Created on 2020-09-10 by the reprex package (v0.3.0)}

Problems in the with_*() family

In the latest commit (0c4c89d), I added the with_*() family of functions. They are supposed to wrap common functions in the purrr::map() family to add new functionalities to them: parallelism, progress tracking and side effect capturing.

Unfortunately, some things aren't working as expected 👎 Parallelism doesn't work with imap() because of the .y placeholder, and nothing works with map_if() because we don't deal with the predicate .p.

Should we simply ignore the situations where the functions don't work or should we try to deal with these edge cases?

# Setup
fun <- function(x, y, z) { x + y + z }
fun2 <- function(x, y, z) { Sys.sleep(1); x + y + z }
l <- list(1, 2, 3, "a", 4, 5)

# Parallel functions
par_map <- with_parallelism(purrr::map)
par_imap <- with_parallelism(purrr::imap)
par_map_if <- with_parallelism(purrr::map_if)

# Parallel tests
par_map(1:10, fun, y = 1, z = 2)
par_map(1:10, ~.x+1)
par_map(1:10, ~fun(1, .x, 2))
par_imap(1:10, fun, z = 2)
par_imap(1:10, ~.x+.y)                         ### COMPLETELY WRONG
par_imap(1:10, ~fun(.y, .x, 2))                ### COMPLETELY WRONG
par_map_if(1:10, ~.x%%2==0, fun, y = 1, z = 2) ### DOESN'T WORK
par_map_if(1:10, ~.x%%2==0, ~.x+1)             ### DOESN'T WORK
par_map_if(1:10, ~.x%%2==0, ~fun(1, .x, 2))    ### DOESN'T WORK

# Progress functions
pro_map <- with_progress(purrr::map)
pro_imap <- with_progress(purrr::imap)
pro_map_if <- with_progress(purrr::map_if)

# Progress tests
pro_map(1:10, fun2, y = 1, z = 2)
pro_map(1:10, ~{ Sys.sleep(1); .x+1 })
pro_map(1:10, ~fun2(1, .x, 2))
pro_imap(1:10, fun2, z = 2)
pro_imap(1:10, ~{ Sys.sleep(1); .x+.y })
pro_imap(1:10, ~fun2(.y, .x, 2))
pro_map_if(1:10, ~.x%%2==0, fun2, y = 1, z = 2)      ### NO PROGRESS
pro_map_if(1:10, ~.x%%2==0, ~{ Sys.sleep(1); .x+1 }) ### NO PROGRESS
pro_map_if(1:10, ~.x%%2==0, ~fun2(1, .x, 2))         ### NO PROGRESS

# Possibly functions
pos_map <- with_possibly(purrr::map)
pos_imap <- with_possibly(purrr::imap)
pos_map_if <- with_possibly(purrr::map_if)

# Possibly tests
pos_map(l, fun, y = 1, z = 2)
pos_map(l, ~.x+1)
pos_map(l, ~fun(1, .x, 2))
pos_imap(l, fun, z = 2)
pos_imap(l, ~.x+.y)
pos_imap(l, ~fun(.y, .x, 2))
pos_map_if(l, ~.x%%2==0, fun, y = 1, z = 2) ### DOESN'T WORK
pos_map_if(l, ~.x%%2==0, ~.x+1)             ### DOESN'T WORK
pos_map_if(l, ~.x%%2==0, ~fun(1, .x, 2))    ### DOESN'T WORK

Correct check_dig

This should i) clean the id, ii) check if it has 20 digits and iii) check if the digit is wright

ii) should raise a warning message and return FALSE. iii) should return TRUE or FALSE

abjutils::check_dig("0005268-75.2013.8.26.100")

When returned object is NULL, pvec() doesn't work

f <- function(x) { if(runif(1)<.1) return(NULL) else return(mtcars)}
res <- abjutils::pvec(1:1000, f)
Error: Columns return, output must be length 1 or 1000, not 890, 890