ducpy.search.search_models¶

Scoped model-element search helpers.

Model elements (DucModelElement, element type "model") embed CAD/BIM content in one of two ways:

Embedded Python code (model_type == "python") whose imports reveal the real engine — ezdxf, ifcopenshell or build123d.
Linked external files (model_type is dxf / ifc / step / …) whose blobs live in the connected external files (file_ids).

Before any content can be searched we must classify each model into the engine that produced it:

Engine	Sources
ezdxf	`model_type` `dxf` / `dwg`, or Python importing `ezdxf`
ifc	`model_type` `ifc`, or Python importing `ifcopenshell`
build123d	`model_type` `step` / `stl`, or Python importing `build123d`
unsupported	anything we can’t classify

This module detects each engine and searches user-facing content for the ezdxf and IFC engines. Build123d elements currently remain searchable through their DUC label and description.

Classes¶

`ModelEngine`	CAD/BIM engine responsible for a model element's content.
`ModelElementInfo`	Engine classification of a single model element.

Functions¶

`extract_python_imports`(→ set[str])	Return the top-level module names imported by `code`.
`detect_model_engine`(→ ModelEngine)	Classify a parsed model element into a `ModelEngine`.
`model_element_info`(→ ModelElementInfo)	Build a `ModelElementInfo` from a parsed model element dict.
`iter_model_elements`(→ Iterator[dict[str, Any]])	Yield non-deleted model elements from parsed duc data.
`resolve_model_search_targets`(→ list[ModelElementInfo])	Classify every live model element in parsed duc data.
`search_duc_models`(...)	Search the user-authored text inside model elements and rank the results.

Module Contents¶

class ducpy.search.search_models.ModelEngine¶

Bases: str, enum.Enum

CAD/BIM engine responsible for a model element’s content.

Initialize self. See help(type(self)) for accurate signature.

EZDXF = 'ezdxf'¶

IFC = 'ifc'¶

BUILD123D = 'build123d'¶

UNSUPPORTED = 'unsupported'¶

class ducpy.search.search_models.ModelElementInfo¶

Engine classification of a single model element.

element_id: str¶

label: str¶

model_type: str¶

engine: ModelEngine¶

is_python: bool¶

has_code: bool¶

file_ids: tuple[str, Ellipsis]¶

ducpy.search.search_models.extract_python_imports(code: str | None) → set[str]¶

Return the top-level module names imported by code.

Uses ast for accuracy and falls back to a line-based regex when the source can’t be parsed (e.g. an extracted fragment or a syntax error).

ducpy.search.search_models.detect_model_engine(element: dict[str, Any]) → ModelEngine¶: Classify a parsed model element into a ModelEngine.

ducpy.search.search_models.model_element_info(element: dict[str, Any]) → ModelElementInfo¶: Build a ModelElementInfo from a parsed model element dict.

ducpy.search.search_models.iter_model_elements(duc_data: dict[str, Any]) → Iterator[dict[str, Any]]¶: Yield non-deleted model elements from parsed duc data.

ducpy.search.search_models.resolve_model_search_targets(duc_data: dict[str, Any]) → list[ModelElementInfo]¶: Classify every live model element in parsed duc data.

ducpy.search.search_models.search_duc_models(duc_path: str | pathlib.Path, query: str, *, output_path: str | pathlib.Path | None = None, limit: int = 50, run_code: bool = False) → ducpy.search.search_elements.DucSearchResponse¶

Search the user-authored text inside model elements and rank the results.

Loads the .duc (SQLite-backed or native binary), classifies each model element, extracts searchable DXF/DWG or IFC content, and scores it against query with the same ranking machinery as search_duc_elements(). Build123d models currently fall back to their label and description.

run_code is a trusted-input opt-in. The default (False) searches linked model files only. Setting it to True executes embedded Python model code in-process to capture generated DXF or IFC content; never enable it for untrusted DUC files. Results are written to output_path (or a default path beside the .duc) and returned as a DucSearchResponse.