No name parsed

[laruellef]

Excellent library, thank you! (Using it for my daughter's startup idea :-)

came across a rare case of no name parsed, see below:

>>> pprint(parse_ingredient('2 pounds (16 to 20 count shrimp) 4 per person, peeled but with the tail on'))
ParsedIngredient(name=None,
                 size=None,
                 amount=[IngredientAmount(quantity=2.0,
                                          quantity_max=2.0,
                                          unit=<Unit('pound')>,
                                          text='2 pounds',
                                          confidence=0.999462,
                                          APPROXIMATE=False,
                                          SINGULAR=False,
                                          RANGE=False,
                                          MULTIPLIER=False)],
                 preparation=IngredientText(text='peeled but with the tail on',
                                            confidence=0.999639),
                 comment=IngredientText(text='(16-20 count shrimp) 4 per '
                                             'person',
                                        confidence=0.813389),
                 sentence='2 pounds (16 to 20 count shrimp) 4 per person, '
                          'peeled but with the tail on')

[strangetom]

Thanks for raising this. It looks like the punctuation in the sentence is causing shrimp to be mislabelled.

Would you mind sharing where the sentence came from (if it's from a website). This might be fixable by including more examples like this one in the training data.

[strangetom]

The latest release beta 11 has some improvements to help with cases like this.

[laruellef]

@strangetom sorry, completely missed your question, it came from foodnetwork

[laruellef]

... and great job!! 👍