docs: KCSolve architecture and Python API reference

- Replace OndselSolver architecture doc with KCSolve pluggable solver architecture covering IKCSolver interface, SolverRegistry, OndselAdapter, Python bindings, file layout, and testing - Add kcsolve Python API reference with full type documentation, module functions, usage examples, and pybind11 vector-copy caveat - Add INTER_SOLVER.md spec (previously untracked) with Phase 1 and Phase 2 marked as complete - Update SUMMARY.md with new page links
2026-02-19 18:59:05 -06:00
parent 7ea0078ba3
commit 406e120180
4 changed files with 1004 additions and 17 deletions
--- a/docs/INTER_SOLVER.md
+++ b/docs/INTER_SOLVER.md
@@ -0,0 +1,568 @@
+# Pluggable Assembly Solver Architecture
+
+**Status:** Phase 2 complete  
+**Last Updated:** 2026-02-19
+
+---
+
+## 1. Problem
+
+Kindred Create currently vendors OndselSolver as a monolithic assembly constraint solver. Different engineering domains benefit from different solver strategies — Lagrangian methods work well for rigid body assemblies but poorly for over-constrained or soft-constraint systems. A pluggable architecture lets us ship multiple solvers (including experimental ones) without touching core assembly logic, and lets the server farm out solve jobs to headless worker processes.
+
+---
+
+## 2. Design Goals
+
+1. **Stable C++ API** — A solver-agnostic interface that the Assembly module calls. Solvers are shared libraries loaded at runtime.
+2. **Python binding layer** — Every C++ solver is exposed to Python via pybind11, enabling rapid prototyping, debugging, and server-side execution without a full GUI build.
+3. **Solver-defined joint types** — Each solver declares its own joint/mate vocabulary, mapped from a common base set (inspired by SOLIDWORKS mates: coincident, concentric, tangent, distance, angle, lock, etc.).
+4. **Semi-deterministic solving** — Consistent results given consistent input ordering, with configurable tolerance and iteration limits.
+5. **Server-compatible** — Solvers run as detached processes claimed by `silorunner` workers via the existing job queue.
+
+---
+
+## 3. Architecture Layers
+
+```
+┌──────────────────────────────────────────────────────┐
+│  Layer 4: Server / Worker                            │
+│  silorunner claims solve jobs, executes via Python   │
+│  Headless Create or standalone solver process        │
+├──────────────────────────────────────────────────────┤
+│  Layer 3: Python Debug & Scripting                   │
+│  pybind11 bindings for all solvers                   │
+│  Introspection, step-through, constraint viz         │
+│  import kcsolve; s = kcsolve.load("ondsel")          │
+├──────────────────────────────────────────────────────┤
+│  Layer 2: Solver Plugins (.so / .dll / .dylib)       │
+│  Each implements IKCSolver interface                  │
+│  Registers joint types via manifest                   │
+│  Loaded by SolverRegistry at runtime                  │
+├──────────────────────────────────────────────────────┤
+│  Layer 1: C++ Solver API (libkcsolve)                │
+│  IKCSolver, JointDef, SolveContext, SolveResult      │
+│  SolverRegistry (discovery, loading, selection)       │
+│  Ships as a shared library linked by Assembly module  │
+└──────────────────────────────────────────────────────┘
+```
+
+---
+
+## 4. Layer 1: C++ Solver API
+
+Located at `src/Mod/Assembly/Solver/` (or `src/Lib/KCSolve/` if we want it independent of Assembly).
+
+### 4.1 Core Types
+
+```cpp
+namespace KCSolve {
+
+// Unique identifier for a joint type within a solver
+struct JointTypeId {
+    std::string solver_id;   // e.g. "ondsel", "gnn", "relaxation"
+    std::string joint_name;  // e.g. "coincident", "distance"
+};
+
+// Base joint categories (SOLIDWORKS-inspired vocabulary)
+enum class BaseJointKind {
+    Coincident,
+    Concentric,
+    Tangent,
+    Distance,
+    Angle,
+    Lock,
+    Parallel,
+    Perpendicular,
+    PointOnLine,
+    SymmetricPlane,
+    Gear,
+    Rack,
+    Cam,
+    Slot,
+    Hinge,
+    Slider,
+    Cylindrical,
+    Planar,
+    Ball,
+    Screw,
+    Universal,
+    Custom   // solver-specific extension
+};
+
+// A joint definition registered by a solver plugin
+struct JointDef {
+    JointTypeId id;
+    BaseJointKind base_kind;          // which vanilla category it maps to
+    std::string display_name;
+    std::string description;
+    uint32_t dof_removed;             // degrees of freedom this joint removes
+    std::vector<std::string> params;  // parameter names (e.g. "distance", "angle")
+    bool supports_limits = false;
+    bool supports_friction = false;
+};
+
+// A constraint instance in a solve problem
+struct Constraint {
+    JointTypeId joint_type;
+    std::string part_a;    // part label or id
+    std::string part_b;
+    // Geometry references (face, edge, vertex indices)
+    std::vector<std::string> refs_a;
+    std::vector<std::string> refs_b;
+    std::map<std::string, double> params;   // param_name -> value
+    bool suppressed = false;
+};
+
+// Input to a solve operation
+struct SolveContext {
