113 | Latest posts 114 | 115 | ... 116 | 117 | 118 | **NB:** when extracting contents from a response we use `css` or `xpath` Selectors to select elements from the downloaded 119 | response. 120 | 121 | I find `xpath` more powerful rather than using `css`, you will also have to read more on yourself to understand `Xpath` 122 | expressions. 123 | 124 | now to the final part of our spider, comments on the code will help you get the idea of what is happening 125 | 126 | 127 | import scrapy 128 | 129 | 130 | class NaiblogSpider(scrapy.Spider): 131 | name = 'naiblog' 132 | 133 | start_urls = [ 134 | 'http://blog.pynbo.or.ke/' 135 | ] 136 | 137 | def parse(self, response): 138 | # lets start by extracting the root node /html 139 | # from the response. 140 | html = response.xpath("/html") 141 | 142 | # select the body element which is inside 143 | # the root node html 144 | body = html.xpath("//body") 145 | 146 | # lets now get the content element 147 | content = body.xpath("//div[@class='content']") 148 | 149 | # extract the posts element 150 | posts = content.xpath("//div[@class='posts']") 151 | 152 | # iterate over the posts element in order to get 153 | # each individual post by extracting section node 154 | 155 | for post in posts.xpath("section[@class='post']"): 156 | 157 | # get the node header 158 | header = post.xpath("header[@class='post-header']") 159 | # finally get the title text 160 | title = header.xpath("h3 /a/text()").extract_first() 161 | 162 | # to get the description is the hardest part, if you inspect the element 163 | # of the page you will notice that there are two

and there 164 | # is a

element without any attribute between them which holds a simple description 165 | # of the post. Basically it makes it hard to extract the description of the post. 166 | # I suggest the developers of the web app should look into it. 167 | # with that issue in mind I did some tweak below in order to get the empty

168 | # tag. 169 | description = header.xpath("p[@class='post-meta']|p/text()").extract()[1] 170 | 171 | # extract post category 172 | category = header.xpath("p[@class='post-meta'] /a/text()").extract_first() 173 | 174 | # extract post date 175 | date = header.xpath("p[@class='post-meta'] /text()")[-1].extract() 176 | 177 | # finally lets return some data 178 | yield { 179 | 'description': description, 'category': category, 'title': title, 'date': date 180 | } 181 | 182 | lets execute our spider 183 | 184 | :::bash 185 | $ scrapy crawl naiblog -o posts.json 186 | 187 | 188 | you will see lots of logs being displayed on your terminal, when there are no more logs displayed navigate to the root of 189 | `naiblog` project and you will see a file named `posts.json`, open it and you will see all the posts in `pynbo blog`. 190 | 191 | Thats all!! 192 | 193 | find the entire project on my git repo [naiblog](https://github.com/gr1d99/naiblog-spider.git). 194 | -------------------------------------------------------------------------------- /content/tracking-vehicles-with-accesskenya-camera.md: -------------------------------------------------------------------------------- 1 | Title: Tracking Vehicles on Access Kenya Cameras 2 | Date: 2016-01-09 18:11:00 3 | Tags: python, image analysis, opencv, access kenya, traffic 4 | Slug: image_analysis_by_chris_orwa 5 | Author: Chris Orwa 6 | Summary: Description of a python API I built to track vehicular movement on Access Kenya cameras 7 | email: chrisorwa@gmail.com 8 | about_author:

Chris Orwa is a Data Scientist currently residing in Nairobi, Kenya. He's main focus is on computational techniques for unstructured data. To that end, he has learned to code in Python and R, and retained both languages as the core-analysis software/language with an occasional dash of C and Java.

