feat: Use recipe_scrapers metadata for more accurate parsing

[eric-hoffmann]

What this PR does / why we need it:

Some recipe websites have additional nutrition information not contained in the schema data. The recipe-scrapers library has a nutrients() function which implements custom parsing for some sites, but Mealie is currently pulling nutrition information directly from the schema. This PR updates Mealie’s recipe-scrapers nutrition parsing to use the nutrients function. Existing parsing logic will gracefully fall back to direct parse from schema if the nutrients() function fails for whatever reason.

Which issue(s) this PR fixes:

Fixes #5164

Testing

Tested parsing for recipes exhibiting this issue with a local Mealie instance:

Example recipe: https://cookwell.com/recipe/crispy-oven-fries

Info captured by recipe-scrapers library:

from urllib.request import urlopen
from recipe_scrapers import scrape_html
url = "https://cookwell.com/recipe/crispy-oven-fries"
html = urlopen(url).read().decode("utf-8")
scraper = scrape_html(html,url)

#Data directly from schema:
scraper.schema.data['nutrition']
{'@type': 'NutritionInformation', 'calories': '2558 calories'}
#Data captured by nutrients function:
scraper.nutrients()
{'calories': '2558', 'carbohydrateContent': '402 g', 'fatContent': '49 g', 'proteinContent': '132 g'}

 
Nutrition ingredients captured from schema (Missing all information except calories):

With update, nutrition info captured from scraper.nutrients():

[michael-genson]

Thanks for this! Would you be able to do the same for other fields that recipe scrapers includes? I know image is one of them (scraped_data.image()) but there are probably more.

[eric-hoffmann]

Sure, updated the other fields that had a direct equivalent parser function. scraped_data.url is always populated by the library verbatim from what Mealie passes it, so I left that one alone.

[michael-genson]

Thank you!