Flexibility in the funs parameter

[philerooski]

While funs is currently spec'd to take a function or list of functions which return a a dataframe with column names (those column names then becoming the names of features in the outputted feature dataframe), it would be useful to allow funs to take a list of functions that return atomic numeric values. A one-column dataframe in a sense (but a length one numeric vector in actuality).

This actually works already, but the feature names outputted aren't very useful.

> gyroscope_features(
+   sensor_data = gyroscope_data,
+   funs = list(mean, median, sd))
Joining, by = "axis"
$extracted_features
# A tibble: 3 x 5
  measurementType axis      data.x    data.y   data
  <chr>           <chr>      <dbl>     <dbl>  <dbl>
1 velocity        x     -0.0000919 -0.000477 0.0141
2 velocity        y     -0.0000397 -0.00197  0.0379
3 velocity        z      0.00156   -0.000832 0.0190

$model_features
NULL

$error
NULL