+    std::vector<Constraint> constraints;
+    // Part placements as 4x4 transforms (initial guess)
+    std::map<std::string, std::array<double, 16>> placements;
+    // Which parts are grounded (fixed)
+    std::set<std::string> grounded;
+    // Solver config
+    double tolerance = 1e-10;
+    uint32_t max_iterations = 500;
+    bool deterministic = true;    // force consistent ordering
+    // Optional: previous solution for warm-starting
+    std::map<std::string, std::array<double, 16>> warm_start;
+};
+
+enum class SolveStatus {
+    Converged,
+    MaxIterationsReached,
+    Overconstrained,
+    Underconstrained,
+    Redundant,
+    Failed
+};
+
+struct ConstraintDiagnostic {
+    std::string constraint_id;
+    double residual;
+    bool satisfied;
+    std::string message;
+};
+
+struct SolveResult {
+    SolveStatus status;
+    uint32_t iterations;
+    double final_residual;
+    double solve_time_ms;
+    std::map<std::string, std::array<double, 16>> placements;
+    std::vector<ConstraintDiagnostic> diagnostics;
+    // For semi-deterministic: hash of input ordering
+    uint64_t input_hash;
+};
+
+} // namespace KCSolve
+```
+
+### 4.2 Solver Interface
+
+```cpp
+namespace KCSolve {
+
+class IKCSolver {
+public:
+    virtual ~IKCSolver() = default;
+
+    // Identity
+    virtual std::string id() const = 0;
+    virtual std::string name() const = 0;
+    virtual std::string version() const = 0;
+
+    // Joint type registry — called once at load
+    virtual std::vector<JointDef> supported_joints() const = 0;
+
+    // Solve
+    virtual SolveResult solve(const SolveContext& ctx) = 0;
+
+    // Incremental: update a single constraint without full re-solve
+    // Default impl falls back to full solve
+    virtual SolveResult update(const SolveContext& ctx,
+                               const std::string& changed_constraint) {
+        return solve(ctx);
+    }
+
+    // Diagnostic: check if a constraint set is well-posed before solving
+    virtual SolveStatus diagnose(const SolveContext& ctx) {
+        return SolveStatus::Converged; // optimistic default
+    }
+
+    // Determinism: given identical input, produce identical output
+    virtual bool is_deterministic() const { return false; }
+};
+
+// Plugin entry point — each .so exports this symbol
+using CreateSolverFn = IKCSolver* (*)();
+
+} // namespace KCSolve
+```
+
+### 4.3 Solver Registry
+
+```cpp
+namespace KCSolve {
+
+class SolverRegistry {
+public:
+    // Scan a directory for solver plugins (*.so / *.dll / *.dylib)
+    void scan(const std::filesystem::path& plugin_dir);
+
+    // Manual registration (for built-in solvers like Ondsel)
+    void register_solver(std::unique_ptr<IKCSolver> solver);
+
+    // Lookup
+    IKCSolver* get(const std::string& solver_id) const;
+    std::vector<std::string> available() const;
+
+    // Joint type resolution: find which solvers support a given base kind
+    std::vector<JointTypeId> joints_for(BaseJointKind kind) const;
+
+    // Global default solver
+    void set_default(const std::string& solver_id);
+    IKCSolver* get_default() const;
+};
+
+} // namespace KCSolve
+```
+
+### 4.4 Plugin Loading
+
+Each solver plugin is a shared library exporting:
+
+```cpp
+extern "C" KCSolve::IKCSolver* kcsolve_create();
+extern "C" const char* kcsolve_api_version();  // "1.0"
+```
+
+The registry `dlopen`s each library, checks `kcsolve_api_version()` compatibility, and calls `kcsolve_create()`. Plugins are discovered from:
+
+1. `<install_prefix>/lib/kcsolve/` — system-installed solvers
+2. `~/.config/KindredCreate/solvers/` — user-installed solvers
+3. `KCSOLVE_PLUGIN_PATH` env var — development overrides
+
+---
+
+## 5. Layer 2: OndselSolver Adapter
+
+The first plugin wraps the existing OndselSolver, mapping its internal constraint types to the `IKCSolver` interface.
+
+```
+src/Mod/Assembly/Solver/
+├── IKCSolver.h            # Interface + types from §4
+├── SolverRegistry.cpp     # Plugin discovery and loading
+├── OndselAdapter.cpp      # Wraps OndselSolver as IKCSolver plugin
+└── CMakeLists.txt
+```
+
+`OndselAdapter` translates between `SolveContext` ↔ OndselSolver's Lagrangian formulation. This is the reference implementation and proves the API works before any new solvers are written.
+
+Joint mapping for OndselAdapter:
+
+| BaseJointKind | Ondsel Constraint | DOF Removed |
+|---------------|-------------------|-------------|
+| Coincident | PointOnPoint | 3 |
+| Concentric | CylindricalOnCylindrical | 4 |
+| Tangent | FaceOnFace (tangent mode) | 1 |
+| Distance | PointOnPoint + offset | 2 |
+| Angle | AxisAngle | 1 |
+| Lock | FullLock | 6 |
+| Hinge | RevoluteJoint | 5 |
+| Slider | PrismaticJoint | 5 |
+| Cylindrical | CylindricalJoint | 4 |
+| Ball | SphericalJoint | 3 |
+
+---
+
+## 6. Layer 3: Python Bindings
+
+### 6.1 pybind11 Module
+
+```
+src/Mod/Assembly/Solver/bindings/
+├── kcsolve_py.cpp         # pybind11 module definition
+└── CMakeLists.txt
+```
+
+```python
+import kcsolve
+
+# List available solvers
+print(kcsolve.available())  # ["ondsel", ...]
