eds: Support custom_options for unknown EDS fields

[bizfsc]

Supersedes #615.

When importing an EDS/DCF file, any key that is not part of the CiA 306 standard field set is now collected into a custom_options dict on the object. This allows applications to round-trip vendor-specific metadata without losing it.

Changes

• ODVariable, ODRecord and ODArray gain a custom_options: dict[str, str] attribute (default {}).
• New _STANDARD_OPTIONS set in eds.py lists all keys that are parsed explicitly. ObjFlags and Denotation are intentionally not in this set — they flow through custom_options and survive round-trips until first-class support is added in eds: Add first-class support for CiA 306 ObjFlags and Denotation #654.
• New _get_custom_options(eds, section) helper collects every key not in _STANDARD_OPTIONS.
• custom_options is populated for top-level VAR/DOMAIN, ARRAY and RECORD objects as well as for sub-variables via build_variable().
• copy_variable() now gives each copied variable its own custom_options dict to avoid shared-state mutations between shallow-copy clones.
• export_variable() and export_record() write custom_options back to the EDS file, guarded by a _STANDARD_OPTIONS check to prevent user-supplied keys from clobbering standard fields.

Tests

• test_reading_custom_options — VAR with Category=Motor, Offset=100
• test_custom_options_standard_keys_excluded — standard keys absent from custom_options
• test_custom_options_empty_for_standard_object — heartbeat time has {}
• test_custom_options_record — RECORD-level custom_options
• test_roundtrip_custom_options — full EDS export/import cycle preserves values
• test_roundtrip_custom_options_not_duplicated_as_standard — standard keys stay out after round-trip

[codecov]

Codecov Report

❌ Patch coverage is 91.30435% with 2 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
canopen/objectdictionary/eds.py	88.88%	0 Missing and 2 partials ⚠️

📢 Thoughts on this report? Let us know!

[Copilot]

Pull request overview

Adds support for round-tripping vendor-specific (non-CiA-306) EDS/DCF keys by collecting them into a new custom_options dict on object dictionary objects and exporting them back out.

Changes:

• Introduces custom_options on ODVariable, ODRecord, and ODArray to store unknown EDS keys.
• Updates EDS import/export to populate and persist custom_options using a _STANDARD_OPTIONS allowlist and _get_custom_options() helper.
• Adds tests and sample EDS entries covering read + export/import round-trip behavior.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 4 comments.

File	Description
canopen/objectdictionary/eds.py	Collects unknown keys into custom_options during import and writes them back during export.
canopen/objectdictionary/__init__.py	Adds custom_options attribute to OD object types.
test/test_eds.py	Adds unit tests for reading and round-tripping custom_options.
test/sample.eds	Adds sample objects containing vendor-specific keys to validate behavior.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 6 comments.

[Copilot]

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

[acolomb]

With all the AI output in between, it's hard to keep track of what changed. I'd really prefer not force-pushing additional fixes (or even merging the master branch). Much easier to see what's going on with simple commits on top, ideally with a one-line mention what was touched. Will be squashed away when merging. Thanks!

[acolomb]

Looks great overall, but one question about the shared mutable dicts, see below.

[acolomb]

On second thought, this attribute always exists. Let's avoid the potential creation of non-shared dicts here as well.

Suggested change

	var.custom_options = getattr(template, "custom_options", {})
	var.custom_options = template.custom_options

[acolomb]

Or handle it as part of the enumerated attributes? The existence check is not really needed in this specific case, but still simpler to use a common code path. If you decide to do that, please switch to getattr / setattr in the process.

[acolomb]

This is what I had in mind: (GutHub won't render it correctly as a suggestion)

            for attr in ("data_type", "unit", "factor", "min", "max", "default",
                         "access_type", "description", "value_descriptions",
                         "bit_definitions", "storage_location", "custom_options"):
                if (template_value := getattr(template, attr)) is not None:
                    setattr(var, attr, template_value)

[acolomb]

The two remaining "partial" lines from the Codecov report can be ignored IMHO, that's only the safety check to not rewrite standard options.