Changeset 1306 for framspy/FramsticksLib.py

Timestamp:

05/13/24 03:59:10 (8 weeks ago)

Author:

Maciej Komosinski

Message:

Introduced symbolic names for dissimilarity estimation methods

File:

• framspy/FramsticksLib.py (modified) (9 diffs)

framspy/FramsticksLib.py

@@ -1 +1 @@
 from typing import List  # to be able to specify a type hint of list(something)
+from enum import Enum, auto, unique
 import json
 import sys, os
@@ -5 +6 @@
 import numpy as np
 import frams
+
+
+@unique
+class DissimMethod(Enum):  # values assigned to fields are irrelevant, hence auto()
+        GENE_LEVENSHTEIN = auto()  # genetic Levenshtein distance
+        PHENE_STRUCT_GREEDY = auto()  # phenetic, graph structure, fast but approximate
+        PHENE_STRUCT_OPTIM = auto()  # phenetic, graph structure, slower for complex creatures but optimal
+        PHENE_DESCRIPTORS = auto()  # phenetic, shape descriptors
+        PHENE_DENSITY_COUNT = auto()  # phenetic, density distribution, count of samples
+        PHENE_DENSITY_FREQ = auto()  # phenetic, density distribution, frequency of count of samples
+        FITNESS = auto()  # fitness value
+
 
 
@@ -67 +80 @@
                 for simfile in self.EVALUATION_SETTINGS_FILE:
                         ec = frams.MessageCatcher.new()  # catch potential errors, warnings, messages - just to detect if there are ERRORs
-                        ec.store = 2;  # store all, because they are caught by MessageCatcher and will not appear in output (which we want)
+                        ec.store = 2  # store all, because they are caught by MessageCatcher and will not appear in output (which we want)
                         frams.Simulator.ximport(simfile, 4 + 8 + 16)
                         ec.close()
@@ -75 +88 @@
 
 
-        def getSimplest(self, genetic_format) -> str:
+        def getSimplest(self, genetic_format: str) -> str:
                 return frams.GenMan.getSimplest(genetic_format).genotype._string()
 
@@ -134 +147 @@
                 if not self.PRINT_FRAMSTICKS_OUTPUT:
                         ec = frams.MessageCatcher.new()  # mute potential errors, warnings, messages
-                        ec.store = 2;  # store all, because they are caught by MessageCatcher and will not appear in output
+                        ec.store = 2  # store all, because they are caught by MessageCatcher and will not appear in output
 
                 frams.GenePools[0].clear()
@@ -192 +205 @@
 
 
-        def dissimilarity(self, genotype_list: List[str], method: int) -> np.ndarray:
-                """
-                        :param method: -1 = genetic Levenshtein distance; 0, 1, 2 = phenetic dissimilarity (SimilMeasureGreedy, SimilMeasureHungarian, SimilMeasureDistribution); -2, -3 = phenetic density distribution (count, frequency).
+        def dissimilarity(self, genotype_list: List[str], method: DissimMethod) -> np.ndarray:
+                """
+                        :param method, see DissimMethod.
                         :return: A square array with dissimilarities of each pair of genotypes.
                 """
@@ -206 +219 @@
                 square_matrix = np.zeros((n, n))
 
-                if method in (0, 1, 2):  # Framsticks phenetic dissimilarity methods
-                        frams.SimilMeasure.simil_type = method
+                if method in (DissimMethod.PHENE_STRUCT_GREEDY, DissimMethod.PHENE_STRUCT_OPTIM, DissimMethod.PHENE_DESCRIPTORS):  # Framsticks phenetic dissimilarity methods
+                        frams.SimilMeasure.simil_type = 0 if method == DissimMethod.PHENE_STRUCT_GREEDY else 1 if method == DissimMethod.PHENE_STRUCT_OPTIM else 2
                         genos = []  # prepare an array of Geno objects so that we don't need to convert raw strings to Geno objects all the time in loops
                         for g in genotype_list:
@@ -215 +228 @@
                                 for j in range(n):  # maybe calculate only one triangle if you really need a 2x speedup
                                         square_matrix[i][j] = frams_evaluateDistance(genos[i], genos[j])._double()
-                elif method == -1:
+                elif method == DissimMethod.GENE_LEVENSHTEIN:
                         import Levenshtein
                         for i in range(n):
                                 for j in range(n):  # maybe calculate only one triangle if you really need a 2x speedup
                                         square_matrix[i][j] = Levenshtein.distance(genotype_list[i], genotype_list[j])
-                elif method in (-2, -3):
+                elif method in (DissimMethod.PHENE_DENSITY_COUNT, DissimMethod.PHENE_DENSITY_FREQ):
                         if self.dissim_measure_density_distribution is None:
                                 from dissimilarity.density_distribution import DensityDistribution
                                 self.dissim_measure_density_distribution = DensityDistribution(frams)
-                        self.dissim_measure_density_distribution.frequency = (method == -3)
+                        self.dissim_measure_density_distribution.frequency = (method == DissimMethod.PHENE_DENSITY_FREQ)
                         square_matrix = self.dissim_measure_density_distribution.getDissimilarityMatrix(genotype_list)
                 else:
-                        raise ValueError("Don't know what to do with dissimilarity method = %d" % method)
+                        raise ValueError("Don't know what to do with dissimilarity method = %s" % method)
 
                 for i in range(n):
@@ -381 +394 @@
         offspring = framsLib.crossOver(parent1, parent2)
         print("\tCrossover (Offspring):", offspring)
-        print('\tDissimilarity of Parent1 and Offspring:', framsLib.dissimilarity([parent1, offspring], 1)[0, 1])
+        print('\tDissimilarity of Parent1 and Offspring:', framsLib.dissimilarity([parent1, offspring], DissimMethod.PHENE_STRUCT_OPTIM)[0, 1])
         print('\tPerformance of Offspring:', framsLib.evaluate([offspring]))
         print('\tValidity (genetic) of Parent1, Parent 2, and Offspring:', framsLib.isValid([parent1, parent2, offspring]))