Category "Python"

Python, Boto and AWS EC2

- - Python, Tutorials

Most if not all software companies have adopted to cloud infrastructure and services. AWS in particular is very popular amongst all. The intentions of this post is to host a few examples on using boto to make use of one of the services available on AWS i.e EC2. It is more likely than not to have need of a mechanism to programatically fire up a few instances, shut them down, filter instances and send remote commands to it to say the least.

Filter instances based on tag names from the AWS inventory

EC2 instances on AWS can have as many tag names key: value as required for purposes like identifying an instance or a set of instances. Also when the instance you are working on quite frequently needs to shut down and boot over again and you haven’t implemented elastic IP, you are bound to changes in the public IP address. Although you could argue to use private IP to filter an instance, it isn’t very effective when you have a lot of instances(>100).

Boto2
import boto.ec2

conn = boto.ec2.connect_to_region('us-east-1', aws_access_key_id='aws_access_id', aws_secret_access_key='aws_secret')
reservations = conn.get_all_instances(filters={'tagName' : 'value'})
public_ips = [each_instance.ip_address for r in reservations for each_instance in r.instances]
# each_instance.private_ip_address  to get the private ip address of the instance
Boto3
import boto3
session = boto3.session.Session(aws_access_key_id=aws_access_id,
                                aws_secret_access_key=aws_secret,
                                region_name='us-east-1')
 
ec2 = session.resource('ec2')
instances = ec2.instances.filter(
    Filters=[{'Name':'tag:purpose', 'Values':['intelligence']}
])
public_ips = [each_instance.public_ip_address for each_instance in instances]
# each_instance.private_ip_address to get the private ip address of the instance
Boot/Shutdown an instance/instances from the AWS inventory

Using boto, you can boot/shutdown/terminate instances.

Boto2
def start_stop_terminate_instance(instance_ids, conn, action='start'):
    if action == 'start':
        conn.start_instances(instance_ids=instance_ids)
    elif action == 'stop':
        conn.stop_instances(instance_ids=instance_ids)
    elif action == 'terminate':
        conn.terminate_instances(instance_ids=ids)
Boto3
def start_stop_terminate_instance(instance_ids, conn, action='start'):
    if action == 'start':
        conn.instances.filter(InstanceIds=instance_ids).start()
    elif action == 'stop':
        conn.instances.filter(InstanceIds=instance_ids).stop()
    elif action == 'terminate':
        conn.instances.filter(InstanceIds=instance_ids).terminate()
Create Instances based on various metrics

Boto makes use of the AWS APIs that also allows creating instances. An EC2 instance can have various properties. The most common is the type of the instance. Types are generally a grouping of instances based on metrics such as power, performance, bandwidth. Commonly used types for general purpose are t2, m4, m3. C5, c4, c3 are compute optimized instances. For a process/application more leaned towards in-memory activities, you’d use x1, r4, r3. There are other types too but the above mentioned are quite common in use. The other properties of an instance are instance id, the memory size (micro, nano, small, large, xlarge, 2xlarge, 4xlarge, 8xlarge, 10xlarge.), the key pair to make a secured connection to the instance, tag names, display names, security groups, attached storage id, etc. Using boto we can create an instance or multiple instances based on the above mentioned parameters.

Boto2
import boto.ec2
conn = boto.ec2.connect_to_region('us-east-1', aws_access_key_id='aws_access_id', aws_secret_access_key='aws_secret')
conn.run_instances(
    'ami-ag139jf',
    min_count=10, 
    max_count=100,
    key_name='myKey',
    instance_type='t2.small',
    security_groups=['sg-4512']
)
Boto3
import boto3
session = boto3.session.Session(aws_access_key_id='aws_access_id',
                                aws_secret_access_key='aws_secret',
                                region_name='us-east-1')
 
ec2 = session.resource('ec2')
ec2.create_instances(
    ImageId='ami-ag139jf', 
    MinCount=10, 
    MaxCount=100, 
    InstanceType='t2.small',
    KeyName='myKey',
    SecurityGroups=['sg-4512']
)
Send remote commands to an EC2 instance

Paramiko can be used for connecting to a remote instance and sending commands to be executed and get the standard output/error to act accordingly.

import paramiko

key = paramiko.RSAKey.from_private_key_file(path_to_pem_file)
client = paramiko.SSHClient()
client.set_missing_host_key_policy(paramiko.AutoAddPolicy())

# Connect to the instance
try:
    # using username, public ip address and the pem file, create connection to the instance
    client.connect(hostname=instance_ip, username="ubuntu", pkey=key)

    # Execute command remotely.
    stdin, stdout, stderr = client.exec_command(“ls -l”)
    print stdout.read()
    client.close()

except Exception, e:
    print e

Google APIs and Python – Part II

- - Python, Tutorials

Google services are cool and you can build products and services around it. We will see through examples how you can use various google services such as spreadsheet, slides and drive through Python. I hope people can take ideas from the following example to do amazing stuffs with Google services. There is a part one to this article where I walked through procedure to enable Google APIs, installation of required packages in Python, authentication and demonstrated individual examples of Sheets, Drive and Slides API. https://www.thetaranights.com/brief-introduction-to-google-apissheets-slides-drive/ . In this article however, we will integrate Sheets, Drive and Slides API altogether.

The Idea

We will use data from a sheet which contains some statistics about a few applications/websites. The end goal is to create a presentation slide, add a background image to it, add content from the sheet to the slide and also some other cool stuffs. All the resources used in the following examples are public so you can follow along.

