Main features:
- User-oriented API for Selenium Webdriver (code like speak common English)
- Ajax support (Smart implicit waiting and retry mechanism)
- PageObjects support (all elements are lazy-evaluated objects)
- Automatic driver management (no need to install and setup driver for quick local execution)
Selene was inspired by Selenide from Java world.
Tests with Selene can be built either in a simple straightforward "selenide' style or with PageObjects composed from Widgets i.e. reusable element components.
-
Latest recommended version to use is 2.0.0a36
- it's a completely new version of selene, with improved API and speed
- supports python >= 3.7
- it's incompatible with 1.x
- current master branch is pointed to 2.x
- yet in pre-alpha stage, refining API, improving "migratability", and testing
- it looks pretty stable, but not fully proven yet
- mainly tested on production code base of a few users who successfully migrated from 1.x to 2.x
-
Latest version marked as stable is: 1.0.2
- it is main version used by most selene users during last 2 years
- it was proven to be stable for production use
- its sources and corresponding README version can be found at 1.x branch.
- supports python 2.7, 3.5, 3.6, 3.7
THIS README DESCRIBES THE USAGE OF THE PRE-RELEASE version of Selene. For older docs look at 1.x branch.
GIVEN on 1.0.1:
- upgrade to python 3.7
- update selene to 2.0.0aLATEST
- find&replace the collection.first() method from
.first()
to.first
- ensure all conditions like
text('foo')
are used viabe.*
orhave.*
syntax- example:
- find&replace all
(text('foo'))
to(have.text('foo'))
(be.visible)
to(be.visible)
- smarter find&replace (with some manual refactoring)
.should(x, timeout=y)
to.with_(timeout=y).should(x)
- and add corresponding imports:
from selene import be, have
- find&replace all
- example:
- fix another broken imports if available
- run tests, read deprecation warnings, and refactor to new style recommended in warning messages
- find&replace the collection.first() method from
Given pyenv installed, installing needed version of Python is pretty simple:
$ pyenv install 3.7.3
$ pyenv global 3.7.3
$ python -V
Python 3.7.3
GIVEN poetry and pyenv installed ...
AND
$ poetry new my-tests-with-selene
$ cd my-tests-with-selene
$ pyenv local 3.7.3
WHEN latest pre-release recommended version:
$ poetry add selene --allow-prereleases
WHEN latest stable version:
$ poetry add selene
THEN
$ poetry install
Latest recommended pre-release alpha version:
$ pip install selene --pre
Latest stable version:
$ pip install selene
GIVEN webdriver and webdriver_manager are already installed
THEN
$ git clone https://github.com/yashaka/selene.git
$ python setup.py install
or using pip:
$ pip install git+https://github.com/yashaka/selene.git
Simply...
from selene.support.shared import browser
from selene import by, be, have
browser.open('https://google.com/ncr')
browser.element(by.name('q')).should(be.blank)\
.type('selenium').press_enter()
browser.all('.srg .g').should(have.size(10))\
.first.should(have.text('Selenium automates browsers'))
OR with custom setup
from selene.support.shared import browser
from selene import by, be, have
browser.config.browser_name = 'firefox'
browser.config.base_url = 'https://google.com'
browser.config.timeout = 2
# browser.config.* = ...
browser.open('/ncr')
browser.element(by.name('q')).should(be.blank)\
.type('selenium').press_enter()
browser.all('.srg .g').should(have.size(10))\
.first.should(have.text('Selenium automates browsers'))
OR more Selenide from java style:
from selene.support.shared import config, browser
from selene import by, be, have
from selene.support.shared.jquery_style import s, ss
config.browser_name = 'firefox'
config.base_url = 'https://google.com'
config.timeout = 2
# config.* = ...
browser.open('/ncr')
s(by.name('q')).should(be.blank)\
.type('selenium').press_enter()
ss('.srg .g').should(have.size(10))\
.first.should(have.text('Selenium automates browsers'))
Given:
from selenium.webdriver import Chrome
AND chromedriver executable available in $PATH
WHEN:
from selene import Browser, Config
browser = Browser(Config(
driver=Chrome(),
base_url='https://google.com',
timeout=2))
AND (uncomment if needed):
# import atexit
# atexit.register(browser.quit)
AND:
browser.open('/ncr')
AND:
