Update the scraper-generator template. Inherit from abc.ABC in Abstra…

…ctScraper (#664)

recipe_scrapers/_abstract.py

@@ -1,5 +1,6 @@
 # mypy: disallow_untyped_defs=False
 import inspect
+from abc import ABC, abstractmethod
 from collections import OrderedDict
 from typing import Dict, List, Optional, Tuple, Union
 from urllib.parse import urljoin
@@ -17,7 +18,7 @@
 }
 
 
-class AbstractScraper:
+class AbstractScraper(ABC):
     page_data: Union[str, bytes]
 
     def __init__(
@@ -61,6 +62,7 @@ def __init__(
             setattr(self.__class__, "plugins_initialized", True)
 
     @classmethod
+    @abstractmethod
     def host(cls) -> str:
         """get the host of the url, so we can use the correct scraper"""
         raise NotImplementedError("This should be implemented.")
@@ -71,12 +73,14 @@ def canonical_url(self):
             return urljoin(self.url, canonical_link["href"])
         return self.url
 
+    @abstractmethod
     def title(self):
         raise NotImplementedError("This should be implemented.")
 
     def category(self):
         raise NotImplementedError("This should be implemented.")
 
+    @abstractmethod
     def total_time(self):
         """total time it takes to preparate and cook the recipe in minutes"""
         raise NotImplementedError("This should be implemented.")
@@ -130,9 +134,11 @@ def language(self):
         # Return the first candidate language
         return candidate_languages.popitem(last=False)[0]
 
+    @abstractmethod
     def ingredients(self):
         raise NotImplementedError("This should be implemented.")
 
+    @abstractmethod
     def instructions(self) -> str:
         """instructions to prepare the recipe"""
         raise NotImplementedError("This should be implemented.")
@@ -149,6 +155,7 @@ def ratings(self):
         raise NotImplementedError("This should be implemented.")
 
     def author(self):
+        # question: should we make this a required field (abstractmethod)?
         raise NotImplementedError("This should be implemented.")
 
     def cuisine(self):

recipe_scrapers/coop.py

@@ -16,6 +16,9 @@ def title(self):
     def category(self):
         return self.schema.category()
 
+    def total_time(self):
+        return self.schema.total_time()
+
     def cook_time(self):
         return self.schema.cook_time()
 

recipe_scrapers/davidlebovitz.py

@@ -1,5 +1,6 @@
 # mypy: disallow_untyped_defs=False
 from ._abstract import AbstractScraper
+from ._exceptions import RecipeScrapersExceptions
 
 
 class DavidLebovitz(AbstractScraper):
@@ -13,6 +14,11 @@ def author(self):
     def title(self):
         return self.schema.title()
 
+    def total_time(self):
+        raise RecipeScrapersExceptions(
+            f"{self.host()} does not provide time information."
+        )
+
     def image(self):
         return self.schema.image()
 

recipe_scrapers/farmhousedelivery.py

@@ -4,6 +4,7 @@
 from bs4 import Tag
 
 from ._abstract import AbstractScraper
+from ._exceptions import RecipeScrapersExceptions
 from ._utils import normalize_string
 
 """
@@ -20,6 +21,11 @@ def host(self, domain="com"):
     def title(self):
         return self.soup.find("h1", {"class": "entry-title"}).get_text(strip=True)
 
+    def total_time(self):
+        raise RecipeScrapersExceptions(
+            f"{self.host()} does not provide time information."
+        )
+
     def ingredients(self):
         # Style 1
         ingredients_marker = self.soup.find("p", string=re.compile(r"Ingredients:"))

recipe_scrapers/fredriksfikaallas.py

@@ -2,6 +2,7 @@
 import re
 
 from ._abstract import AbstractScraper
+from ._exceptions import RecipeScrapersExceptions
 
 
 class FredriksFikaAllas(AbstractScraper):
@@ -15,6 +16,11 @@ def title(self):
     def category(self):
         return self.soup.find("div", {"class": "post_category"}).get_text()
 
