summ() throws "incorrect number of dimensions" error when using random intercepts lmer model with confint = TRUE

**Description:**
When using `summ()` with an random intercept multilevel model and setting `confint = TRUE`, the function fails with an index out of bounds error. This is caused by dropped dimensions when computing the confidence intervals.

**Reprex:**
``` r
library(lme4)
#> Loading required package: Matrix
library(jtools)

model <- lmer(Petal.Length ~ 1 + (1 | Species), data = iris)
summ(model, confint = TRUE)
#> Error in the_cis[, 1]: incorrect number of dimensions
```

^{Created on 2025-06-19 with [reprex v2.1.1](https://reprex.tidyverse.org)}

**Proposed solution:**
When computing `the_cis` in `summ.merMod()`, consider adding the `drop = FALSE` argument to the subset:

``` r
the_cis <- confint(model, parm = "beta_", method = conf.method[1], level = ci.width)
the_cis <- the_cis[rownames(the_sum$coefficients), , drop = FALSE]
```

This argument will prevent `the_cis` to be simplified from a matrix to a vector when there is only one row (i.e. the intercept).

Start in the summ.merMod() implementation and reproduce the failure with the lme4 iris model shown in the issue using confint = TRUE. Inspect the the_cis subset used for the confidence intervals and verify that the single-intercept case remains a matrix. Done means the reprex runs without the incorrect number of dimensions error.

Written by the indexing model from the issue text.