Bump versions to 0.28.0-alpha.3

Add support for server side filtering of DataFusion DataFrames (#12147)

This PR adds support for server side (push down) filtering of DataFrames
from a dataset query. Specifically it adds support for
`rerun_segment_id` and time index. With this PR users can do complicated
queries such as the following:

```python
df = view.reader(index="log_tick").filter(
    (col("rerun_segment_id") == "recording_0") & (col("log_tick") == 123456)
)
```

This will be converted into an update in the dataset query or will
produce multiple queries as needed.

All of the results from the queries will be combined to find a unique
set of required chunks and then sent out to the appropriate DataFusion
partitions for requesting chunks from the server.

This PR adds support for boolean logic (AND/OR) to combine the segment
and time parameters. Additionally it supports `functions.in_list` and
`Expr.between`. If the filters include a non-supported expression, then
we will fall back to querying all data.

---------

Co-authored-by: Antoine Beyeler <49431240+abey79@users.noreply.github.com>

Bump versions to 0.27.3

Bump versions to 0.27.3-rc.4

Bump versions to 0.27.3-rc.2

Bump versions to 0.27.2

Bump versions to 0.27.2-rc.2

Bump versions to 0.27.1

Bump versions to 0.27.1-rc.1

Bump versions to 0.27.0