Add query benchmark module and performance docs page

[hrodmn]

I was inspired to write this up after seeing how well titiler-pgstac can perform with high resolution global datasets in collections with massive STAC item footprints (check out this demo). In contrast, collections with many tiny (<100 km squares) do not perform as well because the query time for low zoom tile (z<=5) searches is too long for titiler-pgstac (check out this demo).

It is not strictly a pgstac problem but the goal is to provide some food for thought to pgstac users who get to choose the size of their STAC items/assets but don't have much context for why any size is good/bad.

• Benchmark module simulates XYZ tile queries for collections with different geographic layouts for items
• Adds a new page to docs outlining tradeoffs between number of items and speed of queries
• Adds mkdocs-jupyter (and a few other dependencies) to the dependencies for the mkdocs site