+
+# Load a solver
+solver = kcsolve.load("ondsel")
+print(solver.name, solver.version)
+print(solver.supported_joints())
+
+# Build a problem
+ctx = kcsolve.SolveContext()
+ctx.add_part("base", placement=..., grounded=True)
+ctx.add_part("arm", placement=...)
+ctx.add_constraint("coincident", "base", "arm",
+                    refs_a=["Face6"], refs_b=["Face1"])
+
+# Solve
+result = solver.solve(ctx)
+print(result.status)          # SolveStatus.Converged
+print(result.iterations)      # 12
+print(result.solve_time_ms)   # 3.4
+print(result.placements["arm"])
+
+# Diagnostics per constraint
+for d in result.diagnostics:
+    print(f"{d.constraint_id}: residual={d.residual:.2e} ok={d.satisfied}")
+```
+
+### 6.2 Debug / Introspection API
+
+The Python layer adds capabilities the C++ interface intentionally omits for performance:
+
+```python
+# Step-through solving (debug mode)
+debugger = kcsolve.Debugger(solver, ctx)
+for step in debugger.iterate():
+    print(f"iter {step.iteration}: residual={step.residual:.6e}")
+    print(f"  moved: {step.parts_moved}")
+    print(f"  worst constraint: {step.worst_constraint}")
+    if step.residual < 1e-8:
+        break
+
+# Constraint dependency graph
+graph = kcsolve.dependency_graph(ctx)
+# Returns dict: constraint_id -> [dependent_constraint_ids]
+
+# DOF analysis
+analysis = kcsolve.dof_analysis(ctx)
+print(f"Total DOF: {analysis.total_dof}")
+print(f"Removed: {analysis.constrained_dof}")
+print(f"Remaining: {analysis.free_dof}")
+for part, dofs in analysis.per_part.items():
+    print(f"  {part}: {dofs} free")
+```
+
+### 6.3 Pure-Python Solver Support
+
+The Python layer also supports solvers written entirely in Python (no C++ required). This is the fast path for prototyping new approaches (GNN, relaxation, etc.):
+
+```python
+class RelaxationSolver(kcsolve.PySolver):
+    """A pure-Python iterative relaxation solver for prototyping."""
+
+    id = "relaxation"
+    name = "Iterative Relaxation"
+    version = "0.1.0"
+
+    def supported_joints(self):
+        return [
+            kcsolve.JointDef("coincident", kcsolve.BaseJointKind.Coincident, dof_removed=3),
+            kcsolve.JointDef("distance", kcsolve.BaseJointKind.Distance, dof_removed=2),
+            # ...
+        ]
+
+    def solve(self, ctx: kcsolve.SolveContext) -> kcsolve.SolveResult:
+        placements = dict(ctx.placements)
+        for i in range(ctx.max_iterations):
+            max_residual = 0.0
+            for c in ctx.constraints:
+                residual = self._eval_constraint(c, placements)
+                correction = self._compute_correction(c, residual)
+                self._apply_correction(placements, c, correction)
+                max_residual = max(max_residual, abs(residual))
+            if max_residual < ctx.tolerance:
+                return kcsolve.SolveResult(
+                    status=kcsolve.SolveStatus.Converged,
+                    iterations=i + 1,
+                    final_residual=max_residual,
+                    placements=placements
+                )
+        return kcsolve.SolveResult(
+            status=kcsolve.SolveStatus.MaxIterationsReached,
+            iterations=ctx.max_iterations,
+            final_residual=max_residual,
+            placements=placements
+        )
+
+# Register at runtime
+kcsolve.register(RelaxationSolver())
+```
+
+Python solvers are discovered from:
+- `<user_macros>/solvers/*.py` — user-written solvers
+- `mods/*/solvers/*.py` — addon-provided solvers
+
+---
+
+## 7. Layer 4: Server Integration
+
+### 7.1 Solve Job Definition
+
+Extends the existing worker system (WORKERS.md) with a new job type:
+
+```yaml
+job:
+  name: assembly-solve
+  version: 1
+  description: "Solve assembly constraints using specified solver"
+  trigger:
+    type: revision_created
+    filter:
+      item_type: assembly
+  scope:
+    type: assembly
+  compute:
+    type: solve
+    command: create-solve
+    args:
+      solver: ondsel           # or "auto" for registry default
+      tolerance: 1e-10
+      max_iterations: 500
+      deterministic: true
+      output_placements: true  # write solved placements back to revision
+      output_diagnostics: true # store constraint diagnostics in job result
+  runner:
+    tags: [create, solver]
+  timeout: 300
+  max_retries: 1
+  priority: 75
+```
+
+### 7.2 Headless Solve via Runner
+
+The `create-solve` command in `silorunner`:
+
+1. Claims job from Silo server
+2. Downloads the assembly `.kc` file
+3. Launches Headless Create (or standalone Python if pure-Python solver)
+4. Loads the assembly, extracts constraint graph → `SolveContext`
+5. Calls `solver.solve(ctx)`
+6. Reports `SolveResult` back via `POST /api/runner/jobs/{id}/complete`
+7. Optionally writes updated placements as a new revision
+
+### 7.3 Standalone Solve Process (No GUI)
+
+For server-side batch solving without Headless Create overhead:
+
+```python
+#!/usr/bin/env python3
+"""Standalone solver worker — no FreeCAD dependency."""
