Merge branch 'main' into v14

Conflicts:
	recipe_scrapers/__init__.py
	recipe_scrapers/americastestkitchen.py
	tests/test_data/americastestkitchen.com/americastestkitchen.json
	tests/test_data/cookscountry.com/cookscountry.json
	tests/test_data/cooksillustrated.com/cooksillustrated.json

docs/how-to-develop-scraper.md

@@ -178,7 +178,6 @@ The generated json file will look something like this, with only the host field
     "language": "",
     "title": "",
     "ingredients": "",
-    "instructions": "",
     "instructions_list": "",
     "total_time": "",
     "yields": "",

generate.py

@@ -40,7 +40,6 @@ def generate_scraper_test(class_name, host_name):
         "language": "",
         "title": "",
         "ingredients": "",
-        "instructions": "",
         "instructions_list": "",
         "total_time": "",
         "yields": "",

recipe_scrapers/__init__.py

@@ -5,7 +5,11 @@
 from typing import Any
 
 from ._abstract import AbstractScraper
-from ._exceptions import NoSchemaFoundInWildMode, WebsiteNotImplementedError
+from ._exceptions import (
+    NoSchemaFoundInWildMode,
+    RecipeSchemaNotFound,
+    WebsiteNotImplementedError,
+)
 from ._factory import SchemaScraperFactory
 from ._utils import get_host_name
 from .aberlehome import AberleHome

recipe_scrapers/_exceptions.py

@@ -55,6 +55,15 @@ class SchemaOrgException(FillPluginException):
     ...
 
 
+class RecipeSchemaNotFound(SchemaOrgException):
+    """No recipe schema metadata found on the page"""
+
+    def __init__(self, url):
+        self.url = url
+        message = f"No Recipe Schema found at {self.url}."
+        super().__init__(message)
+
+
 class StaticValueException(RecipeScrapersExceptions):
     """Error to communicate that the scraper is returning a fixed/static value."""
 

recipe_scrapers/_utils.py

@@ -325,3 +325,20 @@ def get_abstract_methods():
         and name not in {"links", "to_json"}  # ... and not excluded as special-cases.
         or name == "host"  # ... explicitly include the `host` method.
     ]
+
+
+def get_nutrition_keys():
+    return [
+        "servingSize",
+        "calories",
+        "fatContent",
+        "saturatedFatContent",
+        "unsaturatedFatContent",
+        "transFatContent",
+        "carbohydrateContent",
+        "sugarContent",
+        "proteinContent",
+        "sodiumContent",
+        "fiberContent",
+        "cholesterolContent",
+    ]

recipe_scrapers/bakels.py

@@ -13,9 +13,6 @@ def author(self):
     def title(self):
         return self.soup.find("h1").get_text()
 
-    def image(self):
-        return self.schema.image()
-
     def ingredients(self):
         div = self.soup.find("div", id="tab-ingredients_1")
         if not div:

recipe_scrapers/bigoven.py

@@ -22,6 +22,3 @@ def ingredient_groups(self):
     def instructions(self):
         ps = self.soup.find("div", {"class": "instructions"}).findAll("p")
         return "\n".join([normalize_string(p.text) for p in ps])
-
-    def ratings(self):
-        return self.schema.ratings()

recipe_scrapers/chefnini.py

@@ -15,9 +15,6 @@ def author(self):
     def title(self):
         return self.soup.find("span", {"itemprop": "headline"}).get_text()
 
-    def category(self):
-        return self.schema.category()
-
     def total_time(self):
         raise FieldNotProvidedByWebsiteException(return_value=None)
 

recipe_scrapers/cookinglight.py

@@ -1,15 +1,12 @@
 from ._abstract import AbstractScraper
-from ._utils import get_yields, normalize_string
+from ._utils import normalize_string
 
 
 class CookingLight(AbstractScraper):
     @classmethod
     def host(cls):
         return "cookinglight.com"
 
-    def yields(self):
-        return get_yields(self.schema.yields())
-
     def ingredients(self):
         ingredients = self.soup.find("div", {"class": "ingredients"}).ul.findAll("li")
         return "\n".join(

recipe_scrapers/dobruchutaktualitysk.py

@@ -12,27 +12,6 @@ def keywords(self):
         # site has no support for recipe keywords
         raise FieldNotProvidedByWebsiteException(return_value=None)
 
-    def author(self):
-        return self.schema.author()
-
-    def title(self):
-        return self.schema.title()
-
-    def category(self):
-        return self.schema.category()
-
-    def total_time(self):
-        return self.schema.total_time()
-
-    def yields(self):
-        return self.schema.yields()
-
-    def image(self):
-        return self.schema.image()
-
-    def ingredients(self):
-        return self.schema.ingredients()
-
     def instructions(self):
         p_list = self.soup.find("div", {"class": "procedure-list"}).findChildren("p")
 
@@ -44,15 +23,6 @@ def instructions(self):
 
         return "\n".join(instructions)
 
-    def ratings(self):
-        return self.schema.ratings()
-
-    def cuisine(self):
-        return self.schema.cuisine()
-
-    def description(self):
-        return self.schema.description()
-
     def ingredient_groups(self):
         return group_ingredients(
             self.ingredients(),

recipe_scrapers/felixkitchen.py

@@ -30,9 +30,6 @@ def yields(self):
         yields_strong = container_p.find("strong")
         return yields_strong.text
 
