resolve_model#

scikitplot.cleanprompt.resolve_model(language='en', size='sm', explicit=None)[source]#

Resolve a language and size into a spaCy model name.

Parameters:
languagestr, default=’en’

Language code.

sizestr, default=’sm’

Preferred model size; one of MODEL_SIZES.

explicitstr, optional

A model name that overrides everything else.

Returns:
modelstr

The model name to load.

notedict

What was decided: language, requested_size, resolved_size, fallback and, when a substitution happened, reason.

Raises:
PolicyError

If size is not a known size.

Parameters:

  • language (str)

  • size (str)

  • explicit (str | None)

Return type:

tuple[str, dict[str, Any]]

Notes

User notes. When fallback is True the multilingual model was substituted, and it recognises only broad categories. Treat thin results as a limitation of that model rather than as an empty document.

Examples

>>> resolve_model("en", "sm")[0]
'en_core_web_sm'
>>> resolve_model("de", "lg")[0]
'de_core_news_lg'
>>> model, note = resolve_model("tr")
>>> model, note["fallback"]
('xx_ent_wiki_sm', True)
>>> resolve_model("de", explicit="my_model")[0]
'my_model'