+import kcsolve
+import json, sys
+
+problem = json.load(sys.stdin)
+ctx = kcsolve.SolveContext.from_dict(problem)
+
+solver = kcsolve.load(problem.get("solver", "ondsel"))
+result = solver.solve(ctx)
+
+json.dump(result.to_dict(), sys.stdout)
+```
+
+This enables lightweight solver containers that don't need the full Create installation — useful for CI validation, quick constraint checks, and scaling solver capacity independently of geometry workers.
+
+---
+
+## 8. Semi-Deterministic Behavior
+
+"Semi-deterministic" means: given the same constraint set and initial placements, the solver produces the same result. This is achieved by:
+
+1. **Canonical input ordering** — `SolveContext` sorts constraints and parts by a stable key (part label + constraint index) before passing to the solver. The ordering hash is stored in `SolveResult.input_hash`.
+
+2. **Solver contract** — `IKCSolver::is_deterministic()` reports whether the implementation guarantees this. OndselAdapter does (Lagrangian formulation with fixed pivot ordering). A GNN solver might not.
+
+3. **Tolerance-aware comparison** — Two `SolveResult`s are "equivalent" if all placement deltas are within tolerance, even if iteration counts differ. Used for regression testing.
+
+4. **Warm-start stability** — When `warm_start` placements are provided, the solver should converge to the same solution as a cold start (within tolerance), just faster. This is validated in the test suite.
+
+---
+
+## 9. Implementation Phases
+
+### Phase 1: API + OndselAdapter (foundation) -- COMPLETE
+
+- Defined `IKCSolver.h`, core types (`Types.h`), `SolverRegistry`
+- Implemented `OndselAdapter` wrapping existing solver
+- Assembly module calls through `SolverRegistry` instead of directly calling OndselSolver
+- 18 C++ tests, 6 Python integration tests
+- **PR:** #297 (merged)
+
+### Phase 2: pybind11 Bindings -- COMPLETE
+
+- Built `kcsolve` pybind11 module exposing all enums, structs, and classes
+- `PyIKCSolver` trampoline for pure-Python solver subclasses
+- `register_solver()` for runtime Python solver registration
+- `PySolverHolder` for GIL-safe forwarding of virtual calls
+- 16 Python tests covering types, registry, and Python solver round-trips
+- Debug/introspection API (Debugger, `dependency_graph()`, `dof_analysis()`) deferred to Phase 4+
+- Automatic Python solver discovery (`mods/*/solvers/`) deferred -- users call `register_solver()` explicitly
+- **PR:** #298
+- **Docs:** `docs/src/architecture/ondsel-solver.md`, `docs/src/reference/kcsolve-python.md`
+
+### Phase 3: Server Integration
+
+- `create-solve` command for `silorunner`
+- YAML job definition for solve jobs
+- Standalone solver process (no FreeCAD dependency)
+- `SolveContext` JSON serialization for inter-process communication
+- **Deliverable:** Solve jobs run async through the worker system
+
+### Phase 4: Second Solver (validation)
+
+- Implement a simple relaxation or gradient-descent solver as a Python plugin
+- Validates that the API actually supports different solving strategies
+- Benchmark against OndselAdapter for correctness and performance
+- **Deliverable:** Two interchangeable solvers, selectable per-assembly
+
+### Phase 5: GNN Solver (future)
+
+- Graph Neural Network approach from existing roadmap
+- Likely a Python solver wrapping a trained model
+- Focus on fast approximate solutions for interactive editing
+- Falls back to OndselAdapter for final precision solve
+- **Deliverable:** Hybrid solve pipeline (GNN fast-guess → Lagrangian refinement)
+
+---
+
+## 10. File Locations
+
+```
+src/Lib/KCSolve/              # or src/Mod/Assembly/Solver/
+├── include/
+│   └── KCSolve/
+│       ├── IKCSolver.h        # Interface + all types
+│       ├── SolverRegistry.h   # Plugin loading and lookup
+│       └── Types.h            # Enums, structs
+├── src/
+│   ├── SolverRegistry.cpp
+│   └── OndselAdapter.cpp
+├── bindings/
+│   └── kcsolve_py.cpp         # pybind11
+├── plugins/                   # Additional compiled solver plugins
+└── CMakeLists.txt
+```
+
+---
+
+## 11. Open Questions
+
+1. **Location**: `src/Lib/KCSolve/` (independent library, usable without Assembly module) vs `src/Mod/Assembly/Solver/` (tighter coupling, simpler build)? Leaning toward `src/Lib/` since server workers need it without the full Assembly module.
+
+2. **Geometry abstraction**: The C++ API uses string references for faces/edges/vertices. Should we pass actual OCC geometry (TopoDS_Shape) through the interface, or keep it abstract and let each solver adapter resolve references? Abstract is more portable but adds a translation step.