I will be using following resources throughout the example

Create Presentation Slides from Sheets data and Drive images using Python
from googleapiclient import discovery
from httplib2 import Http
from oauth2client import file, client, tools

TEMPLATE_FILE = "TEM_F"

SCOPES = ('https://www.googleapis.com/auth/spreadsheets','https://www.googleapis.com/auth/drive')

CLIENT_SECRET = 'client_secret_760822340075-i0ark1h51pnbhii5dgafug3k4g1nodb8.apps.googleusercontent.com.json' # download from google console after activating apis

store = file.Storage('storage.json') # doesn't matter if not present, you will be prompted to accept access to google resources on your account and a token will be generated that is stored inside storage.json with requested previliges.

credz = store.get()

if not credz or credz.invalid:
    flow = client.flow_from_clientsecrets(CLIENT_SECRET, SCOPES)
    credz = tools.run_flow(flow, store)

HTTP = credz.authorize(Http())

SHEETS = discovery.build('sheets', 'v4', http=HTTP)

SLIDES = discovery.build('slides', 'v1', http=HTTP)

DRIVE = discovery.build('drive', 'v3', http=HTTP)


presentation_template_file_id = "1wLimfuGw1pqZZJvkc15lOfD7LSEAWCjuIlhbrOaiulE" # the template has been made public.
# name of the presentation file
DATA = {'name':'MobileApplicationsReport'}

PRESENTATION_ID = DRIVE.files().copy(body=DATA, fileId="1wLimfuGw1pqZZJvkc15lOfD7LSEAWCjuIlhbrOaiulE").execute()['id']
print(PRESENTATION_ID)

sheet_ID = '1xpjQkF692lNnTsfOckVll2OPTa659ZCuK3JezDSkris' # the sheet where we fetch data from to populate to the slides.

application_statistics = SHEETS.spreadsheets().values().get(range='Sheet1', spreadsheetId=sheet_ID).execute().get('values') # all the data from the sheet as lists including headers.

print(application_statistics)

presentation_details = SLIDES.presentations().get(presentationId=PRESENTATION_ID).execute()

slides_data = presentation_details.get('slides', [])[0]

page_id = slides_data['objectId'] # page id of the first slide of the presentation.

for each_data in application_statistics[1:]: # skip the headers.
    # duplicate slide for the next cycle before replacing content on a slide since we are using method of replacing text from the slide to populate data.
    reqs = [{"duplicateObject": {"objectId": page_id}}]
    copy_slide_rsp = SLIDES.presentations().batchUpdate(body={'requests':reqs}, presentationId=PRESENTATION_ID).execute()
    
    IMG_ID = each_data[10] # the id of the image present on google drive which we intend to have as a background image to this particular slide.
    img_url = '%s&access_token=%s' % (DRIVE.files().get_media(fileId=IMG_ID).uri, credz.access_token)
    print("Image url", img_url)

    # prepare a bulk requests that basically replaces the text from the template with the actual data from the sheets.
    bulk_requests = [
        {'updatePageProperties':{'objectId':page_id, 'pageProperties':{'pageBackgroundFill':{'stretchedPictureFill':{'contentUrl':img_url}}}, 'fields':'pageBackgroundFill'}},
        {'replaceAllText':{'containsText':{'text':'{{SHOWCASE  NAME}}', 'matchCase':True}, 'replaceText':each_data[1], "pageObjectIds": [page_id]}},
        {'replaceAllText':{'containsText':{'text':'{{DESCRIPTION}}', 'matchCase':True}, 'replaceText':each_data[2], "pageObjectIds": [page_id]}},
        {'replaceAllText':{'containsText':{'text':'{{COMPOSITION}}', 'matchCase':True}, 'replaceText':each_data[3], "pageObjectIds": [page_id]}},
        {'replaceAllText':{'containsText':{'text':'{{IMPRESSIONS}}', 'matchCase':True}, 'replaceText':each_data[8], "pageObjectIds": [page_id]}},
        {'replaceAllText':{'containsText':{'text':'{{VIDEO VIEWS}}', 'matchCase':True}, 'replaceText':each_data[7], "pageObjectIds": [page_id]}},
        {'replaceAllText':{'containsText':{'text':'{{USERS}}', 'matchCase':True}, 'replaceText':each_data[6], "pageObjectIds": [page_id]}},
        {'replaceAllText':{'containsText':{'text':'{{MOBILE}}', 'matchCase':True}, 'replaceText':each_data[9], "pageObjectIds": [page_id]}}
    ]
    bulk_update_response = SLIDES.presentations().batchUpdate(body={'requests':bulk_requests}, presentationId=PRESENTATION_ID, fields='').execute().get('replies')

    page_id = copy_slide_rsp['replies'][0]['duplicateObject']['objectId'] # update the page id as the one that was duplicated so we now can work on this slide.

delete_final_page = SLIDES.presentations().batchUpdate(body={'requests':[{"deleteObject": {"objectId": page_id}}]}, presentationId=PRESENTATION_ID, fields='').execute().get('replies')
Output Presentation Created:

After successful running of the above program, following presentation was generated.
https://docs.google.com/presentation/d/1h9YqUnCWu5pxXmW3rs_9rKmMsVeJM9I8nIBGbr25pME/edit?usp=sharing

Brief Introduction to Google APIs(Sheets, Slides, Drive)

- - Python, Tutorials

The intentions of this post is to familiarize usage of Google APIs with Python. Google services are cool and you can build products and services around it. We will see through examples how you can use various google services such as spreadsheet, slides and drive through Python. I hope people can take ideas from the following example to do amazing stuffs with Google services. In order to work with google services via their APIs, first we need to create a project on google console with specific APIs enabled. For the scope of this article, we will need the SHEETS API, SLIDES API and DRIVE API enabled.

I will be using following resources throughout the examples

Installation of libraries and setup

pip install --upgrade google-api-python-client oauth2client

Creating a project on Google Console and enabling APIs

1. Open google console https://console.cloud.google.com/apis/dashboard
2. Create a new project

create_a_project_google_console

Create a new project on google console

3. Name the project

new_project_google_console

Name new project

4. Enable Sheets, Slides and Drive APIs

google_console_enable_apis

Enable Google APIs



google_console_enable_drive_api

Enable Drive API as well as slides and sheets APIs

5. Create Credentials and Download it.

google_console_create_credentials

Create credentials for the project and download it

Authentication

We need the credentials that was downloaded from google console for authentication. Google creates an access token to access and work on the google resources. The token does expire and in case it does, we will be prompted on a browser to provide access to the application for the specified resources on our google account.

>>> from googleapiclient import discovery
>>> from httplib2 import Http
>>> from oauth2client import file, client, tools
>>> SCOPES = ('https://www.googleapis.com/auth/spreadsheets','https://www.googleapis.com/auth/drive')
>>> CLIENT_SECRET = 'client_secret_760822340075-i0ark1h51pnbhii5dgafug3k4g1nodb8.apps.googleusercontent.com.json'
>>> store = file.Storage('token.json')
>>> creds = store.get()
/home/bhishan-1504/googleapis/googleapienv/lib/python3.6/site-packages/oauth2client/_helpers.py:255: UserWarning: Cannot access token.json: No such file or directory
  warnings.warn(_MISSING_FILE_MESSAGE.format(filename))
>>> 

>>> if not credz or credz.invalid:
...     flow = client.flow_from_clientsecrets(CLIENT_SECRET, SCOPES)
...     credz = tools.run_flow(flow, store)
...

