How to use test_click_for_shadow_dom_v1 method in Playwright Python

You can select it by attribute:

self.page.locator('[id=\'#A/BC\']').click()

I had the same issue and it was because of an unclosed page instance, however, I see in your code page is closed.

Please make sure you did the following steps:

1. Close all playwright instances
2. Close all BrowserContext objects
3. Close all opened Page objects

Opened Playwright instance looks like this:

playwright_instance = await async_playwright().start()
browser = await playwright_instance.firefox.launch(headless=headless, proxy=PROXIES[0])
context = await browser.new_context(java_script_enabled=True)

Here is my solution:

await page.close()
await browser.close()
await playwright_instance.stop()

Reading your question I understand that your problem is not the download itself, your problem is the way of clicking the link that you need.

From that moment, you have to options:

1 - You know the date (exact text):

date = "2022-09-26"
with page.expect_download() as download_info:
    page.locator(f"//a[text()='{date}_TEST.zip']").click()
    download = download_info.value
    page.wait_for_url("http://192.xx")
    download_filepath = os.path.join(download_path, suggested_filename)
    download.save_as(download_filepath)
    context.close()
    browser.close()

As you can see you can use f"..{}.." for your locator string in order to use some vars for the locator, basically a locator is a string.

Of course you should know the date before, in my example is hardcoded.

2 - You don't know the date (Contains text):

with page.expect_download() as download_info:
    page.locator(f"//a[contains(text(),'_TEST.zip')]").click()
    download = download_info.value
    page.wait_for_url("http://192.xx")
    download_filepath = os.path.join(download_path, suggested_filename)
    download.save_as(download_filepath)
    context.close()
    browser.close()

And being honest, I think the line page.wait_for_url("http://192.xx") is not needed. Playwright will wait for your download, actually you don't need to wait for the url, I think.

For my examples I used selector with Xpath.

You should be able to get that button by text:

page.locator('"Log in"').click()

If you need to handle new page without event listener (i.e. new tab open by link click), you can try this code:

    from playwright import sync_playwright

    with sync_playwright() as p:
        browser = p.chromium.launch()
        context = browser.newContext()
        page = context.newPage()
        page.goto('<site url>')
        with context.expect_page() as tab:
            page.click('.newTabByLink')
        # do some steps
        ...
        tab.close()
        
        browser.close()

I know you said you're sure the element exists but I would still suggest using page.wait_for_selector() instead because I've run into a few situations where the element simply wasn't ready yet even though it seemed like it should be:

table = page.wait_for_selector('#grvPedidos')
print(table.inner_html())

Note that I'm using synchronous Playwright.

You can not interact with extensions in playwright. You could script windows tasks like clicking, waiting, pressing keys etc and run it with headfull browser mode.

So, it seems that this is simply not directly possible with python-playwright, nor is it going to be. From this discussion, I gathered that the dialog...

"[...] is not a part of the web, it is a part of the browser, so we don't automate it."

There are workarounds, e.g. to trigger a download in a different way, but this would then not involve handling the browser dialog, as stated in the question.

The locator will help you find an element. But from there, if you don't have an element to pierce the Shadow DOM, you might need to do that manually.

text = page.locator('some-element').first.evaluate("node => node.shadowRoot.innerHTML")