+    def total_time(self):
+        raise RecipeScrapersExceptions(
+            f"{self.host()} does not provide time information."
+        )
+
     def image(self):
         return self.soup.find("meta", {"property": "og:image", "content": True}).get(
             "content"

recipe_scrapers/kuchniadomowa.py

@@ -10,10 +10,17 @@ def host(cls):
     def title(self):
         return self.soup.find("h2").get_text().strip()
 
+    def total_time(self):
+        return self.schema.total_time()
+
     def image(self):
         urls = self.soup.findAll("img", {"class": "article-img", "id": "article-img-1"})
         return f"https:{urls[1]['src']}"
 
+    def ingredients(self):
+        # TODO: add implementation
+        raise NotImplementedError("This should be implemented.")
+
     def instructions(self):
         instructions = self.soup.find("div", {"id": "recipe-instructions"}).findAll(
             "li"

recipe_scrapers/kwestiasmaku.py

@@ -1,5 +1,6 @@
 # mypy: disallow_untyped_defs=False
 from ._abstract import AbstractScraper
+from ._exceptions import RecipeScrapersExceptions
 from ._utils import get_yields, normalize_string
 
 
@@ -14,6 +15,11 @@ def author(self):
     def title(self):
         return self.schema.title()
 
+    def total_time(self):
+        raise RecipeScrapersExceptions(
+            f"{self.host()} does not provide time information."
+        )
+
     def yields(self):
         return get_yields(
             self.soup.find("div", {"class": "field-name-field-ilosc-porcji"})

recipe_scrapers/maangchi.py

@@ -1,6 +1,7 @@
 # mypy: allow-untyped-defs
 
 from ._abstract import AbstractScraper
+from ._exceptions import RecipeScrapersExceptions
 from ._utils import normalize_string
 
 
@@ -18,6 +19,11 @@ def title(self):
     def category(self):
         return self.schema.category()
 
+    def total_time(self):
+        raise RecipeScrapersExceptions(
+            f"{self.host()} does not provide time information."
+        )
+
     def yields(self):
         return self.schema.yields()
 

recipe_scrapers/mykitchen101.py

@@ -4,6 +4,7 @@
 from bs4 import BeautifulSoup
 
 from ._abstract import AbstractScraper
+from ._exceptions import RecipeScrapersExceptions
 from ._utils import get_yields, normalize_string
 
 
@@ -18,6 +19,11 @@ def author(self):
     def title(self):
         return self.soup.find("h1", {"class": "entry-title"}).get_text()
 
+    def total_time(self):
+        raise RecipeScrapersExceptions(
+            f"{self.host()} does not provide time information."
+        )
+
     def yields(self):
         return get_yields(self.soup.find("p", string=re.compile("分量：")).get_text())
 

recipe_scrapers/mykitchen101en.py

@@ -1,5 +1,6 @@
 # mypy: disallow_untyped_defs=False
 from ._abstract import AbstractScraper
+from ._exceptions import RecipeScrapersExceptions
 
 
 class MyKitchen101en(AbstractScraper):
@@ -13,6 +14,11 @@ def author(self):
     def title(self):
         return self.schema.title()
 
+    def total_time(self):
+        raise RecipeScrapersExceptions(
+            f"{self.host()} does not provide time information."
+        )
+
     def yields(self):
         return self.schema.yields()
 

recipe_scrapers/owenhan.py

@@ -1,5 +1,6 @@
 # mypy: disallow_untyped_defs=False
 from ._abstract import AbstractScraper
+from ._exceptions import RecipeScrapersExceptions
 
 
 class OwenHan(AbstractScraper):
@@ -13,6 +14,11 @@ def author(self):
     def title(self):
         return self.soup.find("h1", {"class": "entry-title"}).text
 
+    def total_time(self):
+        raise RecipeScrapersExceptions(
+            f"{self.host()} does not provide time information."
+        )
+
     def image(self):
         return self.schema.image()
 

recipe_scrapers/rosannapansino.py

@@ -1,6 +1,7 @@
 # mypy: allow-untyped-defs
 
 from ._abstract import AbstractScraper
+from ._exceptions import RecipeScrapersExceptions
 from ._utils import normalize_string
 