Your browser has been opened to visit:

    https://accounts.google.com/o/oauth2/auth?client_id=760822340075-i0ark1h51pnbhii5dgafug3k4g1nodb8.apps.googleusercontent.com&redirect_uri=http%3A%2F%2Flocalhost%3A8080%2F&scope=https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fspreadsheets+https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive&access_type=offline&response_type=code

If your browser is on a different machine then exit and re-run this
application with the command-line parameter

  --noauth_local_webserver

Created new window in existing browser session.
Authentication successful.
>>>

Note: For all of the examples below, we need authentication. I am skipping them in all of the examples underneath to remove redundancy.

Reading a google spreadsheet using Python
>>> HTTP = credz.authorize(Http())
>>> SHEETS = discovery.build('sheets', 'v4', http=HTTP)
>>> sheet_ID = '1xpjQkF692lNnTsfOckVll2OPTa659ZCuK3JezDSkris'
>>> spreadsheet_read = SHEETS.spreadsheets().values().get(range='Sheet1', spreadsheetId=sheet_ID).execute()
>>> spreadsheet_read
{'range': 'Sheet1!A1:Z1001', 'majorDimension': 'ROWS', 'values': [['Category', 'Showcase Name', 'Description', 'Audience Composition', 'ID', 'Audience Name', 'Users', 'Video Views', 'Impressions', 'Mobile Impressions', 'Image'], ['Sports', 'Sports Fans', 'People who likes sports', 'online data', '12', 'Sports Fans', '1000', '1000', '1000', '1000', 'sports-fans.jpg']]}
>>> spreadsheet_values = spreadsheet_read['values']
>>> spreadsheet_values
[['Category', 'Showcase Name', 'Description', 'Audience Composition', 'ID', 'Audience Name', 'Users', 'Video Views', 'Impressions', 'Mobile Impressions', 'Image'], ['Sports', 'Sports Fans', 'People who likes sports', 'online data', '12', 'Sports Fans', '1000', '1000', '1000', '1000', 'sports-fans.jpg']]
>>>
Search and Download a file from drive using Python
>>> import io
>>> DRIVE = discovery.build('drive', 'v3', http=HTTP)
>>> resp = DRIVE.files().list(q="name='{filepath}'".format(filepath="P1150264.JPG")).execute()
>>> resp
{'kind': 'drive#fileList', 'incompleteSearch': False, 'files': [{'kind': 'drive#file', 'id': '0B54qUrMD2GDIa0FMdkxLMmpoZVU', 'name': 'P1150264.JPG', 'mimeType': 'image/jpeg'}]}
>>> file_id = resp['files'][0]['id']
>>> file_id
'0B54qUrMD2GDIa0FMdkxLMmpoZVU'
>>> file_request = DRIVE.files().get_media(fileId=file_id)
>>> fh = io.BytesIO()
>>> downloader = MediaIoBaseDownload(fh, file_request)
>>> done = False
>>> while done is False:
...     status, done = downloader.next_chunk()
...     print("Download {status}".format(status=status.progress() * 100))
...
Download 100.0
>>>

Google Slides API

Create a blank presentation using Python
>>> SLIDES = discovery.build('slides', 'v1', http=HTTP)
>>> body = {'title': 'AutomatedPresentation'}
>>> presentation_request = SLIDES.presentations().create(body=body).execute()
>>> presentation_request['presentationId']
'1ROJOeVFaA4PbC2voR5EFddohxQlZvUkrdi1dsJUks9c'
>>>

Follow the link to see the presentation the above code snippet creates.
https://docs.google.com/presentation/d/1ROJOeVFaA4PbC2voR5EFddohxQlZvUkrdi1dsJUks9c/edit?usp=sharing

Creating presentation using existing template from drive
>>>TEMPLATE_FILE = 'TEM_F'
>>> SLIDES = discovery.build('slides', 'v1', http=HTTP)
>>> DRIVE = discovery.build('drive', 'v3', http=HTTP)
>>> rsp = DRIVE.files().list(q="name='%s'"% TEMPLATE_FILE).execute()['files'][0]
>>> rsp
{'kind': 'drive#file', 'id': '1wLimfuGw1pqZZJvkc15lOfD7LSEAWCjuIlhbrOaiulE', 'name': TEMPLATE_FILE, 'mimeType': 'application/vnd.google-apps.presentation'}
>>> DATA = {'name': 'PresentationUsingTemplate'}
>>> create_presentation_request = DRIVE.files().copy(body=DATA, fileId=rsp['id']).execute()
>>> presentation_id = create_presentation_request['id']
>>> presentation_id
'10iDjayeyVkVSp5F6eQIqzpISAqjFlbqG4_jdYDAFJG4'
>>>

Follow the link to see the presentation the above code snippet creates. https://docs.google.com/presentation/d/10iDjayeyVkVSp5F6eQIqzpISAqjFlbqG4_jdYDAFJG4/edit?usp=sharing

Adding background image to a slide
>>> SLIDES = discovery.build('slides', 'v1', http=HTTP)
>>> DRIVE = discovery.build('drive', 'v3', http=HTTP)
>>> rsp = DRIVE.files().list(q="name='%s'"% TEMPLATE_FILE).execute()['files'][0]
>>> rsp
{'kind': 'drive#file', 'id': '1wLimfuGw1pqZZJvkc15lOfD7LSEAWCjuIlhbrOaiulE', 'name': 'TEM_F', 'mimeType': 'application/vnd.google-apps.presentation'}
>>> DATA = {'name': 'PresentationUsingTemplatePlusBackgroundImage'}
>>> create_presentation_request = DRIVE.files().copy(body=DATA, fileId=rsp['id']).execute()
>>> presentation_id = create_presentation_request['id']
>>> presentation_id
'1cxpaH19h582Q4Ot3b5GL9U6ETl9myqE3JlX4_Fa35e8'
>>> IMG_FILE = "sports-fans.jpg"
>>> img_file_request = DRIVE.files().list(q="name='%s'" % IMG_FILE).execute()['files'][0]
>>> img_url = '%s&access_token=%s' % (DRIVE.files().get_media(fileId=img_file_request['id']).uri, credz.access_token)
>>> img_url
'https://www.googleapis.com/drive/v3/files/0B54qUrMD2GDIa2syZWF3OE5xSUk?alt=media&access_token=ya29.Glz3BcRtfadsfGzKwUQ-6llroeaMfdasfdaffadsjfdXiOewDdHqhdgBef2euMm9OMxGXyXF-axwZ0gFBwH2-T6qS29qmpc-H3ELcyh7CDZCbfzn7DTNJkugoA'

