Merge branch 'stable-1.4'

james-d-mitchell · james-d-mitchell · commit d3769091a919 · 2026-03-20T12:40:02.000Z
diff --git a/CITATION.cff b/CITATION.cff
@@ -41,5 +41,5 @@ abstract: >-
   A Python package that wraps the functionality of the C++
   library libsemigroups.
 license: GPL-3.0+
-version: 1.4.1
-date-released: "2026-03-17"
+version: 1.4.2
+date-released: "2026-03-20"
diff --git a/README.md b/README.md
@@ -57,7 +57,7 @@ presented, semigroups and monoids. The main algorithms implemented in
 [Efficient Testing of Equivalence of Words in a Free Idempotent Semigroup]: https://link.springer.com/chapter/10.1007/978-3-642-11266-9_55
 [Applications of automata theory to presentations of monoids and inverse monoids]: https://digitalcommons.unl.edu/dissertations/AAI8803771/
 [Todd-Coxeter algorithm]: https://en.wikipedia.org/wiki/Todd%E2%80%93Coxeter_algorithm
-[The Todd-Coxeter algorithm for semigroups and monoids]: https://doi.org/10.1007/s00233-024-10431-z
+[The Todd-Coxeter algorithm for semigroups and monoids]: https://doi.org/10.1007/s00233-024-1.4.2-z
 [Schreier-Sims algorithm]: https://en.wikipedia.org/wiki/Schreier%E2%80%93Sims_algorithm
 [Expository Slides]: https://www.irif.fr/~jep/PDF/Exposes/StAndrews.pdf
 [Semigroupe 2.01]: https://www.irif.fr/~jep/Logiciels/Semigroupe2.0/semigroupe2.html
@@ -150,10 +150,10 @@ uv pip install .
 To build `libsemigroups_pybind11` from a release archive:
 
 ```console
-curl -L -O https://github.com/libsemigroups/libsemigroups_pybind11/releases/latest/download/libsemigroups_pybind11-1.4.1.tar.gz
-tar -xf libsemigroups_pybind11-1.4.1.tar.gz
-rm -f libsemigroups_pybind11-1.4.1.tar.gz
-cd libsemigroups_pybind11-1.4.1
+curl -L -O https://github.com/libsemigroups/libsemigroups_pybind11/releases/latest/download/libsemigroups_pybind11-1.4.2.tar.gz
+tar -xf libsemigroups_pybind11-1.4.2.tar.gz
+rm -f libsemigroups_pybind11-1.4.2.tar.gz
+cd libsemigroups_pybind11-1.4.2
 pip install .
 ```
 
diff --git a/docs/source/changelog-v1.rst b/docs/source/changelog-v1.rst
@@ -8,6 +8,18 @@
 Changelog - version 1
 =====================
 
+v1.4.2 (released 20/03/2026)
+----------------------------
+
+This a minor release which contains a number of minor improvements to the
+documentation, and some fixes to the return types of the ``ToddCoxeter``
+methods related to lookahead and lookbehind.
+
+* Todd coxeter rtypes by @james-d-mitchell in
+  https://github.com/libsemigroups/libsemigroups_pybind11/pull/406
+* Doc fixes by @james-d-mitchell in
+  https://github.com/libsemigroups/libsemigroups_pybind11/pull/407
+
 v1.4.1 (released 17/03/2026)
 ----------------------------
 
@@ -20,15 +32,15 @@ v1.4.0 (released 25/02/2026)
 
 This is a minor release including some improvements and new features:
 
-* Update to libsemigroups v3.5.0 and add the new features it contains.
+* Update to libsemigroups_ v3.5.0 and add the new features it contains.
 * transf: future proof ``__getitem__`` by @james-d-mitchell in https://github.com/libsemigroups/libsemigroups_pybind11/pull/382
 * todd-coxeter: update for changes in libsemigroups by @james-d-mitchell in https://github.com/libsemigroups/libsemigroups_pybind11/pull/388
 
 
 v1.3.0 (released 18/01/2026)
 ----------------------------
 
-* Update to libsemigroups v3.4.0 and add features introduced there by
+* Update to libsemigroups_ v3.4.0 and add features introduced there by
   @james-d-mitchell in
   https://github.com/libsemigroups/libsemigroups_pybind11/pull/380
 
@@ -82,7 +94,7 @@ version 3 of libsemigroups_. The extent of the changes made in this release
 means that it is likely that any code written with earlier versions of
 ``libsemigroups_pybin11`` will no longer work.
 
