██▒ █▓ ▄▄▄ ██▓ ██░ ██ ▄▄▄ ██▓ ██▓ ▄▄▄
▓██░ █▒▒████▄ ▓██▒ ▓██░ ██▒▒████▄ ▓██▒ ▓██▒ ▒████▄
▓██ █▒░▒██ ▀█▄ ▒██░ ▒██▀▀██░▒██ ▀█▄ ▒██░ ▒██░ ▒██ ▀█▄
▒██ █░░░██▄▄▄▄██ ▒██░ ░▓█ ░██ ░██▄▄▄▄██ ▒██░ ▒██░ ░██▄▄▄▄██
▒▀█░ ▓█ ▓██▒░██████▒░▓█▒░██▓ ▓█ ▓██▒░██████▒░██████▒▓█ ▓██▒
░ ▐░ ▒▒ ▓▒█░░ ▒░▓ ░ ▒ ░░▒░▒ ▒▒ ▓▒█░░ ▒░▓ ░░ ▒░▓ ░▒▒ ▓▒█░
░ ░░ ▒ ▒▒ ░░ ░ ▒ ░ ▒ ░▒░ ░ ▒ ▒▒ ░░ ░ ▒ ░░ ░ ▒ ░ ▒ ▒▒ ░
░░ ░ ▒ ░ ░ ░ ░░ ░ ░ ▒ ░ ░ ░ ░ ░ ▒
░ ░ ░ ░ ░ ░ ░ ░ ░ ░ ░ ░ ░ ░ ░ ░
░
Valhalla is an open source routing engine and accompanying libraries for use with Open Street Map data. This library/service, Meili, provides a set of algorithms and datastructures for map matching. It matches a sequence of locations (usually noisy e.g. GPS trajectory) to the underlying road network. In keeping with the Norse mythological theme, the name Meili, Thor's brother, was chosen. Since map matching is closely related to routing and since Thor is the Valhalla routing library, meili seemed most appropriate. Additionally the main author of this software, @ptpt from team Mapillary noted that, mĕilì (美丽) means beautiful in Chinese. This is indeed a beautiful collaboration between team Mapillary and team Valhalla! Open source FTW!
Baldr uses the GNU Build System to configure and build itself. To install on a Debian or Ubuntu system you need to install its dependencies with:
srcipts/dependencies.sh
And then run to install it:
scripts/install.sh
Please see ./configure --help
for more options on how to control the build process.
Follow this tutorial to get the service run in Docker.
We welcome contributions to meili. If you would like to report an issue, or even better fix an existing one, please use the meili issue tracker on GitHub.
If you would like to make an improvement to the code, please be aware that all valhalla projects are written mostly in C++11, in the K&R (1TBS variant) with two spaces as indentation. We generally follow this c++ style guide. We welcome contributions as pull requests to the repository and highly recommend that your pull request include a test to validate the addition/change of functionality.
We highly encourage running and updating the tests to make sure no regressions have been made. We use the Automake test suite to run our tests by simply making the check
target:
make check
You can also build a test coverage report. This requires that the packages lcov
, gcov
and genhtml
be installed. On Ubuntu you can get these with:
sudo apt-get install lcov
To make the coverage report, configure the build for it:
./configure --enable-coverage
And generate an HTML coverage report in the coverage/
directory:
make coverage-report
Note also that, because calculating the coverage requires compiler support, you will need to clean any object files from a non-coverage build by running make clean
before make coverage-report
.