>>> presentation_details = SLIDES.presentations().get(presentationId=presentation_id).execute()
>>> first_slide_data = presentation_details.get('slides', [])[0]
>>> first_slide_id = slides_data['objectId']
>>> first_slide_id
'p3'
>>> bulk_reqs = [{'updatePageProperties':{'objectId':first_slide_id, 'pageProperties':{'pageBackgroundFill':{'stretchedPictureFill':{'contentUrl':img_url}}}, 'fields':'pageBackgroundFill'}}]
>>> bulk_update_req = SLIDES.presentations().batchUpdate(body={'requests':bulk_reqs}, presentationId=presentation_id).execute()

Follow the link to see the presentation the above code snippet creates.
https://docs.google.com/presentation/d/1cxpaH19h582Q4Ot3b5GL9U6ETl9myqE3JlX4_Fa35e8/edit?usp=sharing

On a follow up post to this one, we will focus on integrating slides, sheets and drive API altogether. We shall use spreadsheet data and populate it onto presentation slides.
To be continued…

Update

Published the second part to this article. https://www.thetaranights.com/google-apis-and-python-part-ii/

Python filter() built-in

- - Python, Tutorials

Filter makes an iterator that takes a function and uses the arguments from the following iterable passed to the filter built-in. It returns a filtered iterator which contains only those values for which the function(passed as the first argument to the filter) evaluated truth value. What makes this possible is the equal status of every object in Python. One of the main goals of Python was to have an equal status for all the objects. Remember how even a function is an object in Python and hence it can be assigned to a variable, passed as an argument to an another function, etc.


filter(function or None, iterable)

The first argument is a function that you want each of the elements of the following iterables to be passed as an argument and be evaluated.

Other than the function object, the filter built-in should have one iterable as an argument such that the arguments for the function is taken from the iterable.

Filter takes two arguments
>>> def isdivisibleby2(x):
...     if x % 2 == 0:
...         return True
...     return False
...
>>> filter([1,2,3,4])
Traceback (most recent call last):
  File "", line 1, in 
TypeError: filter expected 2 arguments, got 1
>>> filter(isdivisibleby2)
Traceback (most recent call last):
  File "", line 1, in 
TypeError: filter expected 2 arguments, got 1
>>> filter(isdivisibleby2, [1,2,3,4], [5,6,7,8])
Traceback (most recent call last):
  File "", line 1, in 
TypeError: filter expected 2 arguments, got 3
>>>
Filter Example
>>> def isdivisibleby2(x):
...     if x % 2 == 0:
...         return True
...     return False
...
>>> filtered_list = filter(isdivisibleby2, [1, 2, 3, 4])
>>> filtered_list
<filter object at 0x7f04cb644da0>
>>> list(filtered_list)
[2, 4]
>>>
Filter evaluates Truthy and Falsy

Filter built-in returns a filtered iterator which contains only those values for which the function(passed as the first argument to the filter) evaluated truth value(truthy). An empty sequence such as an empty list [], empty dictionaries, 0 for numeric, None are considered false values or falsy. Almost anything excluding the earlier mentioned are considered truthy. You should read this post on Truthy and Falsy concepts in Python. https://www.thetaranights.com/idiomatic-python-use-of-falsy-and-truthy-concepts/

>>> def arbitrary_function(x):
...     return x
...
>>> filtered_list = filter(arbitrary_function, [1, 2, 3, 4])
>>> filtered_list
<filter object at 0x7f04cb5e9550>
>>> list(filtered_list)
[1, 2, 3, 4]
>>>
>>> def arbitrary_function(x):
...     return 0 # any of False, None, [], {}
...
>>> filtered_list = filter(arbitrary_function, [1, 2, 3, 4])
>>> filtered_list
<filter object at 0x7f04cb5e92b0>
>>> list(filtered_list)
[]
>>>

Python map() built-in

- - Python, Tutorials

Map makes an iterator that takes a function and uses the arguments from the following iterables passed to the map built-in. What makes this possible is the equal status of every object in Python. One of the main goals of Python was to have an equal status for all the objects. Remember how even a function is an object in Python and hence it can be assigned to a variable, passed as an argument to a function, etc.


map(func, *iterables)

The first argument is a function that you want each of the elements of the following iterables to be passed as an argument and be evaluated.

Other than the function object, a map built-in should have at least one iterable and could have iterables as an argument such that the arguments for the function is taken from each of the iterables.

Map takes at least two arguments
>>> def square(x):
...     return x**2
...
>>> map(square)
Traceback (most recent call last):
  File "", line 1, in 
TypeError: map() must have at least two arguments.
>>>
Map Example
>>> def square(x):
...     return x**2
...
>>> squared = map(square, [1,2,3,4,5])
>>> squared
<map object at 0x7f1948bbbef0>
>>> list(squared)
[1, 4, 9, 16, 25]
>>>
Map could take multiple iterables
>>> def add_and_square(x, y):
...     return (x+y)**2
...
>>> added_and_squared = map(add_and_square, [1,2,3,4], [5,6,7,8])
>>> added_and_squared
<map object at 0x7f1948b79518>
>>> list(added_and_squared)
[36, 64, 100, 144]
>>>
When you pass iterables of varying length
>>> def add_and_square(x, y):
...     return (x+y)**2
...
>>> added_and_squared = map(add_and_square, [1,2,3,4], [5,6,7,8, 9])
>>> added_and_squared
<map object at 0x7f1948b795f8>
>>> list(added_and_squared)
[36, 64, 100, 144]
>>>

When you pass iterables of varying length to map built-in, it falls back to the minimum length.

Examples of Browser Automations using Selenium in Python

- - Python, Tutorials

Browser Automation is one of the coolest things to do especially when there is a major purpose to it. Through this post, I intend to host a set of examples on browser automation using selenium in Python so people can take ideas from the code snippets below to perform browser automation as per their need. Selenium allows just about any kinds of interactions with the browser elements and hence is a go for tasks requiring user interaction and javascript support.

