Q: X must be a vector with type <distribution>, Instead, it has type <distribution>.

[Fuco1]

When I try to turn a tsibble into a fable using as_fable I get this odd error:

Error: `fbl[[chr_dist]]` must be a vector with type <distribution>.
Instead, it has type <distribution>.

The thing is that when I look at an output of forecast the column looks the same and is written out as <dist> which I think is this type. To be honest I quite fail to understand the vctrs package but I thought that distributions, models, hilos etc are "vectors" by default (hence they print as <distribution[n]> where n is the length).

I think this should be a complete repro case (might need to load some additional package but I'm sure you have them all somewhere on your system :))

library(tibble)
library(tsibble)
library(fable)
library(fabletools)

tibble(
    .model = "test",
    date = yearmonth(ymd('2019-01-01') + months(1:12 - 1)),
    quantity = dist_normal(1:12 * 10, sqrt(1:12) * 5),
    .mean = 10
) %>%
    as_tsibble(key = ".model") %>%
    as_fable(response = ".mean", distribution = quantity)

[mitchelloharawild]

Works for me. Try installing the latest version of fabletools from CRAN.

library(tibble)
library(tsibble)
library(fable)
#> Loading required package: fabletools
library(fabletools)
library(lubridate)
#> 
#> Attaching package: 'lubridate'
#> The following object is masked from 'package:tsibble':
#> 
#>     interval
#> The following objects are masked from 'package:base':
#> 
#>     date, intersect, setdiff, union
library(distributional)

tibble(
  .model = "test",
  date = yearmonth(ymd('2019-01-01') + months(1:12 - 1)),
  quantity = dist_normal(1:12 * 10, sqrt(1:12) * 5),
  .mean = 10
) %>%
  as_tsibble(key = ".model") %>%
  as_fable(response = ".mean", distribution = quantity)
#> Using `date` as index variable.
#> Warning: The dimnames of the fable's distribution are missing and have been set
#> to match the response variables.
#> # A fable: 12 x 4 [1M]
#> # Key:     .model [1]
#>    .model     date    quantity .mean
#>    <chr>     <mth>      <dist> <dbl>
#>  1 test   2019 Jan   N(10, 25)    10
#>  2 test   2019 Feb   N(20, 50)    10
#>  3 test   2019 Mar   N(30, 75)    10
#>  4 test   2019 Apr  N(40, 100)    10
#>  5 test   2019 May  N(50, 125)    10
#>  6 test   2019 Jun  N(60, 150)    10
#>  7 test   2019 Jul  N(70, 175)    10
#>  8 test   2019 Aug  N(80, 200)    10
#>  9 test   2019 Sep  N(90, 225)    10
#> 10 test   2019 Oct N(100, 250)    10
#> 11 test   2019 Nov N(110, 275)    10
#> 12 test   2019 Dec N(120, 300)    10

^{Created on 2020-09-07 by the reprex package (v0.3.0)}

[Fuco1]

Yep, I was on 2.0.something devel version, after updating it works. Sorry for the noise!