Primitive Logic Directed Acyclic Graph
The Primitive Logic Directed Acyclic Graph (PL-DAG) is a data structure based on a Directed Acyclic Graph (DAG) where each node represents a logical relationship and leaf nodes correspond to literals. Each node encapsulates information about how its incoming nodes or leaves are logically related. For example, a node might represent an AND operation, meaning that if it evaluates to true, all of its incoming nodes or leaves must also evaluate to true.
How It Works
Each composite (node) is represented as a linear inequality of the form A = a + b + c >= 0. A primitive (leaf) is a named alias connected to a literal. A literal is represented as a complex number with two values: -1+3j indicates the lowest value a variable can take (-1) and the highest value (+3). For example, a boolean primitive has the literal value 1j since it can only be 0 or 1. A primitive with value 52j (representing weeks, for instance) could take any discrete value between 0 and 52, though it is expressed only by its bounds.
Example
from pldag import PLDAG
# Initialize model
model = PLDAG()
# Set x, y, and z as boolean variables in the model
model.set_primitives("xyz")
# Create a simple AND operation connected to alias "A"
# This is equivalent to A = x + y + z - 3 >= 0
# The ID for this proposition is returned. We can also assign an alias to it.
id_ref = model.set_and(["x","y","z"], alias="A")
# Later, if we forget the ID, we can retrieve it like this
id_ref_again = model.id_from_alias("A")
assert id_ref == id_ref_again
# When all x, y, and z are set to 1, we expect `id_ref` to be 1+1j
assert model.propagate({"x": 1+1j, "y": 1+1j, "z": 1+1j}).get(id_ref) == 1+1j
# When not all are set, we get 1j (meaning the model doesn't know whether it's true or false)
assert model.propagate({"x": 1+1j, "y": 1+1j, "z": 1j}).get(id_ref) == 1j
# However, if we know that any variable is not set (equal to 0), the model knows the composite is false (0j)
assert model.propagate({"x": 1+1j, "y": 1+1j, "z": 0j}).get(id_ref) == 0j
Using a Solver
There is no built-in solver, but PL-DAG supports integration with existing solvers. Install the package:
pip install pldag
Then use it as follows:
from pldag import Solver
# Maximize [x=1, y=0, z=0] such that the model's rules hold and `id_ref` must be true
solution = next(iter(model.solve(objectives=[{"x": 1}], assume={id_ref: 1+1j}, solver=Solver.DEFAULT)))
# Since x=1 and `id_ref` must be true (i.e., all(x,y,z) must be true),
# we expect all variables to be set
assert solution.get(id_ref) == 1+1j
Release files for pldag 0.10.17
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | |
|---|---|---|---|
| pldag-0.10.17.tar.gz | 22.6 kB | Details |
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Reset |
|---|---|---|---|---|
| pldag-0.10.17-py3-none-any.whl | Python 3 | none | any | Details |
Total release size: 45.2 kB
Release files / pldag-0.10.17.tar.gz
| Download URL | pldag-0.10.17.tar.gz |
|---|---|
| Size | 22.6 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
c1d29584f093c0a3fe481a8e04711f317a2285d39fbb5f8846dd9891cba3ddc2
|
|
BLAKE2b-256 checksum How to use checksums |
476e074899b0ad1ec6d61a56b5d7c211649c1d4902dfc6aadfd3fcfe0e0d3133
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.8.4 CPython/3.10.15 Darwin/24.6.0
|
Release files / pldag-0.10.17-py3-none-any.whl
| Download URL | pldag-0.10.17-py3-none-any.whl |
|---|---|
| Size | 22.6 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
2804c55f132cae59d35b1c99bb842f425d5ec8f2c27bba641191da9b7ab39b0d
|
|
BLAKE2b-256 checksum How to use checksums |
255528c7c2f40879af445d4709aa50d97e08e89e8e3d84d3e2d596031c61b350
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
No |
| Uploaded via |
poetry/1.8.4 CPython/3.10.15 Darwin/24.6.0
|
