business-science / timetk

Time series analysis in the `tidyverse`
https://business-science.github.io/timetk/
613 stars 101 forks source link

Upkeep #153

Closed olivroy closed 1 year ago

olivroy commented 1 year ago

Update gh actions, remove a dependency. Make rlang dependency on 1.1.1 explicit (fix #151 )

Use ggplot2 linewidth instead of size.

Remove roxygen2 and covr dependencies as they are not runtime dependencies.

Use modeltime conditionally on vignettes.

olivroy commented 1 year ago

@mdancho84 , could you trigger the workflow again? seems like I forgot a modeltime::

mdancho84 commented 1 year ago

Just started it.

olivroy commented 1 year ago

CI passing. But I get this warning in tests. this was there before I did some work

── Warning ('test-tk_index.R:36:5'): tk_index(ts) test returns correct format. ──
  'tzone' attributes are inconsistent
  Backtrace:
       ▆
    1. ├─... %>% tk_index(timetk_idx = TRUE) at test-tk_index.R:36:4
    2. ├─timetk::tk_index(., timetk_idx = TRUE)
    3. ├─timetk:::tk_index.ts(., timetk_idx = TRUE) at timetk/R/index-tk_index.R:77:4
    4. │ └─ret %>% lubridate::as_datetime() %>% lubridate::as_date() at timetk/R/index-tk_index.R:111:12
    5. ├─lubridate::as_date(.)
    6. ├─lubridate::as_datetime(.)
    7. └─lubridate::as_datetime(.)
    8.   └─lubridate (local) .local(x, ...)
    9.     ├─base::as.POSIXct(x, origin = origin, tz = tz)
   10.     └─base::as.POSIXct.numeric(x, origin = origin, tz = tz)
   11.       ├─base::.POSIXct(...)
   12.       └─base::`+.POSIXt`(as.POSIXct(origin, tz = "GMT", ...), x)
   13.         ├─base::.POSIXct(unclass(e1) + unclass(e2), .check_tzones(e1, e2))
   14.         └─base::.check_tzones(e1, e2)

In R 4.3.0 news. as.Date.POSIXct(., tz) now treats several tz values, notably "GMT" as equivalent to "UTC", proposed and improved by Michael Chirico and Joshua Ulrich in PR#17674.