-    def image(self):
-        return self.schema.image()
-
     def ingredients(self):
         ingredients = []
 

recipe_scrapers/fifteengram.py

@@ -11,24 +11,3 @@ def canonical_url(self):
 
     def author(self):
         return "15gram"
-
-    def title(self):
-        return self.schema.title()
-
-    def total_time(self):
-        return self.schema.total_time()
-
-    def yields(self):
-        return self.schema.yields()
-
-    def image(self):
-        return self.schema.image()
-
-    def ingredients(self):
-        return self.schema.ingredients()
-
-    def instructions(self):
-        return self.schema.instructions()
-
-    def description(self):
-        return self.schema.description()

recipe_scrapers/gourmettraveller.py

@@ -10,12 +10,6 @@ class GourmetTraveller(AbstractScraper):
     def host(cls):
         return "gourmettraveller.com.au"
 
-    def author(self):
-        return self.schema.author()
-
-    def title(self):
-        return self.schema.title()
-
     def category(self):
         recipe_category_span = self.soup.find(
             "span", {"class": "related-tags__label"}, string=re.compile("Recipe Course")
@@ -25,21 +19,6 @@ def category(self):
         value = recipe_category_span.find_next_sibling("span")
         return normalize_string(value.text)
 
-    def total_time(self):
-        return self.schema.total_time()
-
-    def prep_time(self):
-        return self.schema.prep_time()
-
-    def cook_time(self):
-        return self.schema.cook_time()
-
-    def yields(self):
-        return self.schema.yields()
-
-    def image(self):
-        return self.schema.image()
-
     def ingredients(self):
         group_heading_divs = self.soup.find_all(
             "div", {"class": "recipe-ingredients__title"}
@@ -58,15 +37,6 @@ def ingredients(self):
 
         return ingredients
 
-    def instructions(self):
-        return self.schema.instructions()
-
-    def cuisine(self):
-        return self.schema.cuisine()
-
-    def description(self):
-        return self.schema.description()
-
     def ingredient_groups(self):
         return group_ingredients(
             self.ingredients(),

recipe_scrapers/halfbakedharvest.py

@@ -7,9 +7,6 @@ class HalfBakedHarvest(AbstractScraper):
     def host(cls):
         return "halfbakedharvest.com"
 
-    def nutrients(self):
-        return self.schema.nutrients()
-
     def ingredient_groups(self):
         return group_ingredients(
             self.ingredients(),

recipe_scrapers/hellofresh.py

@@ -33,6 +33,3 @@ def total_time(self):
         prep_time, cook_time = self.prep_time(), self.cook_time()
         if prep_time or cook_time:
             return (prep_time or 0) + (cook_time or 0)
-
-    def nutrients(self):
-        return self.schema.nutrients()

recipe_scrapers/jamieoliver.py

@@ -7,9 +7,6 @@ class JamieOliver(AbstractScraper):
     def host(cls):
         return "jamieoliver.com"
 
-    def image(self):
-        return self.schema.image()
-
     def instructions(self):
         instructions = self.soup.find("ol", {"class": "recipeSteps"}).findAll("li")
         return "\n".join([normalize_string(inst.get_text()) for inst in instructions])

recipe_scrapers/jocooks.py

@@ -8,27 +8,6 @@ class JoCooks(AbstractScraper):
     def host(cls):
         return "jocooks.com"
 
-    def author(self):
-        return self.schema.author()
-
-    def title(self):
-        return self.schema.title()
-
-    def category(self):
-        return self.schema.category()
-
-    def total_time(self):
-        return self.schema.total_time()
-
-    def yields(self):
-        return self.schema.yields()
-
-    def image(self):
-        return self.schema.image()
-
-    def ingredients(self):
-        return self.schema.ingredients()
-
     def ingredient_groups(self):
         return group_ingredients(
             self.ingredients(),
@@ -37,21 +16,6 @@ def ingredient_groups(self):
             ".wprm-recipe-ingredient",
         )
 
-    def instructions(self):
-        return self.schema.instructions()
-
-    def ratings(self):
-        return self.schema.ratings()
-
-    def cuisine(self):
-        return self.schema.cuisine()
-
-    def description(self):
-        return self.schema.description()
-
-    def nutrients(self):
-        return self.schema.nutrients()
-
     def equipment(self):
         equipment_items = [
             item.get_text()

recipe_scrapers/kitchensanctuary.py

@@ -8,33 +8,6 @@ class KitchenSanctuary(AbstractScraper):
     def host(cls):
         return "kitchensanctuary.com"
 
-    def author(self):
-        return self.schema.author()
-
-    def title(self):
-        return self.schema.title()
-
-    def category(self):
-        return self.schema.category()
-
-    def total_time(self):
-        return self.schema.total_time()
-
-    def prep_time(self):
-        return self.schema.prep_time()
-
-    def cook_time(self):
-        return self.schema.cook_time()
-
-    def yields(self):
-        return self.schema.yields()
-
-    def image(self):
-        return self.schema.image()
-
-    def ingredients(self):
-        return self.schema.ingredients()
-
     def ingredient_groups(self):
         ingredient_groups = self.soup.find_all(
             "div", {"class": "wprm-recipe-ingredient-group"}
@@ -68,15 +41,3 @@ def ingredient_groups(self):
             ingredient_index_base = ingredient_count
 
         return results
-
-    def instructions(self):
-        return self.schema.instructions()
-
-    def ratings(self):
-        return self.schema.ratings()
-
-    def cuisine(self):
-        return self.schema.cuisine()
-
-    def description(self):
-        return self.schema.description()

recipe_scrapers/lekkerensimpel.py

@@ -17,9 +17,6 @@ def title(self):
         title = self.soup.find("h1", {"class": "hero__title"}).text
         return normalize_string(title)
 
-    def category(self):
-        return self.schema.category()
-
     def image(self):
         image = self.soup.find("meta", {"property", "og:image"})
         return image["content"] if image else None