Repository navigation
Claude/dreamy goldberg wjzix8 - #2
Merged
DHoffmannOVGU merged 3 commits intoOct 17, 2026
Merged
DHoffmannOVGU merged 3 commits into
DHoffmannOVGU merged 3 commits into
Conversation
The documented Python API builds models with snake_case names, such as
InternalElement(name="Motor"). Pydantic accepts that at runtime through
populate_by_name, but type checkers name a model's __init__ parameters
after the alias of the assigned Field. pyright, and therefore Pylance in
VS Code, rejected every snake_case argument and demanded the PascalCase
alias instead: a 13-line sample using the documented API produced 37
false errors.
Move the 105 CAEX aliases in models.py into Annotated metadata, which
type checkers ignore, and keep the defaults on the right-hand side:
name: Annotated[str, Field(alias="Name")]
unit: Annotated[str | None, Field(alias="Unit")] = None
The rewrite was mechanical. Verified that the runtime model is
unchanged: every field's alias, validation and serialization alias,
annotation, required flag, default, default factory, and description is
identical to main for all models, the generated JSON Schema is
byte-identical, and all tests pass. On the sample, pyright now reports
only the two deliberate mistakes (a missing required name and a
misspelled keyword), as mypy already did. Errors on the SDK's own
source drop from 163 to 58 under pyright and from 167 to 86 under mypy,
with mypy call-arg errors going from 84 to 0.
A test keeps aliases out of assigned Fields, and the design notes
explain the convention for future fields.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VzX3UpZysYhwy1ZMJNcVvq
The builders covered only the instance side. Authoring a library meant constructing models directly, and the obvious attempt fails: SystemUnitClassLib(system_unit_classes=[SystemUnitClass(...)]) raises a Pydantic type error because libraries hold the hierarchical SystemUnitFamily, RoleFamily, and InterfaceFamily models. The README tutorial had to explain that distinction to beginners. Add system_unit_class, role_class, interface_class, and attribute_type, plus a *_lib builder for each, following the existing conventions: attributes accept a mapping or a list of attribute() results (now one shared helper, also used by internal_element), role_paths become SupportedRoleClass entries just as internal_element turns them into RoleRequirements, children holds nested InternalElements, and classes nests further classes in the library tree. Tests build a document from builders alone and check that every kind of reference resolves with no validation issues, that a nested class inherits from its base when instantiated, and the exact AML JSON emitted. The README tutorial now uses the builders, and every snippet still runs on the base install. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VzX3UpZysYhwy1ZMJNcVvq
Three rough edges a newcomer hits early: - Using OPC UA or RDF conversion on a base install failed with a bare "No module named 'lxml'" (or 'rdflib'), with no hint that an extra exists. The modules that need those packages now raise MissingDependencyError, a ModuleNotFoundError subclass, whose message names the command, for example: python -m pip install "automationml[opcua]" - A malformed or non-CAEX XML file raised LegacyConversionError, even for CAEX 3.0 input that has nothing to do with legacy conversion. The three "not a readable CAEX document" cases (unparsable XML, wrong root, unsupported schema version) now raise AMLParseError. It subclasses LegacyConversionError so existing handlers keep working. CAEX 2.15 schema and upgrade failures are unchanged. - The package had no __version__, and source_document() recorded a hard-coded OriginVersion of "0.1.0" while the package is 0.1.0a1. The version now lives once in automationml/_version.py; hatchling reads it through [tool.hatch.version], the package exports it, and source_document() uses it by default. The base AutomationMLError and CAEXValidationError are now importable from the top-level package as well. Verified on a real base install without lxml or rdflib that both conversions raise the new message. The built wheel's metadata reports 0.1.0a1 and twine check passes. The TestPyPI rename step still matches pyproject.toml. All tests pass, every README and OPC UA guide snippet runs, the learning notebooks execute, and the OPC UA release gate stays at 121/121. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01VzX3UpZysYhwy1ZMJNcVvq
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Describe the AutomationML use case and the change made.
Compatibility
Verification
python -m pytestpasses locally.CHANGELOG.mdwas updated for user-visible changes.Additional context
List relevant CAEX paths, schema sections, issues, or design decisions.