Development with VSCode¶
Debugging¶
You can debug Python and C++ by installing the Python C++ Debugger extension. The website provides examples for debugging in Windows and with gdb.
For Apple silicon, install the CodeLLDB extension and create the file .vscode/launch.json with the following configurations.
This extension allows debugging C++ code.
Launch Configurations¶
Change the path to eigenlldb.py to enable pretty printing of C++ types.
{
"configurations": [
{
"name": "* C++ Attach",
"type": "lldb",
"request": "attach",
"pid": "",
"initCommands": [
# ***** CHANGE THIS *****
"command script import \"YOUR-PATH-TO-LUPNT/LuPNT/eigenlldb.py\"",
# ***** CHANGE THIS *****
],
},
{
"name": "* Python",
"type": "debugpy",
"request": "launch",
"program": "${file}",
"cwd": "${fileDirname}",
"console": "integratedTerminal"
},
{
"name": "* Python/C++ Debugger",
"type": "pythoncpp",
"request": "launch",
"pythonLaunchName": "* Python Debugger: Current File",
"cppAttachName": "* Attach",
},
],
}
The Python/C++ Debugger starts your typical debugging session in python and passes the PID of your process to the C++ debugger.
Settings¶
Edit the .vscode/settings.json file to debug all targets using CodeLLDB.
Change the path to eigenlldb.py to enable pretty printing of C++ types.
{
"cmake.sourceDirectory": "${workspaceFolder}/all",
"cmake.debugConfig": {
"name": "* C++ Launch",
"type": "lldb",
"request": "launch",
"initCommands": [
// ***** CHANGE THIS *****
"command script import \"YOUR-PATH-TO-LUPNT/LuPNT/eigenlldb.py\"",
// ***** CHANGE THIS *****
],
},
}
Pretty Printing¶
To use the eigenlldb.py script for pretty printing with the CodeLLDB extension, we need to install numpy within its Python distribution.
Open the VSCode Command Palette by pressing Command + Shift + P (or Ctrl + Shift + P), and select LLDB: Command Prompt.
Once the LLDB command prompt opens, enter pip install numpy.
Debugging Variables¶
In the VSCode debug console use p <variable-name> to print the name of a variables or ? <variable-name> to inspect its raw contents. For example, for position vector and velocity vector would be.
Vec3 r(4338.99, -4338.99, -0.0757297);
Body moon = Body::Moon();
p position
(lupnt::Vec3) (3,1) (static,static)
[[ 4338.99 ]
[-4338.99 ]
[ -0.0757297]]
p body
(const lupnt::Body &) 0x00000001674e0210
? body
{id:MOON, name:"MOON", GM:4902.8001180000001, ...}
id = MOON
name = "MOON"
GM = 4902.8001180000001
R = 1737.4000000000001
n = 20
m = 1
fixed_frame = MOON_PA
inertial_frame = MOON_CI
use_gravity_field = true
gravity_field = {n_max:1, m_max:3, n:1803047168, m:1, ...}
[raw] = const lupnt::Body &
Building the Python Bindings¶
Run the CMake target pylupnt-dev to automatically build pylupnt, i.e., LuPNT and the python bindings, copy the generated _pylupnt.*.so file to source/python/pylupnt, and generate stubs for the bindings. The stubs allow IDEs like VSCode to find the contents of a compiled package such us our bindings. Finally, add source/python/pylupnt to your python path by adding the environment variable
export PYTHONPATH="${PYTHONPATH}:YOUR-PATH-TO-LUPNT/LuPNT/source/python"
You can check whether it was added by fully restarting your terminal session or VSCode and executing the following command
python -c "import sys; print(sys.path)"
Now you can import the package and use it in your python scripts.
python -c "import pylupnt as pnt; print(pnt.R_MOON)"
Pre-Commit Hook¶
To ensure that the code is formatted correctly, we use the pre-commit tool. The configuration file .pre-commit-config.yaml is located in the root directory of the repository. To install the pre-commit hook, run the following commands in the root directory of the repository.
pip install pre-commit
pre-commit install
This command installs the pre-commit hook in the repository. The hook is executed before each commit and checks the code formatting. If the code is not formatted correctly, the commit is rejected. To bypass the hook, use the --no-verify option when committing.
To test the pre-commit hook, run the following command in the root directory of the repository.
pre-commit run --all-files
Python Version¶
You can specify the python version to use in the .vscode/settings.json file.
"cmake.configureArgs": [
# ***** CHANGE THIS *****
"-DPYTHON_EXECUTABLE=YOUR-PATH-TO-PYTHON/bin/python"
# ***** CHANGE THIS *****
],
The path to the python executable can be found by running the following command in the terminal.
which python
Make sure you use Clean Reconfigure All Projects in the CMake extension to apply the changes. You should see the following message in the CMake output.
[cmake] -- pybind11 v2.13.1
[cmake] -- Found PythonInterp: YOUR-PATH-TO-PYTHON/bin/python (found suitable version "3.9.19", minimum required is "3.7")
[cmake] -- Found PythonLibs: YOUR-PATH-TO-PYTHON/lib/libpython3.X.dylib
It is ok to see other versions of python in the output of other libraries like Format.cmake or autodiff, as long as pybind11 finds the correct version.
Python Environment¶
We recommend using micromamba to create a new environment for the project.
Conda and mamba can be used for package management, but mamba is generally preferred due to its faster performance. Mamba is a partial re-implementation of conda in C++, utilizing multi-threading and libsolv for improved speed. It still relies on some conda code and has a dependency on Python. Alternatively, micromamba is a pure C++ re-implementation with no Python dependency, making it a better choice for creating lightweight images from scratch. It is possible to use mamba’s solver from within conda.
To create a new environment with micromamba, run the following commands in the root directory of the repository.
micromamba create -n lupnt python=3.11
micromamba activate lupnt
micromamba install --file requirements.yml