Fix: unparsed ingredients poorly formatted when fed to NLP parser

[gabrielclimb]

Summary

• Fixes [BUG] - Unparsed ingredients with quantities are poorly formatted when fed to NLP #7079
• Unparsed ingredients with a default quantity of 1 and the full text in the note field (e.g. "1/2 cup apples") were being reconstructed as "1 1/2 cup apples", causing the NLP parser to misinterpret the quantity
• Added ingredientToParserString() that prefers originalText, falls back to just the note for unparsed ingredients (no unit/food), and only uses the full reconstruction for properly parsed ingredients

Test plan

• Existing tests pass
• New unit tests for ingredientToParserString covering unparsed, originalText, parsed, and empty cases
• Manual test: create recipe with unparsed ingredient (qty=1, note="1/2 cup apples"), open parse dialog, verify NLP receives "1/2 cup apples"

[alexxxxxxxxxxandria]

Apologies that I can't just contribute this directly right now, but I recommend also adding tests for items that are less ambiguous. The "1/2 cup apples" was my go-to because you could get a fraction you wouldn't expect, but another example was "2 tbsp olive oil" being reconstructed as "1 2 tbsp olive oil" and then being parsed as "1 tbsp olive oil"

[gabrielclimb]

Apologies that I can't just contribute this directly right now, but I recommend also adding tests for items that are less ambiguous. The "1/2 cup apples" was my go-to because you could get a fraction you wouldn't expect, but another example was "2 tbsp olive oil" being reconstructed as "1 2 tbsp olive oil" and then being parsed as "1 tbsp olive oil"

Ok, I will add this

[gabrielclimb]

@alexxxxxxxxxxandria , could you take another look?