Select iframe using Python + Selenium

2019-01-01 15:38发布

So, I was absolutely baffled as to how to do this in Selenium, and couldn't find the answer anywhere, so I'm sharing my experience.

I was trying to select an iframe and having no luck (or not repeatably anyway). The HTML looked like this:

<iframe id="upload_file_frame" width="100%" height="465px" frameborder="0" framemargin="0" name="upload_file_frame" src="/blah/import/">
<html>
    <body>
        <div class="import_devices">
            <div class="import_type">
                <a class="secondary_button" href="/blah/blah/?source=blah">
                    <div class="import_choice_image">
                        <img alt="blah" src="/public/images/blah/import/blah.png">
                    </div>
                    <div class="import_choice_text">Blah Blah</div>
                </a>
            </div>
        </div>
    </body>
</html>

The Python code (using the selenium library) was trying to find this iframe using this:

    @timed(650)
def test_pedometer(self):
    sel = self.selenium
    ...
    time.sleep(10)
    for i in range(5):
        try:
            if sel.select_frame("css=#upload_file_frame"): break
        except: pass
        time.sleep(10)
    else: self.fail("Cannot find upload_file_frame, the iframe for the device upload image buttons")

Repeated fails with every combination of Selenium commands I could find. The occasional success would not be reproducible, so perhaps it was some sort of race condition or something? Never did find the right way to get it in selenium proper.

5条回答
一个人的天荒地老
2楼-- · 2019-01-01 16:15

What finally worked for me was:

        sel.run_script("$('#upload_file_frame').contents().find('img[alt=\"Humana\"]').click();")

Basically, don't use selenium to find the link in the iframe and click on it; use jQuery. Selenium has the capability to run an arbitrary piece of javascript apparently (this is python-selenium, I am guessing the original selenium command is runScript or something), and once I can use jQuery I can do something like this: Selecting a form which is in an iframe using jQuery

查看更多
旧时光的记忆
3楼-- · 2019-01-01 16:22

This worked for me with Python (v. 2.7), webdriver & Selenium when testing with iframes and trying to insert data within an iframe:

self.driver = webdriver.Firefox()

## Give time for iframe to load ##
time.sleep(3)
## You have to switch to the iframe like so: ##
driver.switch_to.frame(driver.find_element_by_tag_name("iframe"))
## Insert text via xpath ##
elem = driver.find_element_by_xpath("/html/body/p")
elem.send_keys("Lorem Ipsum")
## Switch back to the "default content" (that is, out of the iframes) ##
driver.switch_to.default_content()
查看更多
低头抚发
4楼-- · 2019-01-01 16:26

You don't need to use JavascriptExecutor. All you needed to do was switch into the frame and then switch back out, like so:

// do stuff on main window
driver.switch_to.frame(frame_reference)
// then do stuff in the frame
driver.switch_to.default_content()
// then do stuff on main window again

As long as you are careful with this, you will never have a problem. The only time I always use a JavascriptExecutor is to get window focus since I think using Javascript is more reliable in that case.

查看更多
何处买醉
5楼-- · 2019-01-01 16:35

If iframe is dynamic node, it's also possible to wait for iframe appearence explicitly and then switch to it using ExpectedConditions:

from selenium.webdriver.support import expected_conditions as EC
from selenium.webdriver.support.ui import WebDriverWait as wait

driver = webdriver.Chrome()
driver.get(URL)
wait(driver, 10).until(EC.frame_to_be_available_and_switch_to_it("iframe_name_or_id"))

If iframe doesn't have @id or @name it can be found as common WebElement using driver.find_element_by_xpath(), driver.find_element_by_tag_name(), etc..:

wait(driver, 10).until(EC.frame_to_be_available_and_switch_to_it(driver.find_element_by_xpath("//iframe[@class='iframe_class']")))

To switch back from iframe:

driver.switch_to.default_content()
查看更多
妖精总统
6楼-- · 2019-01-01 16:39

Selenium's selectFrame command accepts all the standard locators like css=, but it also has a an extra set of locators that work specifically with FRAME and IFRAME elements.

As the doc says:

selectFrame ( locator ) Selects a frame within the current window. (You may invoke this command multiple times to select nested frames.) To select the parent frame, use "relative=parent" as a locator; to select the top frame, use "relative=top". You can also select a frame by its 0-based index number; select the first frame with "index=0", or the third frame with "index=2".

You may also use a DOM expression to identify the frame you want directly, like this: dom=frames["main"].frames["subframe"]

Arguments: locator - an element locator identifying a frame or iframe

In general, you'll have better luck using the specialized locators, especially if you establish the right context first (e.g., select_frame("relative=top"); select_frame("id=upload_file_frame");).

查看更多
登录 后发表回答