Installation:


pip install selenium
Download chromedriver from http://chromedriver.chromium.org/downloads
Download phantomjs from http://phantomjs.org/download.html

Login to a website using selenium
>>> from selenium import webdriver
>>> from selenium.webdriver.common.keys import Keys
>>> executable_path = "/home/bhishan-1504/Downloads/chromedriver_linux64/chromedriver"
>>> browser = webdriver.Chrome(executable_path=executable_path)
>>> browser.get("https://github.com/login")
>>> username_field = browser.find_element_by_name("login")
>>> password_field = browser.find_element_by_name("password")
>>> username_field.send_keys("bhishan")
>>> password_field.send_keys("password")
>>> password_field.send_keys(Keys.RETURN)
>>>
Switching proxy with selenium

As much as selenium is used for web scraping, it is very effective for web interactions too. Suppose a scenario where you have to cast a vote for a competition, one vote per IP address. Following example demonstrates how you would use selenium to perform a repetitive task(casting a vote in this case) from various IP addresses.

from selenium import webdriver
from selenium.webdriver.support.ui import WebDriverWait
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.common.by import By
url = "somedummysite.com/voting/bhishan.php" # url not made public



def cast_vote(proxy):
    service_args = [
    '--proxy=' + proxy,
    '--proxy-type=http',
    ]
    print(service_args)
    browser = webdriver.PhantomJS(service_args=service_args)
    
    browser.get(each_url)
    try:
        cast_vote_element = WebDriverWait(browser, 10).until(
            EC.presence_of_element_located((By.CLASS_NAME, 'vote'))
        )
    except selenium.common.exceptions.TimeoutException:
        print("Cast vote button not available. Seems like you have used this IP already!")
        return
    cast_vote_element.click()
    browser.quit()

def main():
    with open(proxies.txt', 'rb') as f:
        for each_ip in f:
            cast_vote(each_ip.strip())



if __name__ == '__main__':
    main()
Execute JavaScript using selenium

There could be cases where you’d want to execute javascript on the browser instance. The below example is a depiction of one such scenario. Remember when in your News Feed on facebook, a post has hundreds of thousands of comments and you have to monotonously click to expand the comment threads. The example below does it through selenium but has an even bigger purpose. The following code snippet loops over a few thousand facebook urls(relating to a post) and expands the comment threads and prints the page as a pdf file. This was a part of a larger program that had something to do with the pdf files. However, it isn’t relevant to this post. Here is a link to the JavaScript code which is used in the program below that expands the comments on facebook posts. I don’t even remember where I found it though.

from selenium import webdriver
from selenium.webdriver.common.by import By
from selenium.webdriver.common.keys import Keys
from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.support.wait import WebDriverWait

import json
import time


# get the js to be executed.

with open('js_code.txt', 'r') as f:
    js_code = f.read()

executable_path = '/home/bhishan-1504/Downloads/chromedriver_linux64/chromedriver'


appState = {
    "recentDestinations": [
        {
            "id": "Save as PDF",
            "origin": "local"
        }
    ],
    "selectedDestinationId": "Save as PDF",
    "version": 2
}


profile = {"printing.print_preview_sticky_settings.appState": json.dumps(appState), 'savefile.default_directory': "/home/bhishan-1504/secret_project/"}

profile["download.prompt_for_download"] = False
profile["profile.default_content_setting_values.notifications"] = 2
chrome_options = webdriver.ChromeOptions()

chrome_options.add_experimental_option('prefs', profile)
chrome_options.add_argument("--start-maximized")
chrome_options.add_argument('--kiosk-printing')

# chrome_options.add_argument("download.default_directory=/home/bhishan-1504/secret_project/")
browser = webdriver.Chrome(executable_path=executable_path, chrome_options=chrome_options)

def save_pdf(count):
    browser.execute_script("document.title=" + str(count) + ";")
    browser.execute_script('window.print();')
    time.sleep(1)


def visit_page(url, count):
    browser.get(url)
    try:
        home_btn = WebDriverWait(browser, 10).until(
            EC.presence_of_element_located((By.LINK_TEXT, "Home"))
        )
    except selenium.common.exceptions.TimeoutException:
        print("Didn’t work out!")
        return

    browser.execute_script(js_code)
    time.sleep(7)
    save_pdf(count)




if __name__ == '__main__':
    count = 1
    # loop through the text file and pass to visit page function.
    with open('urls.txt', 'r') as f:
        for each_url in f.readlines():
            visit_page(each_url, count)
            count += 1

I recently published an article on Web Scraping using BeautifulSoup. You should read it.

Web Scraping – BeautifulSoup Python

- - Python, Tutorials

Data collection from public sources is often beneficial to a business or an individual. As such the term “web scraping” isn’t something new. These data are often wrangled within html tags and attributes. Python is often used for data collection from these sources. The intentions of this post is to host example code snippets so people can take ideas from it to build scrapers as per their needs using BeautifulSoup and urllib module in Python. I will be using github’s trending page https://github.com/trending throughout this post for the examples, especially because it best suits for applying various BeautifulSoup methods.

Installation:

