From 31a840feef0e422d9ae6158e31b5d3f2d5533b15 Mon Sep 17 00:00:00 2001 From: ChrisRackauckas-Claude Date: Fri, 19 Jun 2026 04:43:30 -0400 Subject: [PATCH] Fix docs build: resolve collocate_data ambiguity in neural_ode_training.md On Julia 1.12 the docs `build / Build and Deploy Documentation` job fails in the first `@example collocation_cp` block of neural_ode_training.md with: UndefVarError: `collocate_data` not defined ... Hint: ... ambiguity ... also exists in DiffEqFlux ... also exists in DataCollocations. The block does both `using DiffEqFlux` and `using DataCollocations`, and both modules export `collocate_data`. Julia 1.12 no longer silently picks one of two clashing exported bindings; an unqualified reference is an ambiguity error. The failure of the first block cascaded to every subsequent block (du/u/prob_neuralode undefined), terminating the build. Fix: import `collocate_data` explicitly from DataCollocations (alongside the already-explicit EpanechnikovKernel), so the binding is unambiguous regardless of DiffEqFlux's re-export. Verified locally on Julia 1.12.6 against the real DiffEqFlux + DataCollocations packages: the old import reproduces the exact CI UndefVarError, the new import resolves collocate_data to DataCollocations.collocate_data, and the example's collocate_data(...) call executes successfully. Co-Authored-By: Chris Rackauckas Co-Authored-By: Claude Opus 4.8 (1M context) --- docs/src/neural_ode_training.md | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/docs/src/neural_ode_training.md b/docs/src/neural_ode_training.md index e4cd7b1..de5c8fa 100644 --- a/docs/src/neural_ode_training.md +++ b/docs/src/neural_ode_training.md @@ -17,8 +17,7 @@ Before getting to the explanation, here's some code to start with. We will follo ```@example collocation_cp using ComponentArrays, Lux, DiffEqFlux, OrdinaryDiffEq, SciMLSensitivity, Optimization, OptimizationOptimisers, Plots -using DataCollocations -using DataCollocations: EpanechnikovKernel +using DataCollocations: collocate_data, EpanechnikovKernel using Random rng = Xoshiro(0)