His crazy analysis monologues can be found at http://blackorwa.com

9 | 10 | Hello everyone, here's my blog post on how I built an API to return traffic conditions from Access Kenya cameras. The API would work by specifiying a road name at the URL endpoint and a json reponse would have number of cars that moved and the speed of movement. 11 | 12 | #### Data Capture 13 | To begin the process I decided to test if is possible to capture Image from each camera. After exploration of the website I realized each camera had a url with a JPEG file extension at the end. I figured the cameras wrote a new image to the JPEG file on the URL. So, can I captured every new image? Yes - I used the **urllib** library to capture the image and store it on the disk. 14 | 15 | I organized the camera urls into a dict so as to have a means of calling and referencing each road. 16 | 17 | :::python 18 | cameras = dict ( 19 | museum='http://traffic.accesskenya.com/images/traffic/feeds/purshotam.jpg', 20 | ojijo='http://traffic.accesskenya.com/images/traffic/feeds/mhcojijo.jpg', 21 | forest_limuru='http://traffic.accesskenya.com/images/traffic/feeds/forestlimuru.jpg?', 22 | kenyatta_uhuru_valley='http://traffic.accesskenya.com/images/traffic/feeds/barclaysplaza.jpg',) 23 | 24 | After that, I wrote a function that takes a url and extracts three images every 6 seconds. 25 | 26 | :::python 27 | def capture_images(self): 28 | for i in 'abc': 29 | if self.name in way: 30 | urllib.urlretrieve(way[self.name],'img_'+i+'.jpg') 31 | time.sleep(6) 32 | 33 | #### Image Processing 34 | The images captured from the camera are stored on a folder with the name of the road. Next task involves processing the images for analysis. I utlize two libraries for this task; **PIL** for loading the images and **numpy** to convert pixel values to numerical array values. All these are wrapped in a function (shown below) that takes the image folder directory as input then proceeds to load all 3 images, converts them to numpy arrays, deletes the images and returns a python dict holding arrays on all the images. 35 | 36 | :::python 37 | # load images 38 | def load(self): 39 | files = os.listdir(self.path) 40 | a = dict() 41 | b = dict() 42 | k = 0 43 | 44 | while k <= len(files): 45 | for names in files: 46 | if names != '.DS_Store': 47 | a[names] = Image.open(names).convert('L') 48 | a[names].load() 49 | b[names] = np.asarray(a[names]) 50 | k +=1 51 | 52 | # delete image folder 53 | shutil.rmtree(os.getcwd()) 54 | 55 | return b 56 | 57 | #### Motion Detection 58 | The first important step in analysis of the images is checking if there has been movements within the 6 second period. To achieve this, I utilized the concept of differential imaging - a means of measuring motion detection by subtracting the pixel values of subsquent images. In my function, I calculate the number of pixels that have moved, this helps in quantifying the movement (standstill, moderate traffic). 59 | 60 | :::python 61 | # differential imaging 62 | def diffImg(self,img1,img2,img3): 63 | 64 | # calculate absolute difference 65 | d1 = cv2.absdiff(img1,img2) 66 | d2 = cv2.absdiff(img2,img3) 67 | bit = cv2.bitwise_and(d1,d2) 68 | ret,thresh = cv2.threshold(bit,35,255,cv2.THRESH_BINARY) 69 | 70 | #get number of different pixels 71 | moving = list() 72 | for cell in thresh.flat: 73 | if cell == 255: 74 | move = 'True' 75 | moving.append(move) 76 | pixie = len(moving) 77 | 78 | return pixie 79 | 80 | #### Calibrating Movement 81 | Once movement is detected, it is important to then quantify trafficc in km/h. To aid in this calculation is the optical flow algorithm. A concept in computer vision that allow tracking features in an image. I utilized this functionality to find features to track (cars) in the first images, and get their corresponding positions in the second and third image. I then proceeded to calculate the avaerage distance (euclidean distance) that the feature has moved. Dividing the pixel distance by 12 seconds gives me speed at which the objects(cars) are moving. My function returns this value. 