Package throws errors on Python 3.4

[aelzenaar]

@ariymarkowitz please copy the error message here thx

[ariymarkowitz]

Traceback (most recent call last):
  File "atom.py", line 8, in <module>
    from bella import cayley
  File "/home/amar630/bella/cayley.py", line 68
    warnings.warn(f"generator {n} does not seem to have non-unit determinant {det}",NonUnitDeterminantWarning)
                                                                                  ^
SyntaxError: invalid syntax

[ariymarkowitz]

Errors when I run pytest:

______________________________________ ERROR collecting tests/test_cayley.py ______________________________________
../.local/lib/python3.4/site-packages/_pytest/python.py:507: in _importtestmodule
    mod = self.fspath.pyimport(ensuresyspath=importmode)
../.local/lib/python3.4/site-packages/py/_path/local.py:704: in pyimport
    __import__(modname)
E     File "/home/amar630/bella-main/tests/test_cayley.py", line 18
E       assert matrix_almosteq_up_to_sign(orthogonal_gens[0] @ orthogonal_gens[0], mp.eye(2))
E                                                            ^
E   SyntaxError: invalid syntax
______________________________________ ERROR collecting tests/test_farey.py _______________________________________
ImportError while importing test module '/home/amar630/bella-main/tests/test_farey.py'.
Hint: make sure your test modules/packages have valid Python names.
Traceback:
tests/test_farey.py:3: in <module>
    from bella import farey, riley, cayley
E   ImportError: No module named 'bella'
______________________________________ ERROR collecting tests/test_riley.py _______________________________________
ImportError while importing test module '/home/amar630/bella-main/tests/test_riley.py'.
Hint: make sure your test modules/packages have valid Python names.
Traceback:
tests/test_riley.py:2: in <module>
    from bella import riley,farey
E   ImportError: No module named 'bella'

[aelzenaar]

Traceback (most recent call last): File "atom.py", line 8, in from bella import cayley File "/home/amar630/bella/cayley.py", line 68 warnings.warn(f"generator {n} does not seem to have non-unit determinant {det}",NonUnitDeterminantWarning) ^ SyntaxError: invalid syntax

This one is an easy fix, I just need to replace f"...{var}..." with "...{}...".format(var)

[aelzenaar]

__ ERROR collecting tests/test_cayley.py __ ../.local/lib/python3.4/site-packages/_pytest/python.py:507: in _importtestmodule mod = self.fspath.pyimport(ensuresyspath=importmode) ../.local/lib/python3.4/site-packages/py/_path/local.py:704: in pyimport import(modname) E File "/home/amar630/bella-main/tests/test_cayley.py", line 18 E assert matrix_almosteq_up_to_sign(orthogonal_gens[0] @ orthogonal_gens[0], mp.eye(2)) E ^ E SyntaxError: invalid syntax

This is PEP 465, the syntax was introduced in 3.5. I will need to take a little bit of time to think about this because in order to do precision work I use mpmath.matrix() for matrices almost everywhere and use numpy.array only when doing p-adic work. These two classes have interchangeable syntax so I can pass np.arrays with padics around and as long as I do nothing weird everything works interchangeably. But without @ I think they have different syntax for multiplication (one uses * and one has an external dot(.,.) function iirc) so I will need to think about it. (It would be an easy fix if I didn't care about p-adic support.)

[aelzenaar]

Closing as WONTFIX. The issue is that 3.4 came out almost 10 years ago and support ended for it 6 years ago, and all the very standard dependencies like pandas use more recent language features (which they are entitled to do, since all supported versions of Python have them).

Thus I will roll back the changes I made to matrix multiplication etc. since they have made the software significantly slower (instead of dropping down directly to the C++ backend I have to do some introspection to determine which kind of matrix I have which means spending a lot of CPU cycles in Python and not C, even with the significant speedups of 3.11 this is still slow as it cannot be optimised away).

[aelzenaar]

(For reference having to fall back on introspection makes the apollonian_circles.py example 11% slower on my machine.)

[aelzenaar]

today in "should have done that on a branch"...