This is a rather rough script to generate a GPX file for the optimum "burbing" route. Burbing is the process of riding every road in a suburb or locality.
This script take the fun out of trying to work it out yourself.
Given a suburb name, Optiburb fetches data from OpenStreetMap and tries to calculate the optimum route aroud the suburb, and spits out a GPX file.
The solution is based on the Route Inspection Problem (AKA Chinese Postman Problem), which is a well-known mathematical problem and algorithm. There is no amazing code in here - This program is calling code from libraries to do all the heavy lifting for OSM, graph theory and algorithms. This program just glues them together.
There are some options to prune out unnamed gravel tracks, to import boundaries from shapefiles
Some known limitations:
This repo hasn't been set up as a proper installation or package yet - it's just a script and you'll need to sort out your own Python dependencies. You will need to pip install the following packages (tested on late version of Python 3 only so far).
pip install -r requirements.txt
% ./optiburb.py
2020-07-28 21:13:26 optiburb.py:__init__:64 [WARNING] WARNING - this program does not consider the direction of one-way roads or other roads that may be not suitable for your mode of transport. You must confirm the path safe for yourself
usage: optiburb.py [-h] [--debug DEBUG] [--start START] [--prune] [--simplify] [--simplify-gpx] [--select SELECT] [--shapefile SHAPEFILE] [--buffer BUFFER] [--save-fig] [--save-boundary] ...
Optimum Suburb Route Generator
positional arguments:
names suburb names with state, country, etc
optional arguments:
-h, --help show this help message and exit
--debug DEBUG debug level debug, info, warn, etc
--start START optional starting address
--prune prune unnamed gravel tracks
--simplify simplify OSM nodes on load
--simplify-gpx reduce GPX points
--select SELECT select the nth item from the search results. a truely awful hack because i cant work out how to search for administrative boundaries.
--shapefile SHAPEFILE
filename of shapefile to load localities, comma separated by the column to match on
--buffer BUFFER buffer distsance around polygon
--save-fig save an SVG image of the nodes and edges
--save-boundary save a GPX file of the suburb boundary
%
To fetch data from OSM search using overpass API, state the long winded name of the suburb.
./optiburb.py "bellfield, victoria, australia"
You can add multiple adjoining suburbs and they will be merged together (just incase a single suburb isn't big enough).
If the suburb fails with some weird message about no nodes in the polygon, you may have selected a name, instead of the whole locality. I haven't worked out how to specifically search for a administraive boundary yet, so the ugly hack in the short term is to use --select 2 to pick the next search result. Sometimes if you specify the entire name including the broader admin boundary (shire, country, etc) it may help.. You can experiment with the search at https://www.openstreetmap.org/
./optiburb.py --select 2 "footscray, victoria, australia"
You can also import polygon boundaries from shapefiles, but you'll need to know the column name and the key in advance. In this example, the shapefile is from the Australian Government with state locality boundaries.
./optiburb.py --save-fig --save-boundary --prune \
--shapefile ~/Projects/gis/VIC_LOCALITY_POLYGON_shp,vic_loca_2 KEW
Pruning a route will attempt to remove unnamed tracks, which tend to be 4wd tracks and not a lot of fun to do on most bikes.
You save the polygon boundary as a GPX files, which is handy for loading into head units to see where you're going.
You can also save the SVG node file, which shows the various intersections and other OSM points.
If you want to confine your route to a radius around a point (eg, 5km around your house), the following trick should work:
like this:
./optiburb.py --debug=debug --save-fig --save-boundary \
--start "23 main rd, suburb, state, country" \
--buffer 5000 \
"23 main rd, suburb, state, country"
For very dense suburbs, this might be huge.