Implement dictionary support for reading ByteView from parquet

[XiangpengHao]

Which issue does this PR close?

Part of #5904 , sequel to #5968 and #5970 and #5972

This PR is not ready to review until the above PRs are merged.

Rationale for this change

Implement dictionary encoding support for reading ByteView from parquet.

To show that reading ByteViewArray is faster than reading ByteArray:

cargo bench --bench arrow_reader --features="arrow test_common experimental" "arrow_array_reader/Binary.*Array/dictionary encoded"

We should get something like (~3x faster than ByteArray, many times faster than current implementation):

arrow_array_reader/BinaryArray/dictionary encoded, mandatory, no NULLs
                        time:   [270.32 µs 270.34 µs 270.36 µs]

arrow_array_reader/BinaryArray/dictionary encoded, optional, no NULLs
                        time:   [268.11 µs 268.14 µs 268.18 µs]

arrow_array_reader/BinaryArray/dictionary encoded, optional, half NULLs
                        time:   [372.87 µs 372.91 µs 372.95 µs]

arrow_array_reader/BinaryViewArray/dictionary encoded, mandatory, no NULLs
                        time:   [110.81 µs 110.82 µs 110.83 µs]

arrow_array_reader/BinaryViewArray/dictionary encoded, optional, no NULLs
                        time:   [111.89 µs 111.91 µs 111.92 µs]

arrow_array_reader/BinaryViewArray/dictionary encoded, optional, half NULLs
                        time:   [115.55 µs 115.56 µs 115.57 µs]

This is expected as the new implementation never copies data; it tries its best to reuse the buffer from the parquet decoder.

What changes are included in this PR?

This PR is (even) less straightforward than the previous ones, mostly because we need to map the dictionary view buffer indexes to the output view buffer indexes. Please let me know if any point of the code is too difficult to maintain!

Are there any user-facing changes?

[alamb]

I rebased this branch so it only had the changes to support dictionary reading

[alamb]

Thanks @XiangpengHao -- this looks very clever 👌

I wonder if there is some way this code should be tested? Or will be be tested as part of a larger end to end parquet reading exercise later?

[alamb]

That is a clever optimization

[alamb]

Do we know for sure k are valid dictionary indices (aka since this is from untrusted input, do we have to verify that k is within the bound of the dictionary)?

[XiangpengHao]

The input is untrusted, but I think we can do very little to make it safer, e.g., if the view is maliciously crafted we can easily run into memory issues due to too large offset or invalid buffer idx.
But I agree that having this early check is useful, if it doesn't make the code significantly slower. I added the check here, will come back to this when I file the PR to optimize the performance.

[XiangpengHao]

Or will be be tested as part of a larger end to end parquet reading exercise later?

The code to enable and test dictionary encoding is here:
https://github.com/apache/arrow-rs/pull/5973/files#diff-50e48ffbe585053165f5fdc7421dc30b1d1c9abf3ada6aff43dea9f7907489acR482-R486

[alamb]

Looks good to me -- thanks @XiangpengHao