 
@@ -12,6 +13,11 @@ def host(cls):
     def title(self):
         return self.soup.find("meta", {"property": "og:title"})["content"]
 
+    def total_time(self):
+        raise RecipeScrapersExceptions(
+            f"{self.host()} does not provide time information."
+        )
+
     def image(self):
         return self.schema.image()
 

templates/scraper.py

@@ -8,21 +8,12 @@ class Template(AbstractScraper):
     def host(cls):
         return "example.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()
 
@@ -32,6 +23,16 @@ def ingredients(self):
     def instructions(self):
         return self.schema.instructions()
 
+    def author(self):
+        # question: should we make this a required field?
+        return self.schema.author()
+
+    def category(self):
+        return self.schema.category()
+
+    def yields(self):
+        return self.schema.yields()
+
     def ratings(self):
         return self.schema.ratings()
 

templates/test_scraper.py

@@ -11,21 +11,12 @@ class TestTemplateScraper(ScraperTest):
     def test_host(self):
         self.assertEqual("example.com", self.harvester_class.host())
 
-    def test_author(self):
-        self.assertEqual(None, self.harvester_class.author())
-
     def test_title(self):
         self.assertEqual(None, self.harvester_class.title())
 
-    def test_category(self):
-        self.assertEqual(None, self.harvester_class.category())
-
     def test_total_time(self):
         self.assertEqual(None, self.harvester_class.total_time())
 
-    def test_yields(self):
-        self.assertEqual(None, self.harvester_class.yields())
-
     def test_image(self):
         self.assertEqual(None, self.harvester_class.image())
 
@@ -35,6 +26,16 @@ def test_ingredients(self):
     def test_instructions(self):
         self.assertEqual(None, self.harvester_class.instructions())
 
+    def test_author(self):
+        # should we make this a required field?
+        self.assertEqual(None, self.harvester_class.author())
+
+    def test_category(self):
+        self.assertEqual(None, self.harvester_class.category())
+
+    def test_yields(self):
+        self.assertEqual(None, self.harvester_class.yields())
+
     def test_ratings(self):
         self.assertEqual(None, self.harvester_class.ratings())
 

tests/test_davidlebovitz.py

@@ -1,3 +1,4 @@
+from recipe_scrapers._exceptions import RecipeScrapersExceptions
 from recipe_scrapers.davidlebovitz import DavidLebovitz
 from tests import ScraperTest
 
@@ -15,6 +16,9 @@ def test_author(self):
     def test_title(self):
         self.assertEqual("Faux Gras", self.harvester_class.title())
 
+    def test_total_time(self):
+        self.assertRaises(RecipeScrapersExceptions, self.harvester_class.total_time)
+
     def test_image(self):
         self.assertEqual(
             "https://www.davidlebovitz.com/wp-content/uploads/2015/06/Faux-Gras-Lentil-Pate-8.jpg",

tests/test_farmhousedelivery_1.py

@@ -1,3 +1,4 @@
+from recipe_scrapers._exceptions import RecipeScrapersExceptions
 from recipe_scrapers.farmhousedelivery import FarmhouseDelivery
 from tests import ScraperTest
 
@@ -19,6 +20,9 @@ def test_canonical_url(self):
     def test_title(self):
         self.assertEqual("Green Shakshuka", self.harvester_class.title())
 
+    def test_total_time(self):
+        self.assertRaises(RecipeScrapersExceptions, self.harvester_class.total_time)
+
     def test_ingredients(self):
         self.assertEqual(
             [

tests/test_farmhousedelivery_2.py

@@ -1,3 +1,4 @@
+from recipe_scrapers._exceptions import RecipeScrapersExceptions
 from recipe_scrapers.farmhousedelivery import FarmhouseDelivery
 from tests import ScraperTest
 
@@ -22,6 +23,9 @@ def test_title(self):
             self.harvester_class.title(),
         )
 
+    def test_total_time(self):
+        self.assertRaises(RecipeScrapersExceptions, self.harvester_class.total_time)
+
     def test_ingredients(self):
         self.assertEqual(
             [