-The structure of ``libsemigroups_pybind11`` is very tightly linked to the the
+The structure of ``libsemigroups_pybind11`` is very tightly linked to the
 structure of libsemigroups_. Therefore, some of the differences between
 v0.10.1 and v1.0.0 of ``libsemigroups_pyind11`` will be related to the
 differences between v2 and v3 of libsemigroups_, such as changes to class names
diff --git a/docs/source/data-structures/presentations/to-inverse-present.rst b/docs/source/data-structures/presentations/to-inverse-present.rst
@@ -113,7 +113,7 @@ the type of words specified in *rtype*:
        objects of type :any:`InversePresentation` must be converted from one
        type to another sometimes, but not other times.
 
-If the alphabet of of *ip* is :math:`\{a_0, a_1, \dots a_{n-1}\}`, where each
+If the alphabet of *ip* is :math:`\{a_0, a_1, \dots a_{n-1}\}`, where each
 letter is of type ``str``, then the conversion from one type to another is
 :math:`a_i \mapsto` ``human_readable_index(a_i)``. Conversely, if each letter is
 of type ``list[int]``, then the conversion from one type to another is
diff --git a/docs/source/data-structures/presentations/to-present.rst b/docs/source/data-structures/presentations/to-present.rst
@@ -66,7 +66,7 @@ the type of words specified in *rtype*:
        presentations must be converted from one type to another sometimes, but
        not other times.
 
-If the alphabet of of *p* is :math:`\{a_0, a_1, \dots a_{n-1}\}`, where each
+If the alphabet of *p* is :math:`\{a_0, a_1, \dots a_{n-1}\}`, where each
 letter is of type ``str``, then the conversion from one type to another is
 :math:`a_i \mapsto` ``human_readable_index(a_i)``. Conversely, if each letter is
 of type ``list[int]``, then the conversion from one type to another is
diff --git a/docs/source/install.rst b/docs/source/install.rst
@@ -101,10 +101,10 @@ To build ``libsemigroups_pybind11`` from a release archive:
 
 ::
 
-    $ curl -L -O https://github.com/libsemigroups/libsemigroups_pybind11/releases/latest/download/libsemigroups_pybind11-1.4.1.tar.gz
-    $ tar -xf libsemigroups_pybind11-1.4.1.tar.gz
-    $ rm -f libsemigroups_pybind11-1.4.1.tar.gz
-    $ cd libsemigroups_pybind11-1.4.1
+    $ curl -L -O https://github.com/libsemigroups/libsemigroups_pybind11/releases/latest/download/libsemigroups_pybind11-1.4.2.tar.gz
+    $ tar -xf libsemigroups_pybind11-1.4.2.tar.gz
+    $ rm -f libsemigroups_pybind11-1.4.2.tar.gz
+    $ cd libsemigroups_pybind11-1.4.2
     $ pip install .
 
 Building the documentation
diff --git a/docs/source/main-algorithms/froidure-pin/to-froidure-pin.rst b/docs/source/main-algorithms/froidure-pin/to-froidure-pin.rst
@@ -101,7 +101,7 @@ This function throws a :any:`LibsemigroupsError` if the
 .. warning::
 
     The returned :any:`FroidurePin` instance is always infinite, and so any
-    calls to any member functions that that trigger a full enumeration will
+    calls to any member functions that trigger a full enumeration will
     never terminate (or they will when your computer kills the process because
     it has run out of memory).
 
diff --git a/docs/source/main-algorithms/knuth-bendix/to-knuth-bendix.rst b/docs/source/main-algorithms/knuth-bendix/to-knuth-bendix.rst
@@ -100,9 +100,9 @@ following values for *args*:
 Additionally, specify one of the following for *rtype*:
 
     - ``(KnuthBendix, 'RewriteTrie')`` for constructing a :any:`KnuthBendix`
-      with the the ``RewriteTrie'`` rewriter.
+      with the ``RewriteTrie'`` rewriter.
     - ``(KnuthBendix, 'RewriteFromLeft')`` for constructing a :any:`KnuthBendix`
-      with the the ``RewriteFromLeft'`` rewriter.
+      with the ``RewriteFromLeft'`` rewriter.
 
