Added as_tibble argument to read_ods.

This presents a bit of an issue with name repair as there are now three ways column names can be repaired:

1. No column names are given: Here columns are given sequential column names by passing their column number to cellranger::num_to_letter()
2. Only some column names are given: Here data.frame gives default column names to those missing names in the form "Var.%column_number%"
3. Duplicate column names are given and as_tibble's name_repair is invoked: Column names are repaired by prepending . and a number sequentially in the order of columns needing to be repaired.

More than one of these can occur, for example importing a sheet containing

1  a    a    <blank>
2  A    B    C

using arguments as_tibble = TRUE, .name_repair = "unique" (a likely common pairing) gives column names

a...1  a...2  Var.3

which is a little hard to predict.

However, this is likely the output users who have previously been using the pipeline

read_ods(file) |> as_tibble()

will have been expecting, and so a change from this would be a significant and unexpected breaking change, even though it might be preferable from a consistency point of view.