82 | 83 | :::python 84 | # calculate optical flow of points on images 85 | def opticalFlow(self,img1,img2,img3): 86 | 87 | #set variables 88 | lk_params = dict(winSize = (10,10), 89 | maxLevel = 5, 90 | criteria = (cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT,10,0.03)) 91 | 92 | features_param = dict( maxCorners = 3000, 93 | qualityLevel = 0.5, 94 | minDistance = 3, 95 | blockSize = 3) 96 | 97 | # feature extraction of points to track 98 | pt = cv2.goodFeaturesToTrack(img1,**features_param) 99 | p0 =np.float32(pt).reshape(-1,1,2) 100 | 101 | # calaculate average movement 102 | dist = list() 103 | for loop in p0: 104 | p1,st,err =cv2.calcOpticalFlowPyrLK(img1, img2,loop, 105 | None,**lk_params) 106 | 107 | p0r,st,err =cv2.calcOpticalFlowPyrLK(img2,img1,p1, 108 | None,**lk_params) 109 | 110 | if abs(loop-p0r).reshape(-1, 2).max(-1) < 1: 111 | dst = distance.euclidean(loop,p0r) 112 | dist.append(dst) 113 | 114 | return round(max(dist)*10,2) 115 | 116 | #### The API 117 | The API (underconstruction) is based on flask. By specifying a road name at the HTTP endpoint, the API returns speed of traffic and level of movement (stanstill, moderate rate, no traffic). 118 | 119 | :::python 120 | # load required libraries 121 | import image_processing 122 | import numpy as np 123 | from flask import Flask 124 | import links 125 | import json 126 | import scipy as sp 127 | 128 | # create flask web server 129 | app = Flask(__name__) 130 | 131 | # create HTTP endpoint 132 | @app.route('/ImPro/') 133 | 134 | # main function 135 | def get_route(road): 136 | # initialize route class 137 | road = 'sarit' 138 | traffic = image_processing.route(road) 139 | 140 | # setup working directory 141 | traffic.set_dir() 142 | 143 | # get image from traffic camera 144 | traffic.capture_images() 145 | 146 | # load image stack 147 | x = traffic.load() 148 | 149 | # differential imaging 150 | y = traffic.diffImg(x['img_a.jpg'],x['img_b.jpg'],x['img_c.jpg']) 151 | 152 | # calculate optical flow 153 | z = traffic.opticalFlow(x['img_a.jpg'],x['img_b.jpg'],x['img_c.jpg']) 154 | 155 | if __name__ == "__main__": 156 | app.run() 157 | 158 | The code is available at . Feel free to modify and improve as you see fit. 159 | I hope you enjoyed the article. Feel free to drop me an e-mail at chrisorwa@gmail.com for any queries. 160 | -------------------------------------------------------------------------------- /content/decorators-in-python-mbithe-nzomo.md: -------------------------------------------------------------------------------- 1 | --- 2 | Title: Lessons in DRYness - Decorators in Python 3 | Date: 2017-08-30 09:00 4 | Tags: decorators, dry code, python 5 | Category: misc 6 | Slug: decorators_in_python 7 | Author: Mbithe Nzomo 8 | Summary: What decorators are and how to use them for DRYer code 9 | about_author: Mbithe is a software developer at Andela, and a co-organiser and coach at Django Girls Nairobi. She blogs at deCODEgirl.com and tweets at @mbithenzomo. 10 | email: mbithe.nzomo@gmail.com 11 | --- 12 | *This article first appeared 13 | on deCODEgirl.com.* 14 | 15 | Troll Face

16 | 17 | What do clothes and code (and my jokes) have in common? Answer: They're so 18 | much better when they're dry. Troll Face

19 | 20 | What does it mean for code to be **DRY**? Simple: **Don't Repeat Yourself**. 21 | The DRY Principle is key when writing efficient code. It helps keep your 22 | codebase smaller and less complex, saves you time, reduces redundancy... it's 23 | basically sugar, spice, and everything nice. :) In this post, I'm going to demonstrate how decorators in Python can help keep 24 | your code DRY. 25 | 26 | Once upon a time, I was building a RESTful API with Flask (which 28 | you can check it out on GitHub 30 | here) for an online bucket list service. Users could: 31 | 32 | 1. Sign up and login 33 | 2. Create, edit, view, and delete bucket lists 34 | 3. Create, edit, view, and delete bucket list items 35 | 36 | Some endpoints were protected and could only be accessed by authorized users. 37 | For instance, only the user who created a bucket list or bucket list item could 38 | edit, view, or delete it. 39 | 40 | Here's how I initially tackled it. First, I created a method to display an 41 | error message, with the default being "Error: You are not authorized to access 42 | this resource." 43 | 44 | ``` python 45 | def unauthorized(message=None): 46 | """ 47 | Returns an error message. 48 | """ 49 | if not message: 50 | message = "Error: You are not authorized to access this resource." 51 | return jsonify({ 52 | "message": message 53 | }), 403 54 | ``` 55 | 56 | Then, I went ahead and wrote the methods for the endpoints: 57 | 58 | ``` python 59 | class BucketListAPI(Resource): 60 | """ 61 | URL: /api/v1/bucketlists/ 62 | Request methods: GET, PUT, DELETE 63 | """ 64 | 65 | def get(self, id): 66 | """ Get a bucket list """ 67 | bucketlist = Bucketlist.query.filter_by(id=id).first() 68 | if bucketlist: 69 | if bucketlist.created_by == g.user.id: 70 | return marshal(bucketlist, bucketlist_serializer) 71 | else: 72 | return unauthorized() 73 | else: 74 | return unauthorized("Error: The bucket list specified doesn't " 75 | "exist. Please try again!") 76 | 77 | def put(self, id): 78 | """ Edit a bucket list """ 79 | bucketlist = Bucketlist.query.filter_by(id=id).first() 80 | if bucketlist: 81 | if bucketlist.created_by == g.user.id: 82 | parser = reqparse.RequestParser() 83 | parser.add_argument("title", 84 | required=True, 85 | help="No title provided.") 86 | parser.add_argument("description", type=str, default="") 87 | args = parser.parse_args() 88 | title, description = args["title"], args["description"] 89 | bucketlist.title = title 90 | bucketlist.description = description 91 | return edit_item(name="title", 92 | item=bucketlist, 93 | serializer=bucketlist_serializer, 94 | is_user=False, 95 | is_bucketlist=True, 96 | is_item=False) 97 | else: 98 | return unauthorized() 99 | else: 100 | return unauthorized("Error: The bucket list you are trying to " 101 | "edit doesn't exist. Please try again!") 102 | 103 | def delete(self, id): 104 | """ Delete a bucket list """ 105 | bucketlist = Bucketlist.query.filter_by(id=id).first() 106 | if bucketlist: 107 | if bucketlist.created_by == g.user.id: 108 | return delete_item(bucketlist, 109 | bucketlist.title, 110 | is_user=False, 111 | is_bucketlist=True, 112 | is_item=False) 113 | else: 114 | return unauthorized() 115 | else: 116 | return unauthorized("Error: The bucket list you are trying to " 117 | "delete doesn't exist. Please try again!") 118 | ``` 119 | 120 | You can see the problem already: in each of these methods, I basically repeated 121 | the same chunk of code to ensure that: 122 | 123 | 1. The bucket list existed 124 | 2. The current user was authorized to access the endpoint 125 | 3. Appropriate error messages were displayed if the above conditions were false 126 | 127 | ``` python 128 | # Check whether bucket list exists 129 | if bucketlist: 130 | # Check whether bucket list was created by the current user 131 | if bucketlist.created_by == g.user.id: 132 | # Do something here 133 | else: 134 | # Return unauthorized error message 135 | return unauthorized() 136 | else: 137 | # Return non-existent bucket list error message 138 | return unauthorized("Error: The bucket list you are trying to " 139 | "delete doesn't exist. Please try again!") 140 | ``` 141 | 142 | Not DRY at all. Fret not, though: decorators to the rescue! 143 | 144 | A decorator is basically a function that takes a callable (a function or a 145 | class) as a parameter, modifies it, and returns it. You could also say that decorators wrap a callable, modifying its behavior. 146 | 147 | Here's how I used a decorator to eliminate my duplicate code. First, I wrote 148 | the decorator function: 149 | 150 | ``` python 151 | def authorized_user_bucketlist(function): 152 | def auth_wrapper(*args, **kwargs): 153 | g.bucketlist = Bucketlist.query.filter_by(id=kwargs["id"]).first() 154 | try: 155 | if g.bucketlist.created_by == g.user.id: 156 | return function(*args, **kwargs) 157 | return unauthorized() 158 | except: 159 | return unauthorized("Error: The bucket list specified doesn't " 160 | "exist. Please try again!") 161 | return auth_wrapper 162 | ``` 163 | 164 | So here's what's going on in there: The decorator is 165 | `authorized_user_bucketlist`, and it takes some function as a parameter. It has 166 | another function, `auth_wrapper`, where a variable `g.bucketlist` is defined as 167 | the bucket list whose ID is passed to the function as a key word argument. It 168 | then checks if the bucket list was created by the current user. If so, it 169 | invokes the function and returns the result. If not, it returns an unauthorized error message. If the bucket list doesn't exist, it returns a non-existent 170 | bucket list error message. Basically, it does the same thing my duplicated 171 | chunk of code was doing, but with a little more finesse (like, using a 172 | try-except block as opposed to an if-else block, because it's better to ask for forgiveness than to ask for permission). 173 | 174 | The refactored code: 175 | 176 | ``` python 177 | from base import authorized_user_bucketlist 178 | 179 | class BucketListAPI(Resource): 180 | """ 181 | URL: /api/v1/bucketlists/ 182 | Request methods: GET, PUT, DELETE 183 | """ 184 | @authorized_user_bucketlist 185 | def get(self, id): 186 | """ Get a bucket list """ 187 | return marshal(g.bucketlist, bucketlist_serializer) 188 | 189 | @authorized_user_bucketlist 190 | def put(self, id): 191 | """ Edit a bucket list """ 192 | parser = reqparse.RequestParser() 193 | parser.add_argument("title", 194 | required=True, 195 | help="No title provided.") 196 | parser.add_argument("description", type=str, default="") 197 | args = parser.parse_args() 198 | title, description = args["title"], args["description"] 199 | g.bucketlist.title = title 200 | g.bucketlist.description = description 201 | return edit_item(name="title", 202 | item=g.bucketlist, 203 | serializer=bucketlist_serializer, 204 | is_user=False, 205 | is_bucketlist=True, 206 | is_item=False) 207 | 208 | @authorized_user_bucketlist 209 | def delete(self, id): 210 | """ Delete a bucket list """ 211 | return delete_item(g.bucketlist, 212 | g.bucketlist.title, 213 | is_user=False, 214 | is_bucketlist=True, 215 | is_item=False) 216 | ``` 217 | 218 | I imported the decorator function, and then called the decorator using the `@` 219 | symbol before each method. This is some 221 | syntactic sugar that Python provides, 222 | making the calling of decorators much simpler than it otherwise would be. 223 | 224 | The code is much DRYer now! Mission accomplished! It's not perfect, but it's 225 | way better than it was. :) This project is several months old now, and no longer 226 | being maintained, but feel free to clone the repo and test my API! 229 | 230 | Also, you can learn more about decorators in Python here and here. 234 | -------------------------------------------------------------------------------- /content/parametrizedUnitTestingWithPyTest.md: -------------------------------------------------------------------------------- 1 | --- 2 | Title: Python parametrized unit testing with Pytest 3 | Date: 2017-09-04 19:00 4 | Tags: python, pytest 5 | Category: unit testing 6 | Slug: test_parameterization 7 | Author: Paul Upendo 8 | Summary: Create pythonic parametrized tests with pytest 9 | about_author: Upendo is a Software Developer at Andela Kenya 10 | email: lovepaul46@outlook.com 11 | --- 12 | 13 | ## Pre-requisites 14 | > - Beginner level Python 15 | > - Python2* or Python3* installed 16 | > - Basic knowledge on TDD and python UnitTesting 17 | > - Pytest installed 18 | 19 | # Intro 20 | I really did not get a hang of TDD the first time I heard about it. I saw it as an acronym that was maybe too difficult to fathom. A few months down the line I have a clear and better perspective towards it, and even better a more pythonic way to use TDD in my software development. Hang on as we methodically dive into it. 21 | 22 | TDD (Test Driven Development) is a software design methodology. This generally means writing UnitTests for your code even before you write it yourself. TDD, now that you are farmiliar with the acronym, is not limited to a particular language stack but for this instance we will stick to python as our language of choice. 23 | 24 | # Parametrizing your tests with Pytest 25 | There are a couple of UnitTesting frameworks in python: 26 | - Python Unittest (that comes with the standard python library distribution) 27 | - Python Pytest 28 | - Python-nose 29 | 30 | Python UnitTest framework is a member of the XUnit family of testing tools that use a similar pattern for defining testCases. With UnitTest following XUnit Conventions, it makes Pytest and Python-nose a more - Pythonic way to write tests that follow Pep8 style guide. I will focus on paramet`rization of tests using Pytest. 31 | 32 | #### Why parametrize my tests? 