 This function converts a :any:`ToddCoxeter` object *tc* to a :any:`KnuthBendix`
 object with the rewriter as specified above, using
diff --git a/docs/source/main-algorithms/todd-coxeter/class/init.rst b/docs/source/main-algorithms/todd-coxeter/class/init.rst
@@ -10,7 +10,7 @@
 Constructors + initializers
 ===========================
 
-This page documents the constructors and initialisers for the :any:`ToddCoxeter`
+This page documents the constructors and initializers for the :any:`ToddCoxeter`
 class.
 
 Every constructor has a matching init function with the same signature that can
diff --git a/docs/source/main-algorithms/todd-coxeter/class/word_to_index.rst b/docs/source/main-algorithms/todd-coxeter/class/word_to_index.rst
@@ -11,7 +11,7 @@ Word to class index
 ===================
 
 This page contains documentation for the member functions of :any:`ToddCoxeter`
-that can be used to convert a word into the index of congruence class.
+that can be used to convert a word into the index of a congruence class.
 
 .. automethod:: ToddCoxeter.current_index_of
 
diff --git a/docs/source/main-algorithms/todd-coxeter/to-todd-coxeter.rst b/docs/source/main-algorithms/todd-coxeter/to-todd-coxeter.rst
@@ -56,7 +56,7 @@ This function converts the :any:`FroidurePin` object *fpb* into a
 either the :any:`FroidurePin.left_cayley_graph` or the
 :any:`FroidurePin.right_cayley_graph` of *fpb*).
 
-This returned :any:`ToddCoxeter` object represents the trivial congruence over
+The returned :any:`ToddCoxeter` object represents the trivial congruence over
 the semigroup defined by *fpb*.
 
 This will throw a :any:`LibsemigroupsError` if *wg* is not the
@@ -111,7 +111,7 @@ This function converts the :any:`KnuthBendix` object *kb* into a
 :any:`ToddCoxeter` object using the right Cayley graph of the semigroup
 represented by *kb*.
 
-This returned :any:`ToddCoxeter` object represents the trivial congruence over
+The returned :any:`ToddCoxeter` object represents the trivial congruence over
 the semigroup defined by *kb*.
 
 This will throw a :any:`LibsemigroupsError` if either:
diff --git a/src/todd-coxeter-impl.cpp b/src/todd-coxeter-impl.cpp
@@ -83,7 +83,7 @@ The valid values are :
 .. py:attribute:: strategy.R_over_C
   :value: <strategy.R_over_C: 3>
 
-  This strategy is meant to mimic the ACE strategy R/C. The HLT strategy is run until the first lookahead is triggered (when :the number of nodes active is at least :any:`lookahead_next`). A full lookahead is then performed, and then the CR strategy is used.
+  This strategy is meant to mimic the ACE strategy R/C. The HLT strategy is run until the first lookahead is triggered (when the number of nodes active is at least :any:`lookahead_next`). A full lookahead is then performed, and then the CR strategy is used.
 
 .. py:attribute:: strategy.Cr
   :value: <strategy.Cr: 4>
@@ -152,7 +152,7 @@ The valid values are :
   The lookahead will be done in HLT style by following the paths labelled by every relation from every node in the range specified by :any:`lookahead_extent.full` or :any:`lookahead_extent.partial`.
 
 .. py:attribute:: lookahead_style.felsch
-  :value: <lookahead_style.hlt: 0>
+  :value: <lookahead_style.felsch: 1>
 
   The lookahead will be done in Felsch style where every edge is considered in every path labelled by a relation in which it occurs.
 
@@ -196,12 +196,12 @@ The valid values are:
 .. py:attribute:: def_policy.purge_from_top
   :value: <def_policy.purge_from_top: 1>
 
-  If the definition stack has size :any:`def_max` and a new definition is generated, then definitions with dead source node are are popped from the top of the stack (if any).
+  If the definition stack has size :any:`def_max` and a new definition is generated, then definitions with dead source node are popped from the top of the stack (if any).
 
 .. py:attribute:: def_policy.purge_all
   :value: <def_policy.purge_all: 2>
 
-  If the definition stack has size :any:`def_max` and a new definition is generated, then definitions with dead source node are are popped from the entire of the stack (if any).
+  If the definition stack has size :any:`def_max` and a new definition is generated, then definitions with dead source node are popped from the entire of the stack (if any).
 
