Comment 9 for bug 1822455

@costales I'm writing from a position of ignorance with what I say below, so I won't have my feelings hurt if this is a terrible idea, but...

Is there a way to make a UBports specific script/program that took no user input but ran that Docker-based personal server shown at https://github.com/Overv/openstreetmap-tile-server on the mobile device just for long enough to make tiles go into the ~/.local/share/navigator.costales folder?
https://github.com/Overv/openstreetmap-tile-server#preserving-rendered-tiles

And maybe did cron jobs to update the tiles occasionally?

The imagined UBports specific script/program above could use the current GPS to select what region to bring into the device as well as set a maximum reasonable boundary around that center for the tiles being output to uNav's offline maps folder.
e.g.: adjust boundaries in order to get a set that has a total number of tiles smaller than x-many. (mostly because users like me will ask for unreasonable quantities if it's free to do so)

http://tools.geofabrik.de/calc/#type=geofabrik_standard&bbox=-6.266747,54.447649,-5.599081,54.771571

Changing topics a bit here:
My most vital use case of offline-maps with Google Maps is for an extended trip where I fear having no signal during rural or farmland sections of the drive. I will use Google Maps to pull in overlapping segments of the drive at the largest zoom out preview that their app lets me do it for along what I know to be the intended cross-country trip. This is excessively more data than I need. It's so wasteful but I don't have a better way to ensure the drive will be successful.

My secondary use case is my hometown map (about a 10-mile radius around a home address), just to keep the total data consumption to a minimum while on a cellular network connection because it costs money per byte transferred. Most of the time I don't need the map program, but I let it help me to remember/confirm the best route under confusion or to re-route when road problems along my intended path arise.

Just "thinking out loud" here. It would probably be impossible to make a program that could convert a driving route into a minimum set of necessary offline tiles that would accommodate wrong turns, detours, and finding food/fuel/lodging/repairs along the way. That would be fantastic though!

In my youth, long before personal computers, we would go to the AAA office (https://www.aaa.com/) and get a "TripTik" flipbook made by a clerk while we waited. It was only a 15-minute ordeal to have a book made which would take you across the country.
https://pearlsoftravelwisdom.boardingarea.com/2014/01/remember-triptix/
http://wisechapter3.com/2015/04/20/marketing-triptik-why-every-company-needs-one/
Somehow they would be able to quickly go from two endpoints to a selection of ordered sheets to put into the flip-page binding. The narrow books had all pages with North at the top. They did it with paper so maybe that human algorithm and system of identifying pre-printed sheets would solve such a problem of building a minimum set of tiles for an extended length offline trip.

Oh, and the AAA clerk would also hand you the regional paper maps that your TripTik went across. So an ideal offline map system would have the zoom level 1-9 tiles of the trip's overview (mostly for when things went wrong), and the zoom level 14-18 tiles containing the roads which the route touches as well as the adjacent 4 tiles on either side of the one the route is on. Or maybe just every zoom level tile that has the route on it, plus the adjacent tiles pertinent to wrong turns and finding food/fuel/lodging/repairs along the way.