Skip to content

AgentFEM Module Map

This map links FEM concepts to the current Python modules. Module paths in this table are relative to src/agentfem/ unless stated otherwise; runnable examples remain in the repository-level examples/.

Concept Module
Dependency-free public module, Model-verb, CLI-command, and workflow-stage inventories _api_contract.py
Executable Model/Constitutive/State/Operator/Procedure/Backend/Result ownership inventory _architecture_contract.py
Mesh import, boundary/cell regions, tags, summaries, checks, and measures mesh/
External mesh inventory, cell/facet set conversion, and manifests mesh/formats.py
Abaqus migration inventory, element semantics, labels, equations, periodic-cell geometry, and source-order output mesh/abaqus.py
Scope-aware Abaqus engineering migration plans and fail-closed project scaffolds mesh/abaqus_migration.py
Reviewed narrow Abaqus-to-native lowering and one-instance orphan-mesh derivation mesh/abaqus_lowering.py
Study context: analysis type, physics, dimension, and assumptions studies.py
Shared axisymmetric (r,z) kinematics and full-revolution lowering _axisymmetric.py
Standard/Explicit family, equation order, algorithm, and state policy procedures.py
Model registry/facade, amplitudes, material assignments, and readable engineering verbs models.py
Side-effect-free Model validation, inspection/AF-IR views, and shared inspection helpers _model_validation.py, _model_inspection.py, _model_support.py
Stable private facade for built-in scientific Step construction _step_builders.py
Linear-static and steady/transient thermal Step construction _step_builders_thermal.py
Finite-kinematics hyperelastic, mixed, and fabric-membrane Step construction _step_builders_finite_strain.py
Stateful J2, creep, and quasistatic viscoelastic Step construction _step_builders_inelastic.py
Direct-harmonic and harmonic-viscoelastic Step construction _step_builders_frequency.py
Explicit, implicit, finite-strain dynamic, and modal Step construction _step_builders_dynamics.py
Extensible analysis/material lowering and execution-policy normalization behind model.step() step_providers.py
Deterministic provider storage and selection, independent of scientific lowering _step_provider_registry.py
Dependency-light target, material-capability, and procedure-selection predicates _step_provider_support.py
Built-in provider predicates, lowerers, and declarations _builtin_step_providers.py
Function spaces and Lagrange defaults spaces.py
Application-level unknown fields fields.py
Time histories and scale factors for prescribed data amplitudes.py
Low-level dof lookup and field copying kernel/dofs.py
Constraint containers and strong BC construction constraints/
Loads and natural boundary data loads.py
Constitutive response relations and maturity catalog constitutive/
Standard-neutral engineering creep/fatigue assessment consumers and interaction diagrams assessments.py
Quadrature-point committed/trial state constitutive/quadrature.py
Shared restart/replace protocols and first-/second-order transient state state.py
Monotonic cohesive laws, paired interface topology, and physical-keyed state interfaces.py, cohesive_checkpoint.py
Independent cycle coordinate, cyclic cohesive damage, cycle jumps, and 3D fatigue-crack observations fatigue_fracture.py
Finite-strain cohesive forces, named interface composition, energy, wave speeds, and dynamic crack evidence fracture.py
Global stateful and modal solid-mechanics procedures mechanics/
Finite- and small-strain material-point contracts, named parameters, batch response validation, UMAT/UHYPER source inspection, and bridge specifications constitutive/user_material.py, constitutive/small_strain_user_material.py, constitutive/material_driver.py
Named material definitions, project Python assets, physics-role compatibility, reference records, and property containers materials/
Boundary models boundary_models/
UFL weak-form blocks, including stiffness, mass, diffusion, and loads forms.py
Assembly assembly.py
Engineering-level K/M/C/F, registered-assignment and model-first operator lowering, transport, SUPG, and reaction operators operators/, operators/_model_lowering.py
Central difference, Newmark, generalized-alpha, and runtime cadence time/
Common analysis steps, discrete system problems, factories, and compatibility delegates problems.py
Incremental/affine nonlinear procedures, state transactions, cutback, and nonlinear checkpointing _nonlinear_problems.py
First-/second-order transient and explicit-dynamics procedure implementations _transient_problems.py
Shared residual-to-reaction diagnostic-field recovery _problem_fields.py
Structural modal procedure, distributed eigensolver execution, and result assembly mechanics/modal.py, backends/_modal.py, results/_modal.py
Shared transient checkpoint envelope, integrity, accepted-increment cadence, and partition identity checkpointing.py
Automatic/fixed incrementation and cutback policy steps.py
Linear/nonlinear solvers, prepared-solve lifecycle, and convergence evidence _solver_lifecycle.py, solvers.py
Scientific results, MPI-safe point/path/integral QoIs, histories, artifacts, and dataset bridge results/
Rank-reduced timing, workload, solver, and parallel performance evidence results/performance.py, diagnostics.py
Private direct-problem, static, nonlinear, transient, and modal result assembly results/_problem.py, results/_analysis_step.py, results/_nonlinear_step.py, results/_transient_step.py, results/_modal.py
Shared field location and processing semantics results/_field_metadata.py
Finite-strain fields and periodic-cell homogenization results/finite_strain.py
Global and regional L2 projection; engineering-default S/E/MISES, opt-in SENER, and explicit processing metadata results/projection.py, results/core.py
Standard result variables and context-aware aliases results/field_catalog.py
Declarative field requests and unified XDMF/HDF5 time-series writer results/output.py
Combined field/history/diagnostic/presentation output contracts results/plan.py
Reactions, work/energy, thermal balance, progress, and distributed diagnostics diagnostics.py, results/quantities.py
Output writers and scalar logs io.py
Element and integration policies elements/
Test-linked verification obligations benchmarks/
Versioned external scientific contracts integrations/
Runnable workflows examples/
Versioned AF-IR scientific records ir/
Structured validation issues and reports validation.py
Scientific trust, claims, and convergence evidence verification.py
Backend descriptors, registry, and lowering adapters backends/
Explicit installed/private package discovery and staged registration extensions.py
Typed/JSON parameters, sampling, campaign plans, resumable case execution campaigns/
Scientific dataset schemas, arrays, provenance, splits, and FEM-field export datasets/
Unified learning entry; neural-field objectives, conditions, sampling, inferred parameters, learned-constitutive specifications/provider discovery, and compatibility access learning/
Surrogate/ROM models, validation, applicability, observation grids, neural-operator/PINN contracts surrogates/

Application geometry, case inputs, and problem-specific sources should stay in application packages or examples until they become reusable platform concepts.