PAHO Dengue data has the same serotypes coded in different ways

PAHO Dengue data has the same serotypes coded in different ways. This is presumably from the upstream data, but we should probably standardize the spelling/spacing of them.

```
> aa <- pub_paho_dengue(regions = "mx", epiweeks = epirange(201101, 202301))
> View(aa)
> unique(aa$serotype)
[1] "  "          "DEN 1,2"     "DEN 1,2,3,4" "DEN1,2,3,4"  "DEN 3"       "DEN 1,2,3"   "DEN 1,2,34" 
```