+
+3. **Constraint persistence**: Currently constraints live in the FCStd XML. Should the pluggable layer introduce its own serialization, or always read/write through FreeCAD's property system?
+
+4. **API versioning**: `kcsolve_api_version()` returns a string. Semver with major-only breaking changes? How strict on backward compat for the plugin ABI?
+
+5. **License implications**: OndselSolver is LGPL. New solver plugins could be any license since they're loaded at runtime via a stable C API boundary. Confirm this interpretation.
+
+---
+
+## 12. References
+
+- [ondsel-solver.md](ondsel-solver.md) — Current solver documentation
+- [WORKERS.md](WORKERS.md) — Worker/runner job system
+- [MULTI_USER_EDITS.md](MULTI_USER_EDITS.md) — Async validation pipeline
+- [DAG.md](DAG.md) — Dependency graph for incremental recompute
+- [ROADMAP.md](ROADMAP.md) — Tier 3 compute modules, GNN solver plans
--- a/docs/src/SUMMARY.md
+++ b/docs/src/SUMMARY.md
@@ -19,7 +19,7 @@
 - [Python as Source of Truth](./architecture/python-source-of-truth.md)
 - [Silo Server](./architecture/silo-server.md)
 - [Signal Architecture](./architecture/signal-architecture.md)
- [OndselSolver](./architecture/ondsel-solver.md)
+- [KCSolve: Pluggable Solver](./architecture/ondsel-solver.md)

 # Development

@@ -64,3 +64,4 @@
 - [OriginSelectorWidget](./reference/cpp-origin-selector-widget.md)
 - [FileOriginPython Bridge](./reference/cpp-file-origin-python.md)
 - [Creating a Custom Origin (C++)](./reference/cpp-custom-origin-guide.md)
+- [KCSolve Python API](./reference/kcsolve-python.md)
--- a/docs/src/architecture/ondsel-solver.md
+++ b/docs/src/architecture/ondsel-solver.md
@@ -1,27 +1,132 @@
-# OndselSolver
+# KCSolve: Pluggable Solver Architecture

-OndselSolver is the assembly constraint solver used by FreeCAD's Assembly workbench. Kindred Create vendors a fork of the solver as a git submodule.
+KCSolve is the pluggable assembly constraint solver framework for Kindred Create. It defines an abstract solver interface (`IKCSolver`) and a runtime registry (`SolverRegistry`) that lets the Assembly module work with any conforming solver backend. The default backend wraps OndselSolver via `OndselAdapter`.

- **Path:** `src/3rdParty/OndselSolver/`
- **Source:** `git.kindred-systems.com/kindred/solver` (Kindred fork)
+- **Library:** `src/Mod/Assembly/Solver/` (builds `libKCSolve.so`)
+- **Python module:** `src/Mod/Assembly/Solver/bindings/` (builds `kcsolve.so`)
+- **Tests:** `tests/src/Mod/Assembly/Solver/` (C++), `src/Mod/Assembly/AssemblyTests/TestKCSolvePy.py` (Python)

-## How it works
+## Architecture

-The solver uses a **Lagrangian constraint formulation** to resolve assembly constraints (mates, joints, fixed positions). Given a set of parts with geometric constraints between them, it computes positions and orientations that satisfy all constraints simultaneously.
+```
+┌──────────────────────────────────────────────────┐
+│  Assembly Module (AssemblyObject.cpp)             │
+│  Builds SolveContext from FreeCAD document,       │
+│  calls solver via SolverRegistry                  │
+├──────────────────────────────────────────────────┤
+│  SolverRegistry (singleton)                       │
+│  register_solver(), get(), available()            │
+│  Plugin discovery via scan() / scan_default_paths │
+├──────────────┬───────────────────────────────────┤
+│ OndselAdapter │  Python solvers  │  Future plugins │
+│ (C++ built-in)│  (via kcsolve)   │  (.so plugins)  │
+└──────────────┴───────────────────────────────────┘
+```

-The Assembly workbench (`src/Mod/Assembly/`) calls the solver whenever constraints are added or modified. Kindred Create has patches to `Assembly/` that extend `findPlacement()` for better datum and origin handling.
+The Assembly module never references OndselSolver directly. All solver access goes through `SolverRegistry::instance().get()`, which returns a `std::unique_ptr<IKCSolver>`.

-## Why a fork
+## IKCSolver interface

-The solver is forked from the upstream Ondsel project for:
- **Pinned stability** — the submodule is pinned to a known-good commit
- **Potential modifications** — the fork allows Kindred-specific patches if needed
- **Availability** — hosted on Kindred's Gitea instance for reliable access
+A solver backend implements `IKCSolver` (defined in `IKCSolver.h`). Only three methods are pure virtual; all others have sensible defaults:

-## Future: GNN solver
+| Method | Required | Purpose |
+|--------|----------|---------|
+| `name()` | yes | Human-readable solver name |
+| `supported_joints()` | yes | List of `BaseJointKind` values the solver handles |
+| `solve(ctx)` | yes | Solve for static equilibrium |
+| `update(ctx)` | no | Incremental re-solve after parameter changes |
+| `pre_drag(ctx, parts)` | no | Begin interactive drag session |
+| `drag_step(placements)` | no | One mouse-move during drag |
+| `post_drag()` | no | End drag session |
+| `run_kinematic(ctx)` | no | Run kinematic simulation |
+| `num_frames()` | no | Frame count after simulation |
+| `update_for_frame(i)` | no | Retrieve frame placements |
+| `diagnose(ctx)` | no | Detect redundant/conflicting constraints |
+| `is_deterministic()` | no | Whether output is reproducible (default: true) |
+| `export_native(path)` | no | Write solver-native debug file (e.g. ASMT) |
+| `supports_bundle_fixed()` | no | Whether solver handles Fixed-joint bundling internally |

-There are plans to explore a Graph Neural Network (GNN) approach to constraint solving that could complement or supplement the Lagrangian solver for specific use cases. This is not yet implemented.