pip install BeautifulSoup4

Get html of a page:
>>> import urllib
>>> resp = urllib.request.urlopen("https://github.com/trending")
>>> resp.getcode()
200
>>> resp.read() # the html
Using BeautifulSoup to get title from a page
>>> import urllib
>>> import bs4
>>> github_trending = urllib.request.urlopen("https://github.com/trending")
>>> trending_soup = bs4.BeautifulSoup(github_trending.read(), "lxml")
>>> trending_soup.title
<title>Trending  repositories on GitHub today · GitHub</title>
>>> trending_soup.title.string
'Trending  repositories on GitHub today · GitHub'
>>>
Find single element by tag name, find multiple elements by tag name
>>> ordered_list = trending_soup.find('ol') #single element
>>>
>>> type(ordered_list)
<class 'bs4.element.Tag'>
>>>
>>> all_li = ordered_list.find_all('li') # multiple elements
>>>
>>> type(all_li)
<class 'bs4.element.ResultSet'>
>>>
>>> trending_repositories = [each_list.find('h3').text for each_list in all_li]
>>> for each_repository in trending_repositories:
...     print(each_repository.strip())
...
klauscfhq / taskbook
robinhood / faust
Avik-Jain / 100-Days-Of-ML-Code
jxnblk / mdx-deck
faressoft / terminalizer
trekhleb / javascript-algorithms
apexcharts / apexcharts.js
grain-lang / grain
thedaviddias / Front-End-Performance-Checklist
istio / istio
CyC2018 / Interview-Notebook
fivethirtyeight / russian-troll-tweets
boyerjohn / rapidstring
donnemartin / system-design-primer
awslabs / aws-cdk
QUANTAXIS / QUANTAXIS
crossoverJie / Java-Interview
GoogleChromeLabs / ndb
dylanbeattie / rockstar
vuejs / vue
sbussard / canvas-sketch
Microsoft / vscode
flutter / flutter
tensorflow / tensorflow
Snailclimb / Java-Guide
>>>
Getting Attributes of an element
>>> for each_list in all_li:
...     anchor_element = each_list.find('a')
...     print("https://github.com" + anchor_element['href'])
...
https://github.com/klauscfhq/taskbook
https://github.com/robinhood/faust
https://github.com/Avik-Jain/100-Days-Of-ML-Code
https://github.com/jxnblk/mdx-deck
https://github.com/faressoft/terminalizer
https://github.com/trekhleb/javascript-algorithms
https://github.com/apexcharts/apexcharts.js
https://github.com/grain-lang/grain
https://github.com/thedaviddias/Front-End-Performance-Checklist
https://github.com/istio/istio
https://github.com/CyC2018/Interview-Notebook
https://github.com/fivethirtyeight/russian-troll-tweets
https://github.com/boyerjohn/rapidstring
https://github.com/donnemartin/system-design-primer
https://github.com/awslabs/aws-cdk
https://github.com/QUANTAXIS/QUANTAXIS
https://github.com/crossoverJie/Java-Interview
https://github.com/GoogleChromeLabs/ndb
https://github.com/dylanbeattie/rockstar
https://github.com/vuejs/vue
https://github.com/sbussard/canvas-sketch
https://github.com/Microsoft/vscode
https://github.com/flutter/flutter
https://github.com/tensorflow/tensorflow
https://github.com/Snailclimb/Java-Guide
>>>
Using class name or other attributes to get element
>>> for each_list in all_li:
...     total_stars_today = each_list.find(attrs={'class':'float-sm-right'}).text
...     print(total_stars_today.strip())
...
1,063 stars today
846 stars today
596 stars today
484 stars today
459 stars today
429 stars today
443 stars today
366 stars today
330 stars today
282 stars today
182 stars today
190 stars today
200 stars today
190 stars today
166 stars today
164 stars today
144 stars today
158 stars today
157 stars today
144 stars today
144 stars today
142 stars today
132 stars today
101 stars today
108 stars today
>>>
Navigate childrens from an element
>>> for each_children in ordered_list.children:
...     print(each_children.find('h3').text.strip())
...
klauscfhq / taskbook
robinhood / faust
Avik-Jain / 100-Days-Of-ML-Code
jxnblk / mdx-deck
faressoft / terminalizer
trekhleb / javascript-algorithms
apexcharts / apexcharts.js
grain-lang / grain
thedaviddias / Front-End-Performance-Checklist
istio / istio
CyC2018 / Interview-Notebook
fivethirtyeight / russian-troll-tweets
boyerjohn / rapidstring
donnemartin / system-design-primer
awslabs / aws-cdk
QUANTAXIS / QUANTAXIS
crossoverJie / Java-Interview
GoogleChromeLabs / ndb
dylanbeattie / rockstar
vuejs / vue
sbussard / canvas-sketch
Microsoft / vscode
flutter / flutter
tensorflow / tensorflow
Snailclimb / Java-Guide
>>>

The .children will only return the immediate childrens of the parent element. If you’d like to get all of the elements under certain element, you should use .descendent

Navigate descendents from an element
>>> for each_children in ordered_list.descendent:
...     # perform operations
Navigating previous and next siblings of elements
>>> all_li = ordered_list.find_all('li')
>>> fifth_li = all_li[4]
>>> # each li element is separated by '\n' and hence to navigate to the fourth li, we should navigate previous sibling twice
...
>>>
>>> fourth_li = fifth_li.previous_sibling.previous_sibling
>>> fourth_li.find('h3').text.strip()
'jxnblk / mdx-deck'
>>>
>>> # similarly for navigating to the sixth li from fifth li, we would use next_sibling
...
>>> sixth_li = fifth_li.next_sibling.next_sibling
>>> sixth_li.find('h3').text.strip()
'trekhleb / javascript-algorithms'
>>>
Navigate to parent of an element
>>> all_li = ordered_list.find_all('li')
>>> first_li = all_li[0]
>>> li_parent = first_li.parent
>>> # the li_parent is the ordered list <ol>
...
>>>
Putting it all together(Github Trending Scraper)
>>> import urllib
>>> import bs4
>>>
>>> github_trending = urllib.request.urlopen("https://github.com/trending")
>>> trending_soup = bs4.BeautifulSoup(github_trending.read(), "lxml")
>>> ordered_list = trending_soup.find('ol')
>>> for each_list in ordered_list.find_all('li'):
...     repository_name = each_list.find('h3').text.strip()
...     repository_url = "https://github.com" + each_list.find('a')['href']
...     total_stars_today = each_list.find(attrs={'class':'float-sm-right'}).text
…        print(repository_name, repository_url, total_stars_today)

