Map Generator

### Expected behavior

- Users should be able to generate unique maps for their locations if they choose to do so.
- Users should be able to tweak various parameters to influence the map generation.
- Users should be able to manually edit maps after generation.
- Users should be able to "save" their map, either as an image or another reusable format.

### Actual behavior

- Users currently only have the ability to upload existing maps.

### Bonus behavior

- This doesn't actually have to live in Notebook.ai and can be its own app to be integrated later, as long as users can export their map for using on other sites. Bonus if it lives inside Notebook, though.

### Approach

1. We can set up [this fantasy map generator](https://github.com/mewo2/terrain) (algorithm [explanation here](http://mewo2.com/notes/terrain/)) to generate maps with some tweakable parameters.
2. We need a UI for such map generation, parameter tweaking, and previewing/confirming maps, as well as some kind of viewing UI unless we tie it into existing image uploads.
3. [This article](http://mewo2.com/notes/naming-language/) can be used for naming locations, although we could use our existing location name generator and/or pull names from other locations in the universe (especially if generating a map for the universe, as opposed to some specific universe) or other locations linked to the location we're generating a map for.

### Other resources/links:

- http://davidson16807.github.io/tectonics.js/
- http://topps.diku.dk/torbenm/maps.msp
- https://donjon.bin.sh/scifi/world/
- https://donjon.bin.sh/fantasy/world/
- http://pcg.wikidot.com/pcg-algorithm:map-generation
- https://cartographersguild.com/content.php

No files or tests are named. Start by reviewing the existing location and image-upload flows, then evaluate the linked terrain generator and its parameter options. Done would require an agreed scope plus generation, editing, previewing, and export behavior, but the issue does not define the implementation boundaries.

Written by the indexing model from the issue text.