The mesh_tools.py module#

Summary#

mesh_quality_impl

Return live mesh sizing and quality metrics from the connected solver.

Description#

Solve-leaf mesh-introspection domain tools (shared implementation).

mesh_quality is a pure-domain tool: it reads only from Backend.mesh_quality(), Backend.mesh_check() and Backend.mesh_counts(), so it is safe to expose from the standalone MCP leaf. Richer mesh discovery that depends on named-object or schema caches is out of scope for the leaf and lives on host-side callers.

Why cell/face/node counts are always included here: Fluent does not expose mesh-element totals on the settings tree, and PyFluent’s solver.mesh.* namespace has no cell_count attribute. Without these counts, an agent has no reliable accessor for the canonical “how many cells does my mesh have?” question and would resort to guessing non-existent attributes (such as solver.field_info.get_cell_count, solver.mesh.cell_count, and solver.mesh_info.cell_count). The counts come from (inquire-grids) via Backend.mesh_counts(). Surfacing them here makes mesh_quality the one-stop tool for every mesh size/quality/topology question.

Module detail#

async mesh_tools.mesh_quality_impl(backend: ansys.fluent.mcp.common.backend.Backend, *, include_check: bool = False) → dict[str, Any]#

Return live mesh sizing and quality metrics from the connected solver.

Output shape:

  • cell_count/face_count/ node_count: Globals from (inquire-grids). None is returned when the probe failed (no mesh loaded or partition pending). Always present at the top level so “how many cells does my mesh have?” is answered in a single tool call.

  • quality: {min_orthogonal_quality, max_ortho_skew, max_aspect_ratio} from mesh.quality.

  • check (only when include_check=True): Structured mesh.check payload with domain extents, volume/face-area statistics, warnings, errors and a trimmed raw transcript.

Parameters:
backendBackend

Backend to supply to the function.

include_checkbool

Include check to supply to the function.

Returns:
dict[str, Any]

Mapping containing the operation result.