Hi, Just this morning I've done the final "todo" in the polytiles script: now it produces tiles 10 times faster, because it renders them in metatiles. Also MBTiles writer now supports multiprocessing, so it's also 6-8 times more speed. If you need a lot of tiles, do not download them from osm.org or other sources, render them yourself: it is finally faster than downloading.
https://github.com/zverik/polytiles I'd also like to remind of Nik4, the best tool for preparing printed maps of any region, any size and any quality; and of BigMap 2 for downloading and stitching some tiles from any of ~20 sources. The latter can even work online: "Enqueue" button queues a task, and in 2-3 minutes you'll get your image. The limit for that is 100 tiles. https://github.com/zverik/Nik4 http://bigmap.osmz.ru/ As for Nik4, there is "Get Veloroad" service for generating big vector maps online (basically a front-end to Nik4), but the server has only data for Russia and some neighbouring countries. Last week Dmitry Kiselev has made a Docker container for that: just install Docker, pull the container and start it: sudo apt-get install docker.io sudo docker pull dkiselev/nik4web sudo docker run -p 8081:80 -i -t dkiselev/nik4web /bin/bash startup The front-end would be at http://localhost:8081/nik4. You would need to run osm2pgsql with your region before requesting images. With it you can get PNG or vector SVG images (resolution independent), pre-processed for a sane page size and for easier movement of labels. Finally, yesterday I've translated into English a small geocoding exercise: https://github.com/Zverik/visgeocode (live at http://zverik.github.io/visgeocode/en.html ). It takes a CSV file with addresses in one of columns, runs these addresses through MapQuest Nominatim, and allows dragging resulting markers if geocoding was not precise. Then it can turn markers into building contours, but since my server has only Russia, you'd need to set up your own (maybe install the cgi script locally). The result can be downloaded either as CSV or as GeoJSON. We used this page for geocoding 22k addresses in Saint-Petersburg; Russian version uses three geocoders for better quality. IZ _______________________________________________ talk mailing list talk@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk