I'm having trouble understanding unit testing in Python. I have an object, retailer
, which creates another object, deal
. deal
refers to an attribute created in retailer
, so I'm passing it a reference:
class deal():
def __init__(self, deal_container, parent):
The deal_container
attribute also comes from retailer
, which calls its own methods to create it. So how do I create everything I need to easily make a deal
object?
Do I have to create an instance of retailer
in my unit test and then call the method in that object that creates deal
?
Can I use FactoryBoy to create an instance of retailer
and how do I include the method that creates deal
in that object?
What's the best way to approach this?
Here's the unit test. I'm setting up the soup_obj
I need to give deal:
class TestExtractString(TestCase):
fixtures = ['deals_test_data.json']
def setUp(self):
with open('/home/danny/PycharmProjects/askarby/deals/tests/BestBuyTest.html', 'r') as myfile:
text = myfile.read().replace('\n', '')
self.soup_obj = bs4.BeautifulSoup(text,"html.parser")
self.deal = self.soup_obj.find_all('div',attrs={'class':'list-item'})[0]
def test_extracts_title(self):
z = Retailer.objects.get(pk=1)
s = dealscan.retailer(z)
d = dealscan.deal(self.deal,s)
result = d.extract_string(self.deal,'title')
and here's the relevant bit of the deal
class in dealscan
. There's a retailer
class that creates a deal
, but I haven't even written the bit in retailer
that creates deal
yet. I'm hoping I can mock the bits I need for deal
without having to invoke retailer
at all, but then how do I deal with the fact that deal
references retailer
?
class deal():
def __init__(self, deal_container, parent):
'''
Initializes deal object
Precondition: 0 > price
Precondition: 0 > old_price
Precondition: len(currency) = 3
:param deal_container: obj
'''
self.css = self.parent.css
self.deal_container = deal_container
self.parent = parent
self.title = self.extract_string('title')
self.currency = self.parent.currency
self.price = self.extract_price('price')
self.old_price = self.extract_price('old_price')
self.brand = self.extract_string('brand')
self.image = self.extract_image('image')
self.description = self.extract_string('description')
#define amazon category as clearance_url
#define all marketplace deals
def __str__(self):
return self.title
def extract_string(self, element, deal):
'''
:param object deal: deal object to extract title from
:param string element: element to look for in CSS
:return string result: result of string extract from CSS
'''
tag = self.css[element]['tag']
attr = self.css[element]['attr']
name = self.css[element]['name']
result = deal.find(tag, attrs={attr: name})
if result:
if element == 'title':
return result.text
elif element == 'price':
result = self.extract_price(result).text
if result:
return result
elif element == 'image':
result = self.extract_image(result)
return False
The problem is that the deal object is referencing the parent before it sets the self.parent attribute. Use:
self.parent = parent
self.css = self.parent.css
self.deal_container = deal_container
and the AttributeError
goes away.
As for the question about whether it's good form to use an object to create another object in a unit test, the answer is that you can use mocks, but it's fine to do it this way. Using a helper method to set up the parent object once in setUp is acceptable and will make the code easier to read, and may improve test performance a little.