Redesign for 1.0.0

[micahcochran]

Need a class RecipeSchema. This is a class that will store the data in JSON-like dictionary and also as a python-object dictionary in one class.

The functions load(), loads(), scrape(), and scrape_url(), should return an list of RecipeSchema objects.

The RecipeSchema.__init__ will take json_style dictionaries recipe_json, and python object dictionaries recipe_dict. A RecipeSchema class should help code reuse and I think this is a better design paradigm than the entirely functional code. It will add an additional step to decoding, but future class function could likely be added without a major overhaul.

Here's some ideas

>>> import scrape_schema_recipe

>>> url = 'https://www.foodnetwork.com/recipes/alton-brown/honey-mustard-dressing-recipe-1939031'
>>> recipe_list = scrape_schema_recipe.scrape_url(url)  # remove that --> python_objects=True

# OLD WAY
>>> recipe = recipe_list[0]

# NEW WAY
>>> recipe = recipe_list[0].dict
# OR
>>> recipe = dict(recipe_list[0])
# OR if you need the JSON-like dictionary version
>>> recipe = recipe_list[0].json

# to get Name of the recipe
>>> recipe['name']
'Honey Mustard Dressing'

I could still be convinced that one syntax might be better than another.

This is a design decision to entirely remove the python_objects parameter from the functions. It is a little more complicated than it should be. The RecipeSchema.dict and/or dict(RecipeSchema) representation(s) should have that functionality.

These will be breaking changes to the API from version 0.1.5.

TODO:

• Redesign with OOP (above)
• use dataclasses (use dataclasses backport to support Python 3.6 or just leave 3.6 in the dust)
• Create SSR's own TypeError exception that is raised, perhaps SSRTypeError? This way Python TypeErrors won't be caught unnecessarily. -- Done in Release v0.2.0
• Remove code duplication in functions load(), loads(), scrape(), and scrape_url(). (pretty minor)
• Fix sub key date parsing. See edit below for details.

If designed correctly, these changes might make issue #19 a moot point. Might need __set_attr__ function or a specific function like define_default_value to make that happen.

Edit: There are subkeys in which dates that are not being parsed. 'video' -> 'uploadDate'

Here's 0.2.0 code:

In [1]: import scrape_schema_recipe                                             

In [3]: s=scrape_schema_recipe.scrape_url("https://www.midgetmomma.com/olive-garden-at-home-alfredo-sauce-salad-dressing-recipe/", python_objects=True)                      

In [4]: s[0]["video"]["uploadDate"]                            
Out[4]: '2020-04-10T16:41:25.000Z'

Here's proposed 1.0.0 code:

In [1]: import scrape_schema_recipe                                             

In [3]: s=scrape_schema_recipe.scrape_url("https://www.midgetmomma.com/olive-garden-at-home-alfredo-sauce-salad-dressing-recipe/")                      

In [4]: s[0].py["video"]["uploadDate"]                                          
Out[4]: '2020-04-10T16:41:25.000Z'

In both cases those should be returning datetime.datetime(2020, 4, 10, 16, 41, 25)