+## Core types

-## Related: GSL
+All types live in `Types.h` with no FreeCAD dependencies, making the header standalone for future server/worker use.

-The `src/3rdParty/GSL/` submodule is Microsoft's Guidelines Support Library (`github.com/microsoft/GSL`), providing C++ core guidelines utilities like `gsl::span` and `gsl::not_null`. It is a build dependency, not related to the constraint solver.
+**Transform** -- position `[x, y, z]` + unit quaternion `[w, x, y, z]`. Equivalent to `Base::Placement` but independent. Note the quaternion convention differs from `Base::Rotation` which uses `(x, y, z, w)` ordering; the adapter layer handles the swap.
+
+**BaseJointKind** -- 24 primitive constraint types decomposed from FreeCAD's `JointType` and `DistanceType` enums. Covers point constraints (Coincident, PointOnLine, PointInPlane), axis/surface constraints (Concentric, Tangent, Planar), kinematic joints (Fixed, Revolute, Cylindrical, Slider, Ball, Screw, Universal), mechanical elements (Gear, RackPinion), distance variants, and a `Custom` extension point.
+
+**SolveContext** -- complete solver input: parts (with placements, mass, grounded flag), constraints (with markers, parameters, limits), optional motion definitions and simulation parameters.
+
+**SolveResult** -- solver output: status code, updated part placements, DOF count, constraint diagnostics, and simulation frame count.
+
+## SolverRegistry
+
+Thread-safe singleton managing solver backends:
+
+```cpp
+auto& reg = SolverRegistry::instance();
+
+// Registration (at module init)
+reg.register_solver("ondsel", []() {
+    return std::make_unique<OndselAdapter>();
+});
+
+// Retrieval
+auto solver = reg.get();          // default solver
+auto solver = reg.get("ondsel");  // by name
+
+// Queries
+reg.available();          // ["ondsel", ...]
+reg.joints_for("ondsel"); // [Fixed, Revolute, ...]
+reg.set_default("ondsel");
+```
+
+Plugin discovery scans directories for shared libraries exporting `kcsolve_api_version()` and `kcsolve_create()`. Default paths: `KCSOLVE_PLUGIN_PATH` env var and `<prefix>/lib/kcsolve/`.
+
+## OndselAdapter
+
+The built-in solver backend wrapping OndselSolver's Lagrangian constraint formulation. Registered as `"ondsel"` at Assembly module initialization.
+
+Supports all 24 joint types. The adapter translates between `SolveContext`/`SolveResult` and OndselSolver's internal `ASMTAssembly` representation, including:
+
+- Part placement conversion (Transform <-> Base::Placement quaternion ordering)
+- Constraint parameter mapping (BaseJointKind -> OndselSolver joint classes)
+- Interactive drag protocol (pre_drag/drag_step/post_drag)
+- Kinematic simulation (run_kinematic/num_frames/update_for_frame)
+- Constraint diagnostics (redundancy detection via MbD system)
+
+## Python bindings (kcsolve module)
+
+The `kcsolve` pybind11 module exposes the full C++ API to Python. See [KCSolve Python API](../reference/kcsolve-python.md) for details.
+
+Key capabilities:
+- All enums, structs, and classes accessible from Python
+- Subclass `IKCSolver` in pure Python to create new solver backends
+- Register Python solvers at runtime via `kcsolve.register_solver()`
+- Query the registry from the FreeCAD console
+
+## File layout
+
+```
+src/Mod/Assembly/Solver/
+├── Types.h              # Enums and structs (no FreeCAD deps)
+├── IKCSolver.h          # Abstract solver interface
+├── SolverRegistry.h/cpp # Singleton registry + plugin loading
+├── OndselAdapter.h/cpp  # OndselSolver wrapper
+├── KCSolveGlobal.h      # DLL export macros
+├── CMakeLists.txt       # Builds libKCSolve.so
+└── bindings/
+    ├── PyIKCSolver.h    # pybind11 trampoline for Python subclasses
+    ├── kcsolve_py.cpp   # Module definition (enums, structs, classes)
+    └── CMakeLists.txt   # Builds kcsolve.so (pybind11 module)
+```
+
+## Testing
+
+- **18 C++ tests** (`KCSolve_tests_run`) covering SolverRegistry (8 tests) and OndselAdapter (10 tests including drag protocol and redundancy diagnosis)
+- **16 Python tests** (`TestKCSolvePy`) covering module import, type bindings, registry functions, Python solver subclassing, and full register/load/solve round-trips
+- **6 Python integration tests** (`TestSolverIntegration`) testing solver behavior through FreeCAD document objects
+
+## Related
+
+- [KCSolve Python API Reference](../reference/kcsolve-python.md)
+- [INTER_SOLVER.md](../../INTER_SOLVER.md) -- full architecture specification
--- a/docs/src/reference/kcsolve-python.md
+++ b/docs/src/reference/kcsolve-python.md
@@ -0,0 +1,313 @@
+# KCSolve Python API Reference
+
+The `kcsolve` module provides Python access to the KCSolve pluggable solver framework. It is built with pybind11 and installed alongside the Assembly module.
+
+```python
+import kcsolve
+```
+
+## Module constants
+
+| Name | Value | Description |
+|------|-------|-------------|
+| `API_VERSION_MAJOR` | `1` | KCSolve API major version |
+
+## Enums
+
+### BaseJointKind
+
+Primitive constraint types. 24 values:
+
+`Coincident`, `PointOnLine`, `PointInPlane`, `Concentric`, `Tangent`, `Planar`, `LineInPlane`, `Parallel`, `Perpendicular`, `Angle`, `Fixed`, `Revolute`, `Cylindrical`, `Slider`, `Ball`, `Screw`, `Universal`, `Gear`, `RackPinion`, `Cam`, `Slot`, `DistancePointPoint`, `DistanceCylSph`, `Custom`
+
+### SolveStatus
+
+| Value | Meaning |
+|-------|---------|
+| `Success` | Solve converged |
+| `Failed` | Solve did not converge |
+| `InvalidFlip` | Orientation flipped past threshold |
+| `NoGroundedParts` | No grounded parts in assembly |
+
+### DiagnosticKind
+
+`Redundant`, `Conflicting`, `PartiallyRedundant`, `Malformed`
+
+### MotionKind
+
+`Rotational`, `Translational`, `General`
+
+### LimitKind
+
+`TranslationMin`, `TranslationMax`, `RotationMin`, `RotationMax`
+
+## Structs
+
+### Transform
+
+Rigid-body transform: position + unit quaternion.
+
+| Field | Type | Default | Description |
+|-------|------|---------|-------------|
+| `position` | `list[float]` (3) | `[0, 0, 0]` | Translation (x, y, z) |
+| `quaternion` | `list[float]` (4) | `[1, 0, 0, 0]` | Unit quaternion (w, x, y, z) |
+
+```python
+t = kcsolve.Transform()
+t = kcsolve.Transform.identity()  # same as default
+```
+
+Note: quaternion convention is `(w, x, y, z)`, which differs from FreeCAD's `Base.Rotation(x, y, z, w)`. The adapter layer handles conversion.
+
+### Part
+
+| Field | Type | Default |
+|-------|------|---------|
+| `id` | `str` | `""` |
+| `placement` | `Transform` | identity |
+| `mass` | `float` | `1.0` |
+| `grounded` | `bool` | `False` |
+
+### Constraint
+
+| Field | Type | Default |
+|-------|------|---------|
+| `id` | `str` | `""` |
+| `part_i` | `str` | `""` |
+| `marker_i` | `Transform` | identity |
+| `part_j` | `str` | `""` |
+| `marker_j` | `Transform` | identity |
+| `type` | `BaseJointKind` | `Coincident` |
+| `params` | `list[float]` | `[]` |
+| `limits` | `list[Constraint.Limit]` | `[]` |
+| `activated` | `bool` | `True` |
+
+### Constraint.Limit
+
+| Field | Type | Default |
+|-------|------|---------|
+| `kind` | `LimitKind` | `TranslationMin` |
+| `value` | `float` | `0.0` |
+| `tolerance` | `float` | `1e-9` |
+
+### MotionDef
+
+| Field | Type | Default |
+|-------|------|---------|
+| `kind` | `MotionKind` | `Rotational` |
+| `joint_id` | `str` | `""` |
+| `marker_i` | `str` | `""` |
+| `marker_j` | `str` | `""` |
+| `rotation_expr` | `str` | `""` |
+| `translation_expr` | `str` | `""` |
+
+### SimulationParams
+
+| Field | Type | Default |
+|-------|------|---------|
+| `t_start` | `float` | `0.0` |
+| `t_end` | `float` | `1.0` |
+| `h_out` | `float` | `0.01` |
+| `h_min` | `float` | `1e-9` |
+| `h_max` | `float` | `1.0` |
+| `error_tol` | `float` | `1e-6` |
+
+### SolveContext
+
+| Field | Type | Default |
+|-------|------|---------|
+| `parts` | `list[Part]` | `[]` |
+| `constraints` | `list[Constraint]` | `[]` |
+| `motions` | `list[MotionDef]` | `[]` |
+| `simulation` | `SimulationParams` or `None` | `None` |
+| `bundle_fixed` | `bool` | `False` |
+
+**Important:** pybind11 returns copies of `list` fields, not references. Use whole-list assignment:
+
+```python
+ctx = kcsolve.SolveContext()
+p = kcsolve.Part()
+p.id = "box1"
+ctx.parts = [p]          # correct
+# ctx.parts.append(p)    # does NOT modify ctx
+```
+
+### ConstraintDiagnostic
+
+| Field | Type | Default |
+|-------|------|---------|
+| `constraint_id` | `str` | `""` |
+| `kind` | `DiagnosticKind` | `Redundant` |
+| `detail` | `str` | `""` |
+
+### SolveResult
+
+| Field | Type | Default |
+|-------|------|---------|
+| `status` | `SolveStatus` | `Success` |
+| `placements` | `list[SolveResult.PartResult]` | `[]` |
+| `dof` | `int` | `-1` |
+| `diagnostics` | `list[ConstraintDiagnostic]` | `[]` |
+| `num_frames` | `int` | `0` |
+
+### SolveResult.PartResult
+
+| Field | Type | Default |
+|-------|------|---------|
+| `id` | `str` | `""` |
+| `placement` | `Transform` | identity |
+
+## Classes
+
+### IKCSolver
+
+Abstract base class for solver backends. Subclass in Python to create custom solvers.
