Examples that use the lake

[philrz]

Some docs examples populate a scratch SuperDB data lake so that lake-specific operations using loaded pool data can be shown further down the page.

Details

There's two pages that use a similar approach of prepping a scratch lake up top and then referencing it further down. One is for the from operator:

https://github.com/brimdata/super/blob/8e86ac0ff4c9da6c278eb0ec9a8efba9f888c9d5/docs/language/operators/from.md?plain=1#L81-L96 https://github.com/brimdata/super/blob/8e86ac0ff4c9da6c278eb0ec9a8efba9f888c9d5/docs/language/operators/from.md?plain=1#L144-L146

Another is for the load operator:

https://github.com/brimdata/super/blob/8e86ac0ff4c9da6c278eb0ec9a8efba9f888c9d5/docs/language/operators/load.md?plain=1#L37-L49 https://github.com/brimdata/super/blob/8e86ac0ff4c9da6c278eb0ec9a8efba9f888c9d5/docs/language/operators/load.md?plain=1#L62-L67

It should also be noted that the super db tutorial understandably makes heavy use of the lake but has never even tried to be mdtest-protected in the past. Browsing the doc again right now to refresh myself, it's probably because there's too many unique outputs (e.g., commit/pool IDs) that it would be clunky to make it mdtest-runnable while still being readable. So perhaps that one stays as it is no matter what.

Ideas

This one feels like it would be challenging. Perhaps it's a special case of #5631 in that it implies having some separate Linux-like environment where shell commands can run to create input/state and then having that wired up so the Playground example can call on it. I'm quick to admit limited knowledge here as to what's possible.