puan-db-python 0.1.0

No project description provided

A quick what-is-dis

A Python library for connecting to a puan-db instance. Puan DB is a logical reasoning database which draws conclusions from a model and finds combinations which satisfies a model while minimizing/maximizing some objective function.

There are two types of variables to keep in mind:

• Primitives. These are defined by an ID given by you, and a lower and upper bound representing which integer values this variable may take in the future. The bound is a complex number so a boolean variable x would be declared with the bound 0+1j, or just 1j, as such model.set_primitive("x", 1j)
• Composites. These are created using one or more other primitives or composites as a connecter between them. For instance, an OR-connector is defined as model.set_or(["x","y","z"]).

Example

import puan_db_python

model = puan_db_python.PuanClient(
    host="localhost",
    port=50051,
    password=None,
    ssl=False
)

# Set some primitive variables
model.set_primitives(["x","y","z"])

# And a composite
id = model.set_atleast(["x","y","z"], 1, alias="A")

# Same one could be set instead by
other_id = model.set_or(["x","y","z"], alias="A")
assert id == other_id

# If you forget the id but remember the alias, there's a pointer to it
forgot_id = model.id_from_alias("A")
assert (id == forgot_id) and (forgot_id == other_id)

# Now we can do some propagation to verify some values.
# For instance, if x is 1, then A should be true as well.
# Note that you must give the ID and not the alias when getting its value
model.propagate_downstream({"x": 1+1j}).get(id) == 1+1j 

# Or maybe we want to find a combination that trying
# to not include x and even more not y, such that (x or y or z) must be true.
# Note that we can send in multiple "objectives" and there's one solution returned
# by each objective
solutions = model.solve([{"x": -1, "y": -2}], {id: 1})