Getting started in Python
The Python package is published on PyPI as pycadkit and imported as cadkit (the name
cadkit was taken on PyPI). It wraps the Rust library with PyO3: parsing, rendering and
writing run in Rust with the GIL released, so threads can read files in parallel.
Install
Section titled “Install”pip install pycadkitOne abi3 wheel per platform serves Python 3.9 and newer: Linux x86_64 and aarch64
(manylinux), macOS arm64 and x86_64, and Windows x64. Other platforms build from the source
distribution, which needs a Rust toolchain. Type stubs (.pyi, py.typed) are included.
Read a drawing and export SVG
Section titled “Read a drawing and export SVG”import cadkit
doc = cadkit.read("plan.dxf") # DWG, DGN, DXF or CityGML; the format is detectedprint(doc.format, doc.version, doc.units["unit"])for model in doc.models: print(f"{model['name']}: {len(model['entities'])} entities")for warning in doc.warnings: print("warning:", warning["code"], warning["message"])
with open("plan.svg", "w", encoding="utf-8") as f: f.write(doc.to_svg(width=1200))cadkit.read accepts a path (str or os.PathLike) or bytes (bytes, bytearray,
memoryview) and detects the format. The Document exposes format, version,
application, codepage, units, layers, blocks, models and warnings as plain
dicts and lists shaped like the document JSON.
doc.entities(model=0) iterates the entities of one model. A Document is immutable and its
dict properties are rebuilt on each access: keep them in a local variable inside loops.
Other outputs:
| Call | Returns |
|---|---|
doc.to_svg(width=1600, background=None, ...) |
SVG text |
doc.to_dxf(version="r2018") |
ASCII DXF text, r12 to r2018 |
doc.to_json(pretty=True), doc.to_dict() |
the whole document |
doc.to_dgn(seed) |
DGN V8 bytes; see Write DGN V8 from a seed |
doc.to_citygml(srs_name, lod=1) |
CityGML 2.0 bytes; see CityGML 2.0 |
cadkit.inspect(path) |
header-only metadata as a dict |
Errors
Section titled “Errors”cadkit.CadkitError is the base class of UnknownFormatError (not a recognised drawing),
UnsupportedError (valid input using an unimplemented feature), LimitExceededError (over a
resource limit) and InvalidDataError (truncated or malformed data). Missing or unreadable
files raise the usual OSError subclasses. The max_input_bytes, max_decompressed_bytes,
max_objects and max_depth keywords of read lower the limits; see
Handling untrusted files.