Interfacing With Gazebo
Gazebo exposes a Google Protobuf API interface to interface with the simulator, allowing one to read the state of the simulator and also control the simulator. The API documentation can be found here.
The actual Google ProtoBuf message definitions (proto files) for Gazebo can be found in gazebo/msgs/ in the Gazebo source code repository. If you haven’t done so already clone this repo. Once you have all of the proto files you need to create a language bindings for Java, C++ or Python using the `protoc’ tool (I believe other languages are supported too). Install by,
Well now look at how to generate bindings for Python. Official documentation can be read here.
The protoc tool will compile the .proto files into the corresponding Python files that can be interfaced with your code. The following command will place all the generated files in build/gen.
There is a bug currently that will not allow the generated Python code to use relative imports to be compatible with Python 3. For example if you place all of your generated code in a separate subdirectory, if the generated code needs to import one another this currently will not work.
The work around is to place an init.py file in the generated directory to signify the directory is a package and then run the 2to3 tool to automatically fix all the imports. The following command will fix imports (-f import) and overwrite the fixed files in the msg/ directory.
I’ve already done for you in the pygazebo library.
Unfortunately the original project seems abandoned and has outdated messages. My
fork has the current updated
protobuf messages for Gazebo 8. For future releases you just updated the Python
bindings as explained previously.
Just clone this branch and install,
The library uses asyncio and is a bit confusing. I’ll include some examples in a future post.