[arrow-avro] Move the writer-wire planning fully into `codec.rs`

**Is your feature request related to a problem or challenge? Please describe what you are trying to do.**

Improve the boundary between the `arrow-avro` codec and decoder by ensuring planning logic related to the skipper is handled in `codec.rs` and not during decoder construction or parsing.

**Describe the solution you'd like**

I'd recommend pushing the writer-wire planning fully into `codec.rs`, instead of carrying full `AvroDataType`s down into `record.rs` and having `Skipper::from_avro` reconstruct the skip tree there. However this can always be done in a follow-up as well.

_Originally posted by @jecsand838 in https://github.com/apache/arrow-rs/issues/9605#issuecomment-4172638199_

**Describe alternatives you've considered**

1. Leaving the code structure as is, which is fine for now, but overtime will likely reduce maintainability.

**Additional context**

N/A

Start by reading arrow-avro's codec.rs and record.rs, then trace Skipper::from_avro to understand where writer-wire planning is currently reconstructed. Done means planning related to the skipper is handled in codec.rs rather than during decoder construction or parsing, without carrying full AvroDataTypes into record.rs.

Written by the indexing model from the issue text.