[experiment] expose readable datasets as dataframes and arrow tables

[sh-rp]

Description

As an alternative to the ibis integration, we are testing out wether we can create our own data reader with not too much effort that works across all destinations.

Unordered notes:

• Add some kind of interface to the filesystem so the user has some control over which files are considered for a request
• Find out if exposing dataframes from jsonl files should be discouraged as duckdb probably will not be able to only read the exact right places in the file as it can with parquet
• Should we allow the user to select which columns they want (probably yes), allow ordering (also probably yes, but with docs on which types of ordering are performant) or allow joins (I'm on the fence, user interviews will solve this)

[netlify]

✅ Deploy Preview for dlt-hub-docs canceled.

Name	Link
🔨 Latest commit	20bf9ce
🔍 Latest deploy log	https://app.netlify.com/sites/dlt-hub-docs/deploys/667c3e9d6a1f7f0008eea44a