This repo is to set up the pybind11 for the IvP libraries in MOOS-IvP

In particular we make a copy of parts of the ivp source files:

• pyivp/src/lib_geometry
• pyivp/src/lib_mbutils
• pyivp/src_unit_tests/

We intented to use pybind11 to create

• ivp/pybind11/

For the use of well-established functions.

We will use Docker to manage the dependencies, which should be minimal. Docker mounts the volumn ~/pyivp between the host and container.

We will also use movinngpandashttps://github.com/ARG-NCTU/movingpandas

To get started:

git clone --recursive [email protected]:ARG-NCTU/pyivp.git
cd ~/pyivp
source docker_run.sh

Or

source docker_join.sh

Note that we should compile and run in Docker. We use a root account in order to access driver and other hardware. Therefore, everything edited in container will belong to root.

We suggest to edit code in host.

The pyivp repo has been configured to a nbdev repo. Quote from nbdev

nbdev is a library that allows you to develop a python library in Jupyter Notebooks, putting all your code, tests and documentation in one place. That is: you now have a true literate programming environment, as envisioned by Donald Knuth back in 1983!

We will use nbdev to develop "high-level" API based on IvP and test all ivp libraries in Jupyter Notebook.

After you compile and install the library.

source Docker/colab_jupyter.sh

Open a browser, and enter the following

http://127.0.0.1:8888/tree

Now you could open and enjoy IvP in Jupyter Notebooks.

nbdev_clean_nbs

Remember to run this command to clear notebooks

The documentation shows how to use functions in ivp.

import pyivp

pattern = pyivp.string2SegList("format=lawnmower, x=0, y=0, height=60, width=180, lane_width=15, rows=north-south, startx=20, starty=-300, degs=45")

• must use format=lawnmower for pattern block.
• there is 4 entry, and (startx, starty) will decide where is the closest entry.

import pyivp

waypointPloy = pyivp.string2ploy("example")

• example => x=0, y=0, format=radial, radius=10, pts=4
• you can pass args such as : must have -> [x, y, z, radius, pts], snap, label, active, msg
• snap means the size of the grid, and the ploy's vertex will stick on the grid.

waypointSeg = waypointPloy.exportSegList(x0, y0)

• Build an XYSegList from the poly. Make the first point closest to the x0, y0.

waypointSeg.get_spec_pts()

• this will print the vertexs of ploy (clockwise!!) such as 'pts={-1,1:-1,-1:1,-1:1,1}'
• you can use waypointSeg.reverse() or waypointPloy.reverse() for counterclockwise
• you can use waypointSeg.is_clockwise() to check.

waypointSeg.get_vx() and waypointSeg.get_vy() 

• with for loop to get every vertex's information.

The class diagram is drawn by DotUML

Parts of the diagram is shown below: