Environment & introspection
Discover what your Kratos installation provides. Catalog tools (elements, conditions, constitutive laws, processes) parse the Kratos C++/Python sources; runtime tools query the live build through the worker bridge.
kratos_check_installation
needs Kratos (gracefully reports when unavailable)
Verify that Kratos is importable and report the environment.
Parameters: none.
Returns: kratos_root, pythonpath, ld_library_path, pip_installed, importable, version, build_type, num_threads, is_distributed, compiled_applications (list). When Kratos is missing, importable: false plus a hint.
{
"importable": true,
"version": "10.4.2-...-Release-x86_64",
"num_threads": 20,
"compiled_applications": ["ConvectionDiffusionApplication", "FluidDynamicsApplication", "..."]
}When Kratos is missing, the hint field points at kratos_install.
kratos_install
Install Kratos Multiphysics via pip into the server's own Python environment — no local build required. Official wheels exist for Linux and Windows x86_64 only; on macOS or other unsupported platforms this fails and you need a local KRATOS_ROOT build instead (see Installation).
| Parameter | Type | Description |
|---|---|---|
applications | string[]? | application names to install alongside the core, e.g. ["StructuralMechanicsApplication", "LinearSolversApplication"] — each is mapped to its PyPI name by prefixing Kratos (so StructuralMechanicsApplication → KratosStructuralMechanicsApplication) |
all | bool | install the KratosMultiphysics-all omnibus package (essentially every application) instead of naming individual ones; largest download, simplest choice (default false) |
upgrade | bool | pass --upgrade to pip (default false) |
Returns: {ok, packages, returncode}; on failure stderr_tail and a hint; on success a nested check — the same payload kratos_check_installation returns, so a single call installs and confirms. No server restart is needed — the very next tool call sees the new install.
// kratos_install(applications=["StructuralMechanicsApplication", "LinearSolversApplication"])
{
"ok": true,
"packages": ["KratosMultiphysics", "KratosStructuralMechanicsApplication", "KratosLinearSolversApplication"],
"returncode": 0,
"check": { "importable": true, "version": "...", "compiled_applications": ["StructuralMechanicsApplication", "..."] }
}Which applications to install
Match them to the template you're using — kratos_list_solvers and list_templates name the required_applications for each analysis type. LinearSolversApplication is needed by essentially every template (it provides sparse_lu, the default linear solver).
kratos_list_applications
List every application in the source tree with a compiled flag.
Parameters: none. Returns: {applications: [{name, compiled}], num_compiled, num_source}.
kratos_list_elements
List element type names parsed from KRATOS_REGISTER_ELEMENT macros.
| Parameter | Type | Description |
|---|---|---|
application | string? | e.g. StructuralMechanicsApplication (or Core) |
name_filter | string? | case-insensitive substring, e.g. SmallDisplacement |
Returns: [{name, application, compiled}].
kratos_list_conditions
Same parameters/shape as kratos_list_elements, for conditions (the boundary entities used by loads and fluxes — see the MDPA guide for the crucial load-condition vs geometric-condition distinction).
kratos_list_constitutive_laws
Same parameters/shape, for material models (LinearElastic3DLaw, Newtonian2DLaw, ...).
kratos_list_variables
needs Kratos
List Kratos variables grouped by type, from the live kernel.
| Parameter | Type | Description |
|---|---|---|
type_filter | string? | one of double, array_1d_3, bool, int, vector, matrix, ... |
name_filter | string? | substring, e.g. TEMP |
Returns: {type: [names]} — e.g. {"array_1d_3": ["DISPLACEMENT", ...]}.
kratos_list_solvers
List known solver_type values per analysis type with the Python module implementing each, plus all *_solver.py modules found in the source tree.
| Parameter | Type | Description |
|---|---|---|
analysis_type | string? | structural, thermal, fluid, potential_flow |
kratos_list_processes
List Python process modules usable as python_module in ProjectParameters process lists.
| Parameter | Type | Description |
|---|---|---|
application | string? | filter by owning application (or Core) |
name_filter | string? | substring, e.g. assign_vector |
with_defaults | bool | also attach each module's default_settings and param_types, parsed from source (best effort; default false) |
kratos_get_process_defaults
Return a process' default settings — the parameter names, default values, coarse types (bool/number/string/array/json/null) and which parameters are model-part references — parsed straight from the process' Python source (its ValidateAndAssignDefaults block, or a GetDefaultParameters classmethod). This is the schema you need to author a process block in a process list, the same way kratos_get_solver_defaults gives you a solver's schema.
This is pure AST parsing of the Kratos source (no live build needed), ported from the sibling Flowgraph project's parse-processes.py. It also backs the auto-filled defaults in add_boundary_condition / add_output_process.
| Parameter | Type | Description |
|---|---|---|
python_module | string | a value from kratos_list_processes, e.g. assign_scalar_variable_process, vtk_output_process |
Returns: {python_module, default_settings, param_types, input_model_parts, output_params, help}, or {"error": ...} when the source is unavailable, the module is unknown, or it declares defaults in a non-standard way (e.g. C++-validated output processes such as vtk_output_process, whose defaults live in C++).
// kratos_get_process_defaults("assign_scalar_variable_process") → (excerpt)
{
"default_settings": { "model_part_name": "please_specify", "variable_name": "SPECIFY_VARIABLE",
"value": 0.0, "constrained": true, "interval": [0.0, 1e30] },
"param_types": { "variable_name": "string", "value": "number", "constrained": "bool" },
"input_model_parts": ["model_part_name"]
}kratos_get_solver_defaults
needs Kratos
Return the complete default solver_settings for a solver, straight from its GetDefaultParameters() — the authoritative schema of every accepted key.
| Parameter | Type | Description |
|---|---|---|
analysis_type | string | structural / thermal / fluid / potential_flow |
solver_type | string | a value from kratos_list_solvers, e.g. Static, transient, Monolithic |
// kratos_get_solver_defaults("structural", "Static") → (excerpt)
{
"solver_type": "mechanical_solver",
"model_import_settings": { "input_type": "mdpa", "input_filename": "unknown_name" },
"linear_solver_settings": {},
"convergence_criterion": "residual_criterion",
"max_iteration": 10
}