# browser.element('//*[@name="q"]')).type('selenium').press_enter()
# OR...
# browser.element('[name=q]')).type('selenium').press_enter()
# OR...
from selene import by
# browser.element(by.name('q')).type('selenium').press_enter()
# OR...for total readability
query = browser.element(by.name('q')) # actual search doesn't start here, the element is "lazy"
# here the actual webelement is found
query.type('selenium').press_enter()
# and here it's located again, i.e. the element is "dynamic"
AND (in case we need to filter collection of items by some condition like visibility):
from selene import be
results = browser.all('.srg .g').filtered_by(be.visible)
THEN:
from selene import have
# results.should(have.size(10))
# results.first.should(have.text('Selenium automates browsers'))
# OR...
results.should(have.size(10))\
.first.should(have.text('Selenium automates browsers'))
FINALLY (if not registered "atexit" before):
browser.quit()
Instead of:
from selene import Browser, Config
browser = Browser(Config(
driver=Chrome(),
base_url='https://google.com',
timeout=2))
You can simply use the browser and config instance predefined for you in selene.support.shared
module:
from selene.support.shared import browser, config
# ... do the same with browser.*
So you don't need to create you driver instance manually. It will be created for you automatically.
Yet, if you need some special case, like working with remote driver, etc., you can still use shared browser object, while providing driver to it through:
config.driver = my_remote_driver_instance
# or
browser.config.driver = my_remote_driver_instance
Sometimes you might need some extra actions on elements, e.g. for workaround something through js:
from selene import command
browser.element('#not-in-view').perform(command.js.scroll_into_view)
Probably you think that will need something like:
from selene import query
product_text = browser.element('#to-assert-something-non-standard').get(query.text)
price = my_int_from(product_text)
assert price > 100
But usually it's either better to implement your custom condition:
browser.element('#to-assert-something-non-standard').should(have_in_text_the_int_number_more_than(100))
Where the have_in_text_the_int_number_more_than
is your defined custom condition. Such condition-based alternative will be less fragile, because python's assert does not have "implicit waiting", like selene's should ;)
Furthermore, the good test is when you totally control your test data, and instead:
product = browser.element('#to-remember-for-future')
product_text_before = product.get(query.text)
price_before = my_int_from(product_text_before)
# ... do something
product_text_after = product.get(query.text)
price_after = my_int_from(product_text_after)
assert price_after > price_before
Normally, better would be to refactor to something like:
product = browser.element('#to-remember-for-future')
product.should(have.text('100$'))
# ... do something
product.should(have.text('125$'))
You might think you need something like:
from selene import query
if browser.element('#i-might-say-yes-or-no').get(query.text) == 'yes':
# do something...
Or:
from selene import query
if browser.all('.option').get(query.size) >= 2:
# do something...
Maybe one day, you really find a use case:) But for above cases, probably easier would be:
if browser.element('#i-might-say-yes-or-no').wait_until(have.text('yes')):
# do something
# ...
if browser.all('.i-will-appear').wait_until(have.size_greater_than_or_equal(2)):
# do something
Or, by using non-waiting versions, if "you are in a rush:)":
if browser.element('#i-might-say-yes-or-no').matching(have.text('yes')):
# do something
# ...
if browser.all('.i-will-appear').matching(have.size_greater_than_or_equal(2)):
# do something
TBD
TBD
- bump version via
bash .run/bump_version.sh x.x.x
- build via
bash .run/build.sh
- publish via
bash .run/publish.sh
or
bash .run/bump_build_publish.sh x.x.x
or if you want to control all by yourself
- manually bump version in
pyproject.toml
andselene/__init.py:__version__
- poetry publish --build