fix(solvers): pyamgx solvers crash on instantiation due to precon attribute typo

[gaoflow]

Fixes #1199.

Problem

Instantiating any pyamgx-backed solver raises immediately, before any GPU work:

>>> from fipy.solvers.pyamgx import LinearGMRESSolver
>>> LinearGMRESSolver()
AttributeError: 'LinearGMRESSolver' object has no attribute 'precon'

PyAMGXSolver.__init__ checks self.precon, but the base Solver stores the
preconditioner on self.preconditioner (set from the precon argument via
value_or_default(precon, default_preconditioner)). There is no self.precon
attribute, so construction fails.

This matches the maintainer’s diagnosis in #1199 ("it looks like there’s a typo … that should be if self.preconditioner is not None: self.preconditioner._applyToSolver(...)"), and the reporter confirmed that renaming it removes the error.

Fix

1. Use self.preconditioner instead of the non-existent self.precon.
2. Forward precon=precon to the base Solver.__init__, matching scipySolver
   and petscKrylovSolver. Without this, both an explicitly passed precon
   and the pyamgx solvers’ class DEFAULT_PRECONDITIONER were silently
   dropped — the base initializes self.preconditioner from its precon arg,
   which previously defaulted to "default" → None for every pyamgx solver.

super(PyAMGXSolver, self).__init__(tolerance=tolerance, criterion=criterion,
                                   iterations=iterations, precon=precon)
...
if self.preconditioner is not None:
    self.preconditioner._applyToSolver(self.config_dict["solver"])

Verification

• py_compile clean; no remaining self.precon reference.
• Verified by code inspection and consistency with the sibling scipySolver /
  petscKrylovSolver __init__ signatures, plus the maintainer diagnosis and
  reporter confirmation in object has no attribute 'precon' in pyamgx solvers #1199.
• Not runtime-tested locally: pyamgx requires the AMGX library and an NVIDIA
  GPU, which are unavailable in this environment, so I could not add a
  CI-runnable regression test without that dependency. Happy to add one if there
  is a pyamgx-capable CI lane I can target.

[codacy-production]

Up to standards ✅

🟢 Issues 0 issues

Results:
0 new issues

View in Codacy

🟢 Metrics 0 complexity · 0 duplication

Metric	Results
Complexity	0
Duplication	0

View in Codacy

_{NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer
TIP This summary will be updated as you push new changes.}