The goal is to pull the full known list of product descriptions from the live api endpoint, build a search index and relationship graph over the products, and add any custom rules required to coerce products into a hierarchy that closely resembles reality (aided by selective test coverage and examples).
When it comes to importing these into ingreedy-py, rule match order will be important (in general, most-specific product matches will be the goal) - so the data export should be in a format which lends itself well to creation of unambiguous parser rules.
This changeset builds infrastructure which will supply custom product parsing content for https://github.com/openculinary/ingredient-parser/issues/9
The goal is to pull the full known list of product descriptions from the live api endpoint, build a search index and relationship graph over the products, and add any custom rules required to coerce products into a hierarchy that closely resembles reality (aided by selective test coverage and examples).
When it comes to importing these into ingreedy-py, rule match order will be important (in general, most-specific product matches will be the goal) - so the data export should be in a format which lends itself well to creation of unambiguous parser rules.