`epi_df` joins to data frames with richer keys yields invalid `epi_df`

[brookslogan]

• If x is an epi_df with some set of key cols, and y is a data frame with key cols that include all of x's key cols plus some more, and we join the two by x's key cols, we should be outputting an epi_df with y's key cols. But instead, we output an epi_df with only x's key cols, even when this violates the one-row-per-epikeytime constraint (we don't even decay to tibble).
• check_ukey_unique on errors is printing a string instead of actually generating an error (thus why the below needs to cat its result)

library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
library(epiprocess)
#> Loading required package: epidatasets
#> Registered S3 method overwritten by 'tsibble':
#>   method               from 
#>   as_tibble.grouped_df dplyr
#> 
#> Attaching package: 'epiprocess'
#> The following object is masked from 'package:stats':
#> 
#>     filter
target <- tibble(geo_value = 1, time_value = 1, target = 1) %>% as_epi_df()
sensors <- tibble(geo_value = 1, time_value = 1, sensor_name = c("A", "B"), sensor_value = 1:2)
joined <- left_join(target, sensors, by = c("geo_value", "time_value"))
joined
#> An `epi_df` object, 2 x 5 with metadata:
#> * geo_type  = hhs
#> * time_type = integer
#> * as_of     = 2025-06-03 10:24:05.131905
#> 
#> # A tibble: 2 × 5
#>   geo_value time_value target sensor_name sensor_value
#>       <dbl>      <dbl>  <dbl> <chr>              <int>
#> 1         1          1      1 A                      1
#> 2         1          1      1 B                      2
key_colnames(joined)
#> [1] "geo_value"  "time_value"
epiprocess:::check_ukey_unique(joined, key_colnames(joined)) %>% cat()
#> There cannot be more than one row with the same combination of geo_value
#> and time_value.  Problematic rows:
#> An `epi_df` object, 2 x 5 with metadata:
#> * geo_type  = hhs
#> * time_type = integer
#> * as_of     = 2025-06-03 10:24:05.131905
#> 
#> # A tibble: 2 × 5
#>   geo_value time_value target sensor_name sensor_value
#>       <dbl>      <dbl>  <dbl> <chr>              <int>
#> 1         1          1      1 A                      1
#> 2         1          1      1 B                      2

^{Created on 2025-06-03 with reprex v2.1.1}