This is the repository for an open-source re-implementation of the awesomely successful libpmp risk-based estimation and project tracking software used by Jaybridge Robotics from 2013 to 2016.

• This project is not really baked enough for users other than its developers right now. Sorry!

• This project is structured as a python wheel build.
• For building and testing, prefer to use a virtual environment:

python3 -m venv .test_venv
.test_venv/bin/python3 -m pip install --upgrade pip
.test_venv/bin/pip install -e .

Of course you can use other environment managers like uv or poetry; those are not documented here in order to keep this documentation minimal; venv and pip are relatively stable and universal if not actually optimal.

• To perform linter checks run:

.test_venv/bin/ruff check .

• To run all of the tests:

.test_venv/bin/py.test

• To run the cost or progress estimator scripts, run

.test_venv/bin/cost

or

.test_venv/bin/progress

• You cannot commit until I (ggould256) add you as a contributor.
• Your first commit must add your name to the AUTHORS file.
• The tests listed above can be run via ./test/run_tests.sh, which is also run as a github action on each push.

• Separate subdirectories for time tracking, distribution math,
• Top-level directory should contain the executable scripts. docs, progress/leaderboard, emacs fu, etc.
• Time tracking should be via some sort of abstract interface, as it is unlikely any two users will have the same time tracking conventions.
• Use python3, not python2, primarily because this gets us the new division semantics and removes numerical landmines.
• The :closed: RST role in the original libpmp was done as a way to abuse a custom role present in trac. We will not reimplement it; until we see otherwise, we will assume that it is sufficient for people to delete the estimates from completed tasks.