Mini-Weaver

[yelenacox]

• Added command for inline expansion, mini
• Looks for exclusions in the YAML file and does not include those codes in the expanded output
  • Tree:
    • minus - permissible_values - [curie]
    • minus - reachable_from - source_nodes - [curie]
• Added function to remove permissible_values for easier re-expansion

[torstees]

This looks good, and does seem to do the trick once you get the input correct. However, that expand function should probably be split up. A general rule of thumb is to have each function have a single overarching purpose. Breaking things into smaller, more specialized pieces makes your code more modular, is easier to read since the code that sits together serves a common objective, etc.

[torstees]

I recommend dropping the local init_logging and add car-utils to this and use that one instead.

[torstees]

This function is doing a lot — parsing the model file, resolving enum imports, computing exclusion codes, shelling out to dragon_search per node, and writing results back to disk, all in one ~100 line function with 5+ levels of nesting. Might be worth splitting into a few smaller pieces, e.g.:

_resolve_enum_imports(model_parsed, local_filepath) — the import-filtering/glob-matching block
_compute_minus_codes(reachable) — the minus/minus_codes logic is a nice self-contained unit already
_expand_enum_for_node(node, ontology, expanded_enum, endpoint, iri) — the dragon_search subprocess call + parsed_csv handling
_write_expanded_enum(...) — the yaml dump/write at the end

That would make expand_mini read as an orchestrator (loop over imports → loop over enums → loop over nodes) rather than mixing I/O, subprocess calls, and logic together. Not blocking, just flagging since it'll get harder to touch safely as it grows.