33 | Let's think of a scenario where we have to write tests for a particular function but with multiple input values/arguments. Rewriting my assertion statements with the different arguments works quite well but what if we could simplify this to only one assertion statement for that function? Then run the test multiple times with an array of parameters? Smart right? Python's Pytest provides an excellent Pythonic way to achieve this. 34 | 35 | `@pytest.mark.parametrize` allows one to define multiple sets of arguments and fixtures at the test function or class. 36 | 37 | Let us write parametrized tests for a function that accepts integer input from the user and determines whether the input is odd or even. 38 | ```python 39 | # File test_parametrization.py 40 | import pytest 41 | 42 | 43 | @pytest.mark.parametrize("test_input,expected", [ 44 | (3, "odd"), 45 | (4, "even"), 46 | (9, "even"), 47 | ]) 48 | 49 | def test_even_and_odd_numbers(test_input, expected): 50 | assert (runOddEvenCheck(test_input)) == expected 51 | 52 | ``` 53 | To run this, use the command `pytest`, which executes the pytest test runner for us. The result for this would be a series of failing tests as we can see below. The beautiful thing about this is that by implementing test parameterization, Pytest created 3 individual tests for us from our test function :smile: . 54 | ```python 55 | # Commandline output 56 | (pytest_blog) ~/D/t/p/parameterized_testing ❯❯❯ pytest ⏎ develop ✱ ◼ 57 | =================================================================================================================================== test session starts =================================================================================================================================== 58 | platform darwin -- Python 3.6.1, pytest-3.2.1, py-1.4.34, pluggy-0.4.0 59 | rootdir: /Users/Upendo/Documents/tech_blogs/pytest_blog/parameterized_testing, inifile: 60 | collected 3 items 61 | 62 | test_parametrization.py FFF 63 | 64 | ======================================================================================================================================== FAILURES ========================================================================================================================================= 65 | ____________________________________________________________________________________________________________________________ test_even_and_odd_numbers[3-odd] _____________________________________________________________________________________________________________________________ 66 | 67 | test_input = 3, expected = 'odd' 68 | 69 | @pytest.mark.parametrize("test_input,expected", [ 70 | (3, "odd"), 71 | (4, "even"), 72 | (9, "even"), 73 | ]) 74 | 75 | def test_even_and_odd_numbers(test_input, expected): 76 | > assert (runOddEvenCheck(test_input)) == expected 77 | E NameError: name 'runOddEvenCheck' is not defined 78 | 79 | test_parametrization.py:11: NameError 80 | ____________________________________________________________________________________________________________________________ test_even_and_odd_numbers[4-even] ____________________________________________________________________________________________________________________________ 81 | 82 | test_input = 4, expected = 'even' 83 | 84 | @pytest.mark.parametrize("test_input,expected", [ 85 | (3, "odd"), 86 | (4, "even"), 87 | (9, "even"), 88 | ]) 89 | 90 | def test_even_and_odd_numbers(test_input, expected): 91 | > assert (runOddEvenCheck(test_input)) == expected 92 | E NameError: name 'runOddEvenCheck' is not defined 93 | 94 | test_parametrization.py:11: NameError 95 | ____________________________________________________________________________________________________________________________ test_even_and_odd_numbers[9-even] ____________________________________________________________________________________________________________________________ 96 | 97 | test_input = 9, expected = 'even' 98 | 99 | @pytest.mark.parametrize("test_input,expected", [ 100 | (3, "odd"), 101 | (4, "even"), 102 | (9, "even"), 103 | ]) 104 | 105 | def test_even_and_odd_numbers(test_input, expected): 106 | > assert (runOddEvenCheck(test_input)) == expected 107 | E NameError: name 'runOddEvenCheck' is not defined 108 | 109 | test_parametrization.py:11: