Running with python3.10 produces typeError

[DrLynch]

Greetings, when I run a basic example of the zebra problem with miniKanren 1.0.3 under python3.10 I receive a type error. The minimal example is:

  def left(X, Y, list):
     return kanren.membero((X, Y), zip(list, list[1:]))

  def next(X, Y, list):
     return kanren.conde([left(Y, X, list)], [left(X, Y, list)])

   Houses = kanren.var()

   Zebra_Rules = kanren.lall(
      (kanren.eq, (kanren.var(), kanren.var(), kanren.var(), kanren.var(), kanren.var()), Houses),

    (kanren.membero, ('Englishman', kanren.var(), kanren.var(), kanren.var(), 'red'), Houses),
    (kanren.membero, ('Swede', kanren.var(), kanren.var(), 'dog', kanren.var()), Houses),
    (kanren.membero, ('Dane', kanren.var(), 'tea', kanren.var(), kanren.var()), Houses),
    (left,(kanren.var(), kanren.var(), kanren.var(), kanren.var(), 'green'),
    (kanren.var(), kanren.var(), kanren.var(), kanren.var(), 'white'), Houses),
    (kanren.membero, (kanren.var(), kanren.var(), 'coffee', kanren.var(), 'green'), Houses),
    (kanren.membero, (kanren.var(), 'Pall Mall', kanren.var(), 'birds', kanren.var()), Houses),
    (kanren.membero, (kanren.var(), 'Dunhill', kanren.var(), kanren.var(), 'yellow'), Houses),
    (kanren.eq,(kanren.var(), kanren.var(), (kanren.var(), kanren.var(), 'milk', kanren.var(), kanren.var()), kanren.var(), kanren.var()), Houses),
    (kanren.eq,(('Norwegian', kanren.var(), kanren.var(), kanren.var(), kanren.var()), kanren.var(), kanren.var(), kanren.var(), kanren.var()), Houses),
    (next,(kanren.var(), 'Blend', kanren.var(), kanren.var(), kanren.var()),
    (kanren.var(), kanren.var(), kanren.var(), 'cats', kanren.var()), Houses),
    (next,(kanren.var(), 'Dunhill', kanren.var(), kanren.var(), kanren.var()),
    (kanren.var(), kanren.var(), kanren.var(), 'horse', kanren.var()), Houses),
    (kanren.membero, (kanren.var(), 'Blue Master', 'beer', kanren.var(), kanren.var()), Houses),
    (kanren.membero, ('German', 'Prince', kanren.var(), kanren.var(), kanren.var()), Houses),
    (next,('Norwegian', kanren.var(), kanren.var(), kanren.var(), kanren.var()),
    (kanren.var(), kanren.var(), kanren.var(), kanren.var(), 'blue'), Houses),
    (next,(kanren.var(), 'Blend', kanren.var(), kanren.var(), kanren.var()),
    (kanren.var(), kanren.var(), 'water', kanren.var(), kanren.var()), Houses),
    (kanren.membero, (kanren.var(), kanren.var(), kanren.var(), 'zebra', kanren.var()), Houses)
)

Solution = kanren.run(0, Houses, Zebra_Rules)

Executing this code in the interpreter or a file yields the following error:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/local/share/anaconda3/install/lib/python3.10/site-packages/kanren/core.py", line 202, in run
    return tuple(results)
  File "/usr/local/share/anaconda3/install/lib/python3.10/site-packages/kanren/core.py", line 83, in lconj_seq_goal
    yield from reduce(bind, _goals, g0(S))
  File "/usr/local/share/anaconda3/install/lib/python3.10/site-packages/kanren/core.py", line 83, in lconj_seq_goal
    yield from reduce(bind, _goals, g0(S))
TypeError: 'tuple' object is not callable

[DrLynch]

This problem persists with the latest GitHub sources as well:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/username/Teaching/AIAcademy/Intro-to-AI-Cohort-3-2022/WorkshopMaterials/AIAWorkshopRepo/Week-04/Code/kanren/kanren/core.py", line 238, in run
    return tuple(results)
  File "/home/username/Teaching/AIAcademy/Intro-to-AI-Cohort-3-2022/WorkshopMaterials/AIAWorkshopRepo/Week-04/Code/kanren/kanren/core.py", line 104, in lconj_seq_goal
    yield from reduce(bind, _goals, g0(S))
  File "/home/username/Teaching/AIAcademy/Intro-to-AI-Cohort-3-2022/WorkshopMaterials/AIAWorkshopRepo/Week-04/Code/kanren/kanren/core.py", line 104, in lconj_seq_goal
    yield from reduce(bind, _goals, g0(S))
TypeError: 'tuple' object is not callable

[brandonwillard]

It looks like you're using the tuple-based approach to defining goals from LogPy (e.g. (eq, a, b)). When using this fork, those tuples can be replaced by actual calls on the goals (e.g. eq(a, b)).

[brandonwillard]

The rewritten zebra example is here and it appears to work in Python 3.10, so I'll close this for now.