+
+Three methods must be implemented:
+
+```python
+class MySolver(kcsolve.IKCSolver):
+    def name(self):
+        return "My Solver"
+
+    def supported_joints(self):
+        return [kcsolve.BaseJointKind.Fixed, kcsolve.BaseJointKind.Revolute]
+
+    def solve(self, ctx):
+        result = kcsolve.SolveResult()
+        result.status = kcsolve.SolveStatus.Success
+        return result
+```
+
+Optional overrides (all have default implementations):
+
+| Method | Default behavior |
+|--------|-----------------|
+| `update(ctx)` | Delegates to `solve()` |
+| `pre_drag(ctx, drag_parts)` | Delegates to `solve()` |
+| `drag_step(drag_placements)` | Returns Success with no placements |
+| `post_drag()` | No-op |
+| `run_kinematic(ctx)` | Returns Failed |
+| `num_frames()` | Returns 0 |
+| `update_for_frame(index)` | Returns Failed |
+| `diagnose(ctx)` | Returns empty list |
+| `is_deterministic()` | Returns `True` |
+| `export_native(path)` | No-op |
+| `supports_bundle_fixed()` | Returns `False` |
+
+### OndselAdapter
+
+Built-in solver wrapping OndselSolver's Lagrangian constraint formulation. Inherits `IKCSolver`.
+
+```python
+solver = kcsolve.OndselAdapter()
+solver.name()  # "OndselSolver (Lagrangian)"
+```
+
+In practice, use `kcsolve.load("ondsel")` rather than constructing directly, as this goes through the registry.
+
+## Module functions
+
+### available()
+
+Return names of all registered solvers.
+
+```python
+kcsolve.available()  # ["ondsel"]
+```
+
+### load(name="")
+
+Create an instance of the named solver. If `name` is empty, uses the default. Returns `None` if the solver is not found.
+
+```python
+solver = kcsolve.load("ondsel")
+solver = kcsolve.load()  # default solver
+```
+
+### joints_for(name)
+
+Query supported joint types for a registered solver.
+
+```python
+joints = kcsolve.joints_for("ondsel")
+# [BaseJointKind.Coincident, BaseJointKind.Fixed, ...]
+```
+
+### set_default(name)
+
+Set the default solver name. Returns `True` if the name is registered.
+
+```python
+kcsolve.set_default("ondsel")  # True
+kcsolve.set_default("unknown")  # False
+```
+
+### get_default()
+
+Get the current default solver name.
+
+```python
+kcsolve.get_default()  # "ondsel"
+```
+
+### register_solver(name, solver_class)
+
+Register a Python solver class with the SolverRegistry. `solver_class` must be a callable that returns an `IKCSolver` subclass instance.
+
+```python
+class MySolver(kcsolve.IKCSolver):
+    def name(self): return "MySolver"
+    def supported_joints(self): return [kcsolve.BaseJointKind.Fixed]
+    def solve(self, ctx):
+        r = kcsolve.SolveResult()
+        r.status = kcsolve.SolveStatus.Success
+        return r
+
+kcsolve.register_solver("my_solver", MySolver)
+solver = kcsolve.load("my_solver")
+```
+
+## Complete example
+
+```python
+import kcsolve
+
+# Build a two-part assembly with a Fixed joint
+ctx = kcsolve.SolveContext()
+
+base = kcsolve.Part()
+base.id = "base"
+base.grounded = True
+
+arm = kcsolve.Part()
+arm.id = "arm"
+arm.placement.position = [100.0, 0.0, 0.0]
+
+joint = kcsolve.Constraint()
+joint.id = "Joint001"
+joint.part_i = "base"
+joint.part_j = "arm"
+joint.type = kcsolve.BaseJointKind.Fixed
+
+ctx.parts = [base, arm]
+ctx.constraints = [joint]
+
+# Solve
+solver = kcsolve.load("ondsel")
+result = solver.solve(ctx)
+
+print(result.status)  # SolveStatus.Success
+for pr in result.placements:
+    print(f"{pr.id}: pos={list(pr.placement.position)}")
+
+# Diagnostics
+diags = solver.diagnose(ctx)
+for d in diags:
+    print(f"{d.constraint_id}: {d.kind} - {d.detail}")
+```
+
+## Related
+
+- [KCSolve Architecture](../architecture/ondsel-solver.md)
+- [INTER_SOLVER.md](../../INTER_SOLVER.md) -- full architecture specification