PR review medium effort fixes: deduplicate canonicalizers and optimiz…

[Transurgeon]

…e bounds

• Add _common.py with canonicalize_unary_smooth() helper to reduce code duplication across hyperbolic, trig, exp, and prod canonicalizers
• Refactor canonicalizer files to use the common helper
• Preallocate numpy arrays in Bounds class instead of list extend + convert
• Fix hessian() to return typed empty array (dtype=np.float64)
• Cache empty hessian structure in hessianstructure() for quasi-Newton

Description

Please include a short summary of the change.
Issue link (if applicable):

Type of change

• New feature (backwards compatible)
• New feature (breaking API changes)
• Bug fix
• Other (Documentation, CI, ...)

Contribution checklist

• Add our license to new files.
• Check that your code adheres to our coding style.
• Write unittests.
• Run the unittests and check that they’re passing.
• Run the benchmarks to make sure your change doesn’t introduce a regression.

[github-actions]

Benchmarks that have stayed the same:

   before           after         ratio
 [a87c4f1f]       [20f9f050]
      333±0ms          355±0ms     1.07  gini_portfolio.Murray.time_compile_problem
      509±0ms          530±0ms     1.04  semidefinite_programming.SemidefiniteProgramming.time_compile_problem
      5.29±0s          5.45±0s     1.03  optimal_advertising.OptimalAdvertising.time_compile_problem
      2.13±0s          2.18±0s     1.02  quantum_hilbert_matrix.QuantumHilbertMatrix.time_compile_problem
      910±0ms          931±0ms     1.02  gini_portfolio.Cajas.time_compile_problem
      418±0ms          428±0ms     1.02  gini_portfolio.Yitzhaki.time_compile_problem
     14.1±0ms         14.4±0ms     1.02  simple_LP_benchmarks.SimpleFullyParametrizedLPBenchmark.time_compile_problem
     34.5±0ms         35.2±0ms     1.02  matrix_stuffing.SmallMatrixStuffing.time_compile_problem
      35.0±0s          35.5±0s     1.01  sdp_segfault_1132_benchmark.SDPSegfault1132Benchmark.time_compile_problem
      16.4±0s          16.6±0s     1.01  simple_LP_benchmarks.SimpleLPBenchmark.time_compile_problem
      1.33±0s          1.34±0s     1.01  matrix_stuffing.ParamConeMatrixStuffing.time_compile_problem
      1.18±0s          1.19±0s     1.01  simple_LP_benchmarks.SimpleScalarParametrizedLPBenchmark.time_compile_problem
      1.74±0s          1.75±0s     1.01  finance.FactorCovarianceModel.time_compile_problem
      783±0ms          787±0ms     1.01  matrix_stuffing.ConeMatrixStuffingBench.time_compile_problem
      216±0ms          217±0ms     1.00  high_dim_convex_plasticity.ConvexPlasticity.time_compile_problem
      272±0ms          273±0ms     1.00  matrix_stuffing.ParamSmallMatrixStuffing.time_compile_problem
      6.55±0s          6.56±0s     1.00  huber_regression.HuberRegression.time_compile_problem
     13.8±0ms         13.9±0ms     1.00  simple_QP_benchmarks.ParametrizedQPBenchmark.time_compile_problem
      21.2±0s          21.2±0s     1.00  finance.CVaRBenchmark.time_compile_problem
      1.32±0s          1.32±0s     0.99  simple_QP_benchmarks.LeastSquares.time_compile_problem
      289±0ms          286±0ms     0.99  simple_QP_benchmarks.SimpleQPBenchmark.time_compile_problem
      3.35±0s          3.32±0s     0.99  simple_QP_benchmarks.UnconstrainedQP.time_compile_problem
      7.84±0s          7.75±0s     0.99  svm_l1_regularization.SVMWithL1Regularization.time_compile_problem
      1.91±0s          1.87±0s     0.98  tv_inpainting.TvInpainting.time_compile_problem
      406±0ms          395±0ms     0.97  slow_pruning_1668_benchmark.SlowPruningBenchmark.time_compile_problem