klauscfhq / taskbook                             https://github.com/klauscfhq/taskbook                             1,404 stars today
robinhood / faust                                https://github.com/robinhood/faust                                960 stars today
Avik-Jain / 100-Days-Of-ML-Code 	         https://github.com/Avik-Jain/100-Days-Of-ML-Code                  566 stars today
trekhleb / javascript-algorithms 	         https://github.com/trekhleb/javascript-algorithms                 431 stars today
jxnblk / mdx-deck 			         https://github.com/jxnblk/mdx-deck 	                           416 stars today
apexcharts / apexcharts.js 		         https://github.com/apexcharts/apexcharts.js 	                   411 stars today
faressoft / terminalizer 		         https://github.com/faressoft/terminalizer 	                   406 stars today
istio / istio 			                 https://github.com/istio/istio 	                           309 stars today
thedaviddias / Front-End-Performance-Checklist 	 https://github.com/thedaviddias/Front-End-Performance-Checklist   315 stars today
grain-lang / grain 			         https://github.com/grain-lang/grain 	                           301 stars today
boyerjohn / rapidstring 			 https://github.com/boyerjohn/rapidstring 	                   232 stars today
CyC2018 / Interview-Notebook 			 https://github.com/CyC2018/Interview-Notebook 	                   186 stars today
donnemartin / system-design-primer 		 https://github.com/donnemartin/system-design-primer 	           189 stars today
awslabs / aws-cdk 			         https://github.com/awslabs/aws-cdk 	                           186 stars today
fivethirtyeight / russian-troll-tweets 		 https://github.com/fivethirtyeight/russian-troll-tweets 	   159 stars today
GoogleChromeLabs / ndb 			         https://github.com/GoogleChromeLabs/ndb 	                   172 stars today
crossoverJie / Java-Interview 			 https://github.com/crossoverJie/Java-Interview 	           148 stars today
vuejs / vue 			                 https://github.com/vuejs/vue 	                                   137 stars today
Microsoft / vscode 			         https://github.com/Microsoft/vscode 	                           137 stars today
flutter / flutter 			         https://github.com/flutter/flutter 	                           132 stars today
QUANTAXIS / QUANTAXIS 			         https://github.com/QUANTAXIS/QUANTAXIS 	                   132 stars today
dylanbeattie / rockstar 			 https://github.com/dylanbeattie/rockstar 	                   130 stars today
tensorflow / tensorflow 			 https://github.com/tensorflow/tensorflow 	                   106 stars today
Snailclimb / Java-Guide 			 https://github.com/Snailclimb/Java-Guide 	                   111 stars today
WeTransfer / WeScan 			         https://github.com/WeTransfer/WeScan 	                           118 stars today


Python Lists

- - Python, Tutorials

The intentions of this article is to host a set of example operations that can be performed around lists, a crucial data structure in Python.

Lists

In Python, List is an object that contains a sequence of other arbitrary objects. Lists unlike tuples are mutable objects.

Defining a list

Lists are defined by enclosing a sequence of objects inside square brackets, “[” and “]”. A list can contain sequence of mixed data types.

>>> # empty list
...
>>> a = []
>>>
>>> type(a)
<class 'list'>
>>>
>>> # list containing same data types
...
>>>
>>> a = [1, 4, 9, 16]
>>>
>>> type(a)
<class 'list'>
>>>
>>> # list containing different data types
...
>>> a = [1, "python", 7.4, True]
>>>
>>> type(a)
<class 'list'>
>>>
A list can be nested
>>> # nested list
...
>>> a = [[1, 4, 9], ["thetaranights.com", "blog", "python"]]
>>> type(a)
<class 'list'>
>>>
>>>
>>> a = [[1, 4, 9], "thetaranights.com"]
>>>
>>> type(a)
<class 'list'>
>>>
Accessing elements from a list via index

List index is used to access elements of a list. List index starts from 0 and should be an integer.

>>> a = [[1, 4, 9], ["thetaranights.com", "blog", "python"]]
>>> a[0]
[1, 4, 9]
>>> a[0][2]
9
>>> a[1][0]
'thetaranights.com'
>>>
Negative Indexing

Python allows accessing elements from a list via negative indexes such that the last element would be accessed via list_name[-1] and second last element would be accessed via list_name[-2]

>>> a = [1, 4, 9, 16]
>>> a[-1]
16
>>> a[-2]
9
>>> a[-3]
4
>>>
Slicing
>>> a = [1, 4, 9, 16, 25]
>>> a[1:3]
[4, 9]
>>> a[:4]
[1, 4, 9, 16]
>>> a[3:]
[16, 25]
>>> a[:-1]
[1, 4, 9, 16]
>>>
Lists are mutable
>>> a = [1, 4, 9, 16, 25]
>>>
>>> a [0] = "mutable"
>>> a
['mutable', 4, 9, 16, 25]
>>>
>>> a[:2] = [256, 1024] # changing a range of elements of a list to the sequence given in the right of assignment operator
>>> a
[256, 1024, 9, 16, 25]
>>>
Adding elements to an existing list
>>> a = [1, 4, 9, 16, 25]
>>> a.append(36)
>>> a
[1, 4, 9, 16, 25, 36]
>>>
Extending a list with another sequence
>>> a = [1, 4, 9, 16]
>>> a.extend([25, 36, 49])
>>> a
[1, 4, 9, 16, 25, 36, 49]
>>> a.extend((64, 91))
>>> a
[1, 4, 9, 16, 25, 36, 49, 64, 91]
>>>
Concatenation and Multiplication
>>> a = [1, 4, 9, 16]
>>> a + [25, 36, 49]
[1, 4, 9, 16, 25, 36, 49]
>>>

>>> # multiplication
...
>>> a * 7
[1, 4, 9, 16, 1, 4, 9, 16, 1, 4, 9, 16, 1, 4, 9, 16, 1, 4, 9, 16, 1, 4, 9, 16, 1, 4, 9, 16]
>>>
Add items to a list before certain index
>>> # first item to the insert() is the index and the later is the value to insert
...
>>> a = [1, 4, 9, 16, 25]
>>> a.insert(2, "new value")
>>> a
[1, 4, 'new value', 9, 16, 25]
>>>
Various other methods on a list
method description usage
append() Append object to the end of list L.append(object)
clear() Remove all the items from list L.clear()
copy() A shallow copy of list L.copy()
count() Return number of occurrences of value passed as argument to the method L.count(value)
extend() Extend list by appending elements from the iterable L.extend(iterable)
index() Return first index of the value L.index(value, [start, [stop]])
insert() Insert object before index L.insert(index, object)
pop() Remove and return item at index (defaults to last) L.pop([index])
remove() Remove first occurrence of value L.remove(value)
reverse() Reverse the list in-place L.reverse()
sort() Sort in-place L.sort(key=None, reverse=False)
List built-ins
built-in description
len() Return the number of elements in a list
max() Returns the largest element in the list
min() Returns the smallest element in the list
sorted() Returns the sorted version of the list. It does not sort the given list itself.
sum() Returns the sum of all the elements of the list
all() Returns true if all the elements of the list evaluate to true (See truthy and falsy concepts)
any() Returns true if any element of the list evaluates to true
enumerate() Returns enumerate object that contains the index and corresponding values of an iterable.
list() Converts an iterable (tuple, string, set, dictionary) to a list.
List Comprehension

One of the major features of python is list comprehension. It is a natural way of creating a new list where each element is the result of some operations applied to each member of another sequence of an iterable. The construct of a list comprehension is such that it consists of square brackets containing an expression followed by a for clause then by zero or more for or if clause. List comprehensions always returns a list.

