A vector clock is a data structure used for determining the partial ordering of events in a distributed system and detecting causality violations.
See https://en.wikipedia.org/wiki/Vector_clock.
This implements a vector where each process has its own clock.
A typical description will have the counter start from 1 and increment by 1 every time there is a change.
However, this implementation allows the counter increase by whatever the client asks for. Our processes can set the counter to their own clock, hence allow us to resolve conflicts (unordered changes) by leaning towards later (more recent) object versions.
>>> from vectorclock import VectorClock
>>> vca1 = VectorClock({"A":1})
>>> vca2 = VectorClock.from_string('{"A":2}')
>>> print(vca1 < vca2)
True
>>> print(vca1 != vca2)
True
>>> print(vca1 == vca2)
False
>>> print(vca1 > vca2)
False
>>> print(str(vca1))
{"A":1}
# these two clocks are not ordered
>>> vcb = VectorClock({"B": 1})
>>> print(vca1 == vcb)
False
>>> print(vca1 < vcb)
False
>>> print(vca1 > vcb)
False
# If tie-breaking is off, they are not ordered (but not equal!)
>>> print(vca1.compare(vcb, tiebreak=False))
0
# If tie-breaking is on, we pick an order (deterministically)
>>> print(vca1.compare(vcb, tiebreak=True))
-1
The rest of this README is for vectorclock developers.
python -m unittest discover -s tests
python -m pip install --upgrade build
python -m build
python -m pip install twine
twine check dist/*
twine upload -r testpypi dist/*
Check that the distribution looks as expected. Now:
twine upload -r pypi dist/*