Add ToddCoxeter alphabet order example (#448)

Author: Joseph-Edwards

examples/alphabet-order/M12-distribution.png

@@ -0,0 +1,157 @@
+# pylint: disable=redefined-outer-name, redefined-loop-name, import-error
+"""In this file, we demonstrate the impact of changing the order of a ToddCoxeter alphabet"""
+
+from collections.abc import Callable
+from datetime import timedelta
+from itertools import permutations
+from math import factorial
+from time import perf_counter
+
+import matplotlib.pyplot as plt
+import numpy as np
+import seaborn as sns
+from numpy.typing import NDArray
+
+from libsemigroups_pybind11 import (
+    Presentation,
+    ReportGuard,
+    ToddCoxeter,
+    congruence_kind,
+    presentation,
+)
+from libsemigroups_pybind11.words import parse
+
+
+def generate_alphabet_order_data(
+    p: Presentation, f: Callable[[ToddCoxeter], float], t: int | None = None, repeats: int = 5
+) -> tuple[NDArray, NDArray]:
+    """For each permutation of the alphabet of <p>, construct a ToddCoxeter. On
+    this instance, call the function <f> both before and after running, and
+    record the difference between these two values. If <t> is not specified, the
+    runner will be run using `run()`. Otherwise, the runner will be run for <t>
+    seconds.
+
+    Return the list of alphabets and the corresponding list of data.
+    """
+
+    alphabet = p.alphabet()
+    n = factorial(len(alphabet))
+    alphabets = np.empty((n,), dtype=type(alphabet))
+    data = np.empty((n, repeats), dtype=float)
+
+    for i, new_alphabet in enumerate(permutations(alphabet)):
+        if isinstance(alphabet, list):
+            new_alphabet = list(new_alphabet)
+        else:
+            new_alphabet = "".join(new_alphabet)
+
+        alphabets[i] = new_alphabet
+        p.alphabet(new_alphabet)
+
+        print(f"alphabet: {new_alphabet}")
+
+        for j in range(repeats):
+            tc = ToddCoxeter(congruence_kind.twosided, p)
+            # We record an initial value so the that the time taken can be measured,
+            # if desired.
+            initial = f(tc)
+            if t is None:
+                tc.run()
+            else:
+                tc.run_for(timedelta(seconds=t))
+
+            final = f(tc)
+            print(f"run {j + 1}: {final - initial}")
+            data[i, j] = final - initial
+
+    return alphabets, data
+
+
+ReportGuard(False)
+
+#########################################################################
+# Mathieu Group 12
+#########################################################################
+
+p = Presentation("abAB")
+p.contains_empty_word(True)
+presentation.add_inverse_rules(p, "ABab")
+presentation.add_rule(p, parse("a^2"), "")
+presentation.add_rule(p, parse("b^3"), "")
+presentation.add_rule(p, parse("(ab)^11"), "")
+presentation.add_rule(p, parse("(a,b)^6"), "")
+presentation.add_rule(p, parse("(ababaB)^6"), "")
+
+alphabets, times = generate_alphabet_order_data(p, lambda tc: perf_counter())
+times = times.mean(axis=1)
+
+# Plot the alphabet vs time
+ax = sns.barplot(x=alphabets, y=times)
+ax.set(xlabel="alphabet", ylabel="time", title="Mathieu 12")
+plt.show()
+
+# Plot the time distribution
+ax = sns.displot(x=times, kde=True)
+ax.set(xlabel="time", title="Mathieu Group 12")
+plt.show()
+
+#########################################################################
+# Mathieu Group 22
+#########################################################################
+
+p = Presentation("abAB")
+p.contains_empty_word(True)
+presentation.add_inverse_rules(p, "ABab")
+presentation.add_rule(p, parse("a^2"), "")
+presentation.add_rule(p, parse("b^4"), "")
+presentation.add_rule(p, parse("(ab)^11"), "")
+presentation.add_rule(p, parse("(ab^2)^5"), "")
+presentation.add_rule(p, parse("(a,bab)^3"), "")
+presentation.add_rule(p, parse("(ababaB)^5"), "")
+
+alphabets, times = generate_alphabet_order_data(p, lambda tc: perf_counter())
+times = times.mean(axis=1)
+
+# Plot the time distribution
+ax = sns.displot(x=times, kde=True)
+ax.set(xlabel="time", title="Mathieu Group 22")
+plt.show()
+
+#########################################################################
+# Shutov's partial transformation monoid of degree 5
+#########################################################################
+
+p = presentation.examples.partial_transformation_monoid_Shu60(5)
+alphabets, times = generate_alphabet_order_data(p, lambda tc: perf_counter())
+times = times.mean(axis=1)
+
+# Plot the time distribution
+ax = sns.displot(x=times, kde=True)
+ax.set(xlabel="time", title="Partial transformation monoid degree 5")
+plt.show()
+
+#########################################################################
+# Example 6.6 in Sims
+#########################################################################
+
+p = Presentation("abcd")
+presentation.add_rule(p, "aa", "a")
+presentation.add_rule(p, "ba", "b")
+presentation.add_rule(p, "ab", "b")
+presentation.add_rule(p, "ca", "c")
+presentation.add_rule(p, "ac", "c")
+presentation.add_rule(p, "da", "d")
+presentation.add_rule(p, "ad", "d")
+presentation.add_rule(p, "bb", "a")
+presentation.add_rule(p, "cd", "a")
+presentation.add_rule(p, "ccc", "a")
+presentation.add_rule(p, "bcbcbcbcbcbcbc", "a")
+presentation.add_rule(p, "bcbdbcbdbcbdbcbdbcbdbcbdbcbdbcbd", "a")
+
+alphabets, times = generate_alphabet_order_data(p, lambda tc: perf_counter())
+times = times.mean(axis=1)
+
+# Plot the time distribution
+ax = sns.displot(x=times, kde=True)
+ax.set(xlabel="time", title="Example 6.6 in Sims")
+plt.show()