>>> [x ** 2 for x in range(1, 11)]
[1, 4, 9, 16, 25, 36, 49, 64, 81, 100]
>>>

In a rather real usage scenarios, the expression after the bracket ‘[‘ is a call to a method/function.

some_list = [function_name(x) for x in some_iterable]

Python Tuples

- - Python, Tutorials

This is an introductory post about tuples in python. We will see through examples what are tuples, its immutable property, use cases, various operations on it. Rather than a blog, it is a set of examples on tuples in python

Tuples

It is a sequence of objects in python. Unlike lists, tuple are immutable which means the contents of a tuple can’t be changed once assigned. We will see in a bit through example immutable property of tuples.

Defining a Tuple

Tuples are generally created by enclosing a sequence of objects inside parentheses. “(” and “)”

>>> ip_addresses = ("172.19.56.90", "172.37.57.32", "172.54.21.23")
>>> type(ip_addresses)

>>>
Defining an empty tuple vs single element tuple vs multi-element tuple
>>> # empty tuple
...
>>> ip_addresses = ()
>>> type(ip_addresses)
<class 'tuple'>
>>>
>>> # multi-element tuple
...
>>> ip_addresses = ("172.19.56.90", "172.37.57.32", "172.54.21.23")
>>> type(ip_addresses)
<class 'tuple'>
>>>
>>> # single element tuple
...
>>> ip_addresses = ("172.19.56.90") # incorrect
>>> type(ip_addresses)
<class 'str'>
>>>
>>> ip_addresses = ("172.19.56.90",)
>>> type(ip_addresses)
<class 'tuple'>
>>>

From the code snippet above, the method of defining an empty tuple and multi-element tuple seems obvious. However, what’s not obvious is that ip_addresses = (“172.19.56.90”) evaluates to a str type instead of a tuple. Although a single element tuple is very rare to come in use, there had to be a way to define it. Hence, a single element tuple should end with a comma “,” for the interpreter to evaluate it as a tuple.

Parentheses is also optional
>>> ip_addresses = "172.19.56.90", "172.37.57.32", "172.54.21.23"
>>> type(ip_addresses)
<class 'tuple'>
>>>

It is also optional to have the parentheses to define a tuple. This is possible due to a mechanism called packing which is one of the many useful features of Python.

Packing and Unpacking

Packing as the name suggest is creating an object by packing multiple other objects to make one compact object. Unpacking on the other hand is the vice-versa such that an object is unpacked and assigned to variables the elements of the tuple.

>>> # packing example
...
>>> ip_addresses = "172.19.56.90", "172.37.57.32", "172.54.21.23"
>>>

>>> # unpacking example
...
>>> ip, ip2, ip3 = ip_addresses
>>> ip
'172.19.56.90'
>>> ip2
'172.37.57.32'
>>> ip3
'172.54.21.23'
>>>
The number of variables on the left should equal the number of elements to be unpacked
>>> ip, ip2 = ip_addresses
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: too many values to unpack (expected 2)
>>>
>>>
>>> ip, ip2, ip3, ip4 = ip_addresses
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: not enough values to unpack (expected 4, got 3)
>>>
Accessing elements from a tuple through index
>>> ip_addresses = ("172.19.56.90", "172.37.57.32", "172.54.21.23")
>>>
>>> ip_addresses[1]
'172.37.57.32'
>>>
>>> ip_addresses[-1]
'172.54.21.23'
>>>
Looping over elements of a tuple
>>> ip_addresses = ("172.19.56.90", "172.37.57.32", "172.54.21.23")
>>> for each_ip in ip_addresses:
...     print(each_ip)
...
172.19.56.90
172.37.57.32
172.54.21.23
>>>
>>>
Slicing
>>> ip_addresses[1:]
('172.37.57.32', '172.54.21.23')
>>>

>>> ip_addresses[:-1]
('172.19.56.90', '172.37.57.32')
>>>
Tuples are immutable
>>> ip_addresses[0] = "Accidental edit"
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: 'tuple' object does not support item assignment
>>>

Tuples are immutable. This avoids accidental data change attempts.

Concatenation and Multiplication
>>> us_east_ips = ("172.19.56.90", "172.37.57.32", "172.54.21.23")
>>>
>>> us_west_ips = ("172.18.11.22", "172.99.22.3")
>>>
>>> type(us_east_ips)
<class 'tuple'>
>>> type(us_west_ips)
<class 'tuple'>
>>>
>>> all_ips = us_east_ips + us_west_ips
>>> type(all_ips)
<class 'tuple'>
>>> all_ips
('172.19.56.90', '172.37.57.32', '172.54.21.23', '172.18.11.22', '172.99.22.3')
>>>

Concatenation of two tuples returns a third tuple which contains the all the contents of the both tuples copied in order.

>>> duplicate_values = ("duplicate",)
>>> type(duplicate_values)
<class 'tuple'>
>>> duplicate_values * 5
('duplicate', 'duplicate', 'duplicate', 'duplicate', 'duplicate')
>>>
Count elements in a tuple
>>> arbitrary_values = ('thetaranights.com', 'python', 'python', 'tutorials')
>>>
>>> arbitrary_values.count('python')
2
>>>
>>>
Find index of an element
>>> arbitrary_values = ('thetaranights.com', 'python', 'python', 'tutorials')
>>>
>>> arbitrary_values.index('tutorials')
3
>>>
Comparison of tuples

Comparison operators work for tuples too. The evaluation starts by comparing the first elements from the either tuples and proceeds on further elements until conclusive.

>>> tup = (7, 14, 20)
>>>
>>> tup2 = (7, 14, 21)
>>>
>>> tup < tup2
True
>>>
>>> tup > tup2
False
>>>

For tup < tup2

It compares the first elements from either tuples 7 < 7 which is inconclusive, it then proceeds to comparing 14 < 14, still inconclusive, finally 20 < 21, hence True
Similar is the case for tup > tup2.