Bump dependencies on Delphi packages #12
Conversation
Bump versions, update renv.lock. Adjust for breaking changes: - Some obvious renames. - Replace both `pivot_quantiles` and `unnest` `pivot_wider` patterns to `pivot_quantiles_wider`. - Fix some `epi_recipe` and `frosting` printing that doesn't play well with knitr now. Update _freeze files where a first pass visual diff identified "real" differences rather than just a tiny visual offset.
A first pass visual diff did not identify any "real" differences with the current versions, and suggests that some of these changes might just be a tiny offset and/or size adjustment, but this would need a double check before trying to avoid the updates to the _freeze & corresponding repo bloat.
* remove references to R6 and mutation * use epiprocess correctly * fix the authors section of DESCRIPTION * upgrade renv * update all packages in renv * integrate Rprofile with user Rprofile
feat: update for epiprocess R6 refactor
* add a README file * fix broken formatting in packages.bib * get missing data for sliding-forecasters.qmd online instead of local files
|
@dajmcdon this should be good to go. Mostly just updating the code to the most up to date versions of our packages. This includes my updates to the new |
| #| echo: false | ||
| #| collapse: true | ||
| extract_recipe(out_gb$epi_workflow) | ||
| with_messages_cat_to_stdout(print(extract_recipe(out_gb$epi_workflow))) |
There was a problem hiding this comment.
Generic comment:
with all these patterns, we lose the display of extract_recipe() (or similar) usage. Maybe we want to think about showing the feature in 1 chunk and printing with a second chunk?
There was a problem hiding this comment.
If I understand correctly, you want both code saying
extract_recipe(........)and the better-formatted output from with_messages_......... that doesn't turn it into a bunch of chunks?
I think current state of PR should do that. Every replacement was of something that was already echo: false, except for five_days_ahead$epi_workflow, which was split into a results: false chunk + echo: false chunk.
[That is, unless the pre-existing material was missing some results: false chunks where we actually want them. I think I just assumed they were omitted to avoid clogging the page.]
| } | ||
|
|
||
| @Manual{R-epipredict, | ||
| title = {epipredict: Basic epidemiology forecasting methods}, | ||
| author = {Daniel McDonald and Ryan Tibshirani and Logan Brooks and Rachel Lobay and Maggie Liu and Ken Mawer and Chloe You}, | ||
| author = {Daniel McDonald and Ryan Tibshirani and Logan Brooks and Rachel Lobay}, |
There was a problem hiding this comment.
We may want to rethink author credit in both of our packages. These both strike me as overly exclusive.
Co-authored-by: Daniel McDonald <[email protected]>
Co-authored-by: Daniel McDonald <[email protected]>
Summary: bump epidatr & epipredict dependency versions, fix breaking changes, update
_freeze.Breaking changes:
pivot_quantilesandunnestpivot_widerpatterns topivot_quantiles_wider.epi_recipeandfrostingprinting that doesn't play well withknitr now. (Workaround from here.)
_freezeupdates seemed like they might have both trivial and nontrivial changes. I've tried to divide between commits based on whether I thought were significant vs. those that potentially could be pruned if they seem trivial on a second pass.Some other notes / points to check; if someone else can take a second look that would be great:
lm_wf_log %>% fit(penguins)output is missing step numbering still (messages probably being filtered out), but it's inside a larger chunk and I didn't want to duplicate it to apply the workaround; it looks fine without the numbering anyway.