 .. py:attribute:: def_policy.discard_all_if_no_space
   :value: <def_policy.discard_all_if_no_space: 3>
@@ -475,7 +475,7 @@ This function can be used to specify how to handle definitions. For details see
         R"pbdoc(
 :sig=(self: ToddCoxeter) -> ToddCoxeter.options.def_version:
 
-The current value of the definition policy setting.
+The current value of the definition version setting.
 
 :returns:
    The current value of the setting.
@@ -657,7 +657,7 @@ The default value of this setting is ``10 ** 5``.
 Set the size of a large collapse.
 
 This function can be used to set what should be considered a "large"
-collapse.By default when processing coincidences nodes are merged in the word
+collapse. By default when processing coincidences nodes are merged in the word
 graph one pair at a time, and the in-neighbours of the surviving node are
 updated at the same time. If the number of coincidences is large, then it might
 be that a pair of nodes are merged at one step, then the surviving node is
@@ -969,7 +969,7 @@ lookaheads to be stopped early if the number of nodes being killed is too small
 want to stop the lookahead early, since lookaheads take some time but may not
 result in many nodes being killed).
 
-The default value is `0.01`
+The default value is ``0.01``.
 
 :param val: the proportion of active nodes.
 :type val: float
@@ -1165,7 +1165,7 @@ Set whether or not to perform an HLT-style push of the defining relations at the
 If a :any:`ToddCoxeter` instance is defined over a finitely presented semigroup
 or monoid and the Felsch strategy is being used, it can be useful to follow all
 the paths from the identity labelled by the underlying relations. This setting
-specifies whether or not to do this.The default value of this setting is
+specifies whether or not to do this. The default value of this setting is
 ``False``.
 
 :param val: the boolean value.
@@ -1323,25 +1323,22 @@ The return value of this function indicates the following:
 - :any:`Order.none` implies that no standardization has been performed
   and:
 
-   - the return value of :any:`reduce` will essentially arbitrary;
-   - the return values of :any:`todd_coxeter.normal_forms` or
-     :any:`todd_coxeter.normal_forms` will be essentially arbitrary;
+   - the return value of :any:`reduce` will be essentially arbitrary;
+   - the return values of :any:`todd_coxeter.normal_forms` will be essentially arbitrary;
    - the classes of the congruence will be indexed in an arbitrary order;
 
 - :any:`Order.shortlex` implies that:
 
    - the return value of :any:`reduce` will be the short-lex least word
      belonging to a given congruence class;
-   - the return values of :any:`todd_coxeter.normal_forms` and
-     :any:`todd_coxeter.normal_forms` will be in
+   - the return values of :any:`todd_coxeter.normal_forms` will be in
      short-lex order;
    - the classes of the congruence will be indexed in short-lex order on
      the short-lex least word;
 
 - :any:`Order.lex` implies that:
 
-   - the return values of :any:`todd_coxeter.normal_forms` and
-     :any:`todd_coxeter.normal_forms` will be ordered lexicographically.
+   - the return values of :any:`todd_coxeter.normal_forms` will be ordered lexicographically.
    - the return values of :any:`reduce` and the indexes of class are
      essentially arbitrary because there is not necessarily a
      lexicographically least word in every class;
@@ -1350,8 +1347,7 @@ The return value of this function indicates the following:
 
    - the return value of :any:`reduce` will be the recursive path least
      word belonging to a given congruence class;
-   - the return values of :any:`todd_coxeter.normal_forms` and
-     :any:`todd_coxeter.normal_forms` will be ordered by the
+   - the return values of :any:`todd_coxeter.normal_forms` will be ordered by the
      recursive path order;
    - the classes of the congruence will be indexed in recursive path order
      on the recursive path least word.
@@ -1375,7 +1371,7 @@ Get the word graph after performing a full congruence enumeration.
 In some sense, the purpose of the Todd-Coxeter algorithm is to produce a
 :any:`WordGraph` of the action of a set of generators on the classes of a
 congruence. This function can be used to obtain that
-:any:`WordGraph` instance. This function triggers a full enumeration.The
+:any:`WordGraph` instance. This function triggers a full enumeration. The
 :any:`WordGraph` returned by this function may be in a rather complicated
 state. The active nodes (and nodes) will be :math:`\{0, \ldots, n - 1\}` where
 :math:`n` is the number of classes in the congruence if :any:`presentation`
diff --git a/src/todd-coxeter.cpp b/src/todd-coxeter.cpp
diff --git a/tests/test_todd_coxeter.py b/tests/test_todd_coxeter.py
