San Francisco looks nice, but there seems to be a problem with the projection in some of the sample images. It looks as if it isn't UTM but a global sphere projection, which isn't suitable for local renders. It's suspicious that the word 'projection' isn't mentioned in the Readme.
It looks like the final images have some kind of vignetting to make the corners and outter edges fade away. Probably grabbing OSM tiles and doing some image processing.
Great project! It takes a bit of fiddling to get the distance parameter right in some cases, and two possible improvements would be to allow for coordinate input rather than city name so I can point it at a specific spot as the center of the map. Another one would be to output all styles simultaneously so I can visually select from the generated images.
I believe (from a quick code check on my phone) it should be possible to output the images to SVG with a little tweak, thanks to your use of matplotlib?
Is there a reason you’ve defaulted to PNG that I’m missing?
Holy heck exporting to PDF/SVG for large cities is laggy, just patched the repo and did it for one over half a million people and it’s insanely slow lol
There is throttling in the code when making api request to respect the map provider as per their terms and conditions. Can make few thing parallel but it's a hobby project how many posters would one person want to generate?.
For what it's worth, I use Vistaprint. That's not based on an extensive comparison, just that it always works for me and seems comparable in price and quality to other services.
That's splendid. I've long wanted to make a jigsaw puzzle out of Sydney's road map, so I can familiarise myself with the layout of roads while having fun. That way I can reduce my reliance on nav app and become one of those old-school drivers.
There's a reason Where 2 -> Google Maps happened in Sydney. The sheer number of one-way roads combined with the imposition of the harbour and the messy tunnel system make internalising Sydney navigation a life-long endeavour
What happens if there are multiple cities with the same name. Maybe this is a US specific problem, but there are many states with cities of the same name from another state. As a few examples:
If you just type "Salem" for the city, it just picks one and goes with it. If you type "Salem, Oregon" vs "Salem, MA", it gives the desired result (at least for the cases I've tried), and then uses your input text as the caption.
The bar only updates once that entire step is complete (ie, if step 1 of 3 is downloading roads, it won't tell you what % of roads have been downloaded, but rather it will remain at 0% until all roads are downloaded at which point it will jump to 33%).
It will work, give it time. Also default distance is 29000. Give distance of 10000 to see faster results. It certainly is working as many users have generated maps.
Are you using a coding package or graph-based approach for this, or is it standard image generation? I’ve seen very similar results on https://picxstudio.com —it’s an AI image generation platform.
Pretty cool! It would be great if there was a way to set coordinates manually, since Nominatim can sometimes produce mediocre results. Also, would be nice to have a way to render the same map in all themes, not just one.
The initial version were coordinates actually but then realised people mostly love their cities more than anything and it's easy that way. Will add this feature back as optional parameters.
I was curious about this too and it’s actually a bug— these blue dots are random parks or park features. Most parks are polygons, but less than 5% or them are points from volunteer mappers who didn’t make them polygons and so these stray points seem to slip through and get plotted as the matplotlib default color…
If only I knew this before I printed it out and attached it all together with rasterbator lol
Yes, the blue and orange dots are from the water and parks Nodes and Ways in the OSM data.
It doesn't look like the orange and blue colors are part of the theme definitions, so the rendering library may be using some default values. This is why they are rendered in the same color on images using different theme files.
# Get the directory where THIS script is located
SCRIPT_DIR=$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" &> /dev/null && pwd)
cd "$SCRIPT_DIR" || exit 1
# TODO: kolkata is done already, change to some other city both here and on that notepad.pw link in Chrome
readonly CITY="kolkata"
readonly COUNTRY="india"
readonly CONTAINER_NAME="map_to_poster"
if ! docker buildx build \
--file "Dockerfile" \
--progress none \
--tag "${CONTAINER_NAME}" \
--quiet .; then
printf "Error: %s" "while attempting to build the docker container:${CONTAINER_NAME}"
exit 1
fi
if ! docker container run \
--detach \
--interactive \
--name "${CONTAINER_NAME}" \
--tty \
"${CONTAINER_NAME}"; then
printf "Error: %s" "while attempting to run the docker container:${CONTAINER_NAME}"
exit 1
fi
if ! docker cp "${container_id}":/home/python/app/maptoposter/posters "${HOME}/map_to_poster"; then
printf "Error: %s" "while attempting to copy map files from container:${CONTAINER_NAME}"
fi
if ! docker stop "${CONTAINER_NAME}"; then
printf "Error: %s" "stopping container:${CONTAINER_NAME}"
fi
if ! docker rm "${CONTAINER_NAME}"; then
printf "Error: %s" "removing container:${CONTAINER_NAME}"
fi
https://github.com/anvaka/city-roads
Edit: it outputs SVGs and you can try it online. Anvaka (author) has other really cool dataviz projects.
reply