upper and lower)wedge and hartree; wedge requires $\beta<2$, since its coordinate conversion is singular at $\beta=2$)from generate import Displacement
generator = Displacement()
print(generator.value({'beta': '0', 'branch': 'upper', 'normalisation': 'wedge'}, digits=100))
All 20 entries have 100 significant decimal digits. The attached generate.py adapter uses the unchanged falkner_skan_core.py dependency. Displacement is integrated from the velocity Taylor series on the monotone upper branch. Its convergence is checked separately from the companion quantities, using 160 working decimal digits, order 140, step 1/8 and endpoint 32, and 210 digits, order 180, step 1/10 and endpoint 36. The largest relative displacement difference in either normalisation is 9.645e-149. The wedge values are obtained by multiplying Hartree displacement by $\sqrt{2-\beta}$ using 240-digit Decimal arithmetic before rounding.
The momentum identity is recomputed from displacement, momentum thickness and wall shear at both settings. Boundary residuals and sampled monotonicity are checked at every gradient. Independent SciPy collocation and adaptive quadrature check all ten profiles and both normalisations at ordinary floating-point precision. For Blasius flow, an additional 210-digit initial-value integration with unit wall shear uses Toepfer scaling and the asymptotic profile intercept to check displacement without shooting or displacement quadrature. It shares the Taylor recurrence.
The two main runs share an algorithm. Finite-domain and Taylor truncation errors, including the omitted displacement tail, are convergence-tested rather than rigorously enclosed; Arb midpoints are retained at each step. These checks justify heuristic agreement, not proven 100-digit error bounds.