Toots for sckottie@fosstodon.org account

Written by Scott Chamberlain on 2025-02-04 at 23:26

[#]python habanero v2.2 released https://pypi.org/project/habanero/

release notes https://github.com/sckott/habanero/releases/tag/v2.2.0

small release, impacts tests only

@crossref

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2025-02-04 at 21:17

Follow up blog post on this

https://recology.info/2025/02/r-examples-internal/

Keeping internal function examples alive

[#]rstats https://fosstodon.org/@sckottie/113913038108676372

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2025-01-29 at 19:53

TIL - aha! git clean

clean out untracked files/dirs after you did some stuff and just want to nuke the untracked stuff

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2025-01-29 at 18:23

@maelle given your post about internal #rstats fxns https://blog.r-hub.io/2019/12/12/internal-functions/ what do you think is best practice for examples for internal functions? seems like I see this pattern alot: export the function, but use keywords internal - thoughts?

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2025-01-26 at 18:23

Best goodreads alternative not owned by Amazon?

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2025-01-24 at 18:02

No one told me one of the worst parts of having kids is being up in the middle of the night flicking through your greatest life regrets because somehow the mind just has to go there constantly

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2025-01-19 at 23:39

what might #rstats folks use to manually read an Rd file, modify it, then write it back replacing contents on disk? the use case is to sanitize secrets that may have ended up in example output. maybe it's easier to just use cli tools (sed, etc.)?

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2025-01-13 at 20:03

TIL run a program w/o access to your env vars, what do you use?

env - /path/to/program

env -i /path/to/program

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2024-12-09 at 22:06

[#]rstats cowsay v1 is out! blog post: https://recology.info/2024/12/cowsay-v1/

Written by Scott Chamberlain on 2024-11-27 at 14:56

Local first pod https://localfirst.fm/

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2024-11-25 at 18:40

woohoo, GitHub/PyPi trusted publishing worked like a dream

[#]python habanero v2 out https://pypi.org/project/habanero/ - @crossref client

release notes: https://github.com/sckott/habanero/blob/main/Changelog.rst

=> View attached media

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2024-11-23 at 01:00

any good reason why #rstats says a zero length string has length 1?

[#]python: len(“”) == 0

[#]ruby: "".length == 0

[#]node/#deno: x.length == 0

[#]julia: length(“”) == 0

etc.

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2024-10-31 at 16:31

[#]rstats Is there a best practice or just opinions for how to present errors?

e.g. in https://gist.github.com/sckott/4d220f06cead8dc8a4d8ee2dbce5546a some_fun1 points to actual fxn where error occured & last_trace points to it as well. Whereas some_fun2 points to the user facing fxn, but need last_trace(drop=FALSE) 4 where

I feel like some_fun1 is better b/c you have to jump through fewer hoops to find the actual source of the error, but maybe some_fun2 is considered better for users? maybe there's a some_fun3 that's even better 🤷‍♂️

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2024-10-18 at 21:46

this was it from @elinwaring https://mastodon.social/@elinwaring/113320396475741126 - seems so simple, just hadn't thought about what tbl(con, “table”) was doing #rstats

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2024-10-16 at 22:51

[#]rstats friends - is there a dplyr/ & friends way to do select on tables when privileges mean there's only partial access for a user? I imagine i'm just missing something obvious. Are privileges checked immediately in tbl on the entire table rather than checking for any column level rules? https://gist.github.com/sckott/35ef6ec03cd0fe848742c0c947e88457

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2024-10-15 at 16:14

So let's say you use Makefile for common #rstats pkg building tasks. If you have a Remotes dependency, how do you ensure it's installed before or duing R CMD INSTALL (with minimal or no additional deps)?

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2024-10-10 at 16:37

Is there an #rstats withr thing for temporarily unloading a namespace in a test block? Seems like there's with_package for temporarily load a package, but not the reverse

=> More informations about this toot | View the thread

Written by Scott Chamberlain on 2024-10-10 at 04:31

[#]rstats thoughts on this behavior for intersect ? it’s clear the code within intersect has changed, but it's not clear to me if it's intended that list names are dropped?

x <- list(fruit = "pear", meat = "chicken")

y <- list(fruit = "pear")

intersect(x, y)

[#] R 4.4.1 (retains list names)

[#]> $fruit

[#]> [1] "pear"

[#] R-devel, 4.5.0 (drops list names)

[#]> [[1]]

[#]> [1] "pear"

=> More informations about this toot | View the thread