datasets forces only one dataset

On some modules (teal.osprey, at `tm_g_patient_profile()` but there are others) there is a multiple choice of datasets, where multiple datasets can be selected.
I thought we could use something like :

``` r
library("teal.picks")
teal.picks::datasets(choices = c("dataset1", "dataset2"), multiple = TRUE)
#> Error in `.pick()`:
#> ! formal argument "multiple" matched by multiple actual arguments
```

^{Created on 2026-07-24 with [reprex v2.1.1](https://reprex.tidyverse.org)}

As you can see we can't force to select multiple datasets because internally it uses `.pick(multiple = FALSE, ....)`.
Without it we could set something like:

```r
teal.picks::datasets(choices = starts_with("AD"), select = everything(), multiple = TRUE)
```

And then resolve as usual.

I think the default of `multiple = FALSE` is sensible but we might want to make it a bit more flexible so we can work around the issue.
So the new API could be `function(choices = tidyselect::everything(), selected = 1L, fixed = NULL, multiple = FALSE, ...) {` and use that argument for `.picks()`

Start by reproducing the shown teal.picks::datasets call, then trace datasets and its internal .pick/.picks calls, including the usage from tm_g_patient_profile(). The change is complete when callers can request multiple datasets through the proposed API while the default single-dataset behavior still works.

Written by the indexing model from the issue text.