fca-algorithms 1.0.2

FCA basic algorithms

FCA algorithms

This is a module providing a set of commonly used algorithms in FCA, RCA, and some of its variants. Its general intention is to provide an easy to use API so that it's easier to create other programs using these algorithms. The main algorithm that calculates formal concepts is inclose, and, in this version, it is implemented in C++. Considering this, the API is expected to behave somewhat acceptably.

API Reference

CLI

FCA

Plot a hasse diagram from a context

fca_cli -c input.csv --show_hasse

The context is expected to be a csv with the following format

name	attr1	attr2
obj1	x
obj2		x
obj3	x	x
obj4

Output files

fca_cli -c input.csv --show_hasse --output_dir path/to/folder/ 

Will create two files, one representing the hasse graph, the other one with a concept for each line. The line is the index in the hasse graph.

RCA

To plot the hasse diagrams of the contexts 1 and 2 after applying RCA with exists

fca_cli -k context_1.csv context_2.csv -r relation_1_2.csv relation_2_1.csv --show_hasse

to specify operator

fca_cli -k context_1.csv context_2.csv -r relation_1_2.csv relation_2_1.csv --show_hasse -o forall

FCA utils

Module for FCA basics such as retrieving concepts, drawing a hasse diagram, etc

Getting formal concepts

In batch

from fca.api_models import Context, Concept

c = Context(O : List[str], A : List[str], I : List[List[int]])
concepts = c.get_concepts(c) List[Concept]

Incrementally

from fca.api_models import IncLattice

l = IncLattice(attributes=['a', 'b', 'c', 'd'])
l.add_intent('o1', [0, 2])  # numbers are the indices of the attributes
l.add_intent('o2', [1, 2]) 
.
.
.

Getting association rules

from fca.api_models import Context

c = Context(O, A, I)
c.get_association_rules(min_support=0.4, min_confidence=1)

Drawing hasse diagram

from fca.plot.plot import plot_from_hasse
from fca.api_models import Context


k = Context(O, A, I)
k.get_lattice().plot()
# plot receives a number of kwargs such as print_latex=True|False


l = IncLattice(attributes=['a', 'b', 'c', 'd'])
l.add_intent('o1', [0, 2])  # numbers are the indices of the attributes
l.add_intent('o2', [1, 2])
.
.
.
l.plot()

Contributors

• Ramshell (Nicolas Leutwyler)