NameError 110 | ================================================================================================================================ 3 failed in 0.07 seconds ================================================================================================================================= 111 | (pytest_blog) ~/D/t/p/parameterized_testing ❯❯❯ 112 | ``` 113 | This just made the process to write our tests less time consuming, not that this will always be the case as it also depends on the complexity or requirements of the project/code we plan to write after the tests for. 114 | Let us get back to fixing our failing tests as per the output we have received form Pytest test runner :bowtie: 115 | All our asserts fail as a NameError exception is raised. So let's fix that by creating a function `runOddEvenCheck` to check for number state whether odd or even when user input is provided. Then import the function for our tests to use. 116 | ``` python 117 | # File numberWork.py 118 | def runOddEvenCheck(nums): 119 | return "even" if nums % 2 == 0 else "odd" 120 | 121 | ``` 122 | Then let us import this function to our test file: 123 | ``` python 124 | import pytest 125 | from numberWork import runOddEvenCheck 126 | 127 | 128 | @pytest.mark.parametrize("test_input,expected", [ 129 | (3, "odd"), 130 | (4, "even"), 131 | (9, "even"), 132 | ]) 133 | def test_even_and_odd_numbers(test_input, expected): 134 | assert (runOddEvenCheck(test_input)) == expected 135 | 136 | ``` 137 | And we can finally run our tests with `pytest`: 138 | ``` python 139 | (pytest_blog) ~/D/t/p/parameterized_testing ❯❯❯ pytest ⏎ develop ✱ ◼ 140 | =================================================================================================================================== test session starts =================================================================================================================================== 141 | platform darwin -- Python 3.6.1, pytest-3.2.1, py-1.4.34, pluggy-0.4.0 142 | rootdir: /Users/Upendo/Documents/tech_blogs/pytest_blog/parameterized_testing, inifile: 143 | collected 3 items 144 | 145 | test_parametrization.py ..F 146 | 147 | ======================================================================================================================================== FAILURES ========================================================================================================================================= 148 | ____________________________________________________________________________________________________________________________ test_even_and_odd_numbers[9-even] ____________________________________________________________________________________________________________________________ 149 | 150 | test_input = 9, expected = 'even' 151 | 152 | @pytest.mark.parametrize("test_input,expected", [ 153 | (3, "odd"), 154 | (4, "even"), 155 | (9, "even"), 156 | ]) 157 | def test_even_and_odd_numbers(test_input, expected): 158 | > assert (runOddEvenCheck(test_input)) == expected 159 | E AssertionError: assert 'odd' == 'even' 160 | E - odd 161 | E + even 162 | 163 | test_parametrization.py:11: AssertionError 164 | =========================================================================================================================== 1 failed, 2 passed in 0.06 seconds ============================================================================================================================ 165 | (pytest_blog) ~/D/t/p/parameterized_testing ❯❯❯ 166 | ``` 167 | Wow, we do have passing and failing tests now from our Pytest report. Two tests pass, with one failing test (Fact being that 9 is not an even number from the AssertionError message above). 168 | From these example we are now able to give parametrized testing a try I hope :smile:. The topic is quite vast, this is just but a sneak peak at the good we can do with the Pythonic Pytest. 169 | If you would like to try this example out, [here is a link to the source code on github](https://github.com/paulupendo/parametrized_testing/tree/master). `Fork`, `Clone`, `install requirements` and `pytest` :smile:. 170 | 171 | Happy coding! 172 | 173 | 174 | 175 | 176 | 177 | 178 | --------------------------------------------------------------------------------