fix: issues with save function

[rich-iannone]

This adds a workflow job that tests whether the .save() method is properly producing a table image when using the "firefox" webdriver.

Also fixes: #241

[machow]

I would just run pip install pandas for now after this, since we need it for the datasets

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.71%. Comparing base (ed09f9d) to head (c4bae56).
Report is 27 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #276      +/-   ##
==========================================
- Coverage   81.72%   81.71%   -0.01%     
==========================================
  Files          41       41              
  Lines        4323     4321       -2     
==========================================
- Hits         3533     3531       -2     
  Misses        790      790              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[rich-iannone]

This fails with:

python .github/scripts/save_browser_table.py
Traceback (most recent call last):
  File "/home/runner/work/great-tables/great-tables/.github/scripts/save_browser_table.py", line 3, in <module>
    GT(exibble).save("exibble_firefox.png", web_driver="firefox")
  File "/home/runner/work/great-tables/great-tables/great_tables/_export.py", line 181, in save
    wdriver(options=wd_options) as headless_browser,
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/hostedtoolcache/Python/3.12.2/x64/lib/python3.12/site-packages/selenium/webdriver/firefox/webdriver.py", line 69, in __init__
    super().__init__(command_executor=executor, options=options)
  File "/opt/hostedtoolcache/Python/3.12.2/x64/lib/python3.12/site-packages/selenium/webdriver/remote/webdriver.py", line 208, in __init__
    self.start_session(capabilities)
  File "/opt/hostedtoolcache/Python/3.12.2/x64/lib/python3.12/site-packages/selenium/webdriver/remote/webdriver.py", line 292, in start_session
    response = self.execute(Command.NEW_SESSION, caps)["value"]
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/hostedtoolcache/Python/3.12.2/x64/lib/python3.12/site-packages/selenium/webdriver/remote/webdriver.py", line 347, in execute
    self.error_handler.check_response(response)
  File "/opt/hostedtoolcache/Python/3.12.2/x64/lib/python3.12/site-packages/selenium/webdriver/remote/errorhandler.py", line 229, in check_response
    raise exception_class(message, screen, stacktrace)
selenium.common.exceptions.WebDriverException: Message: Process unexpectedly closed with status 1

make: *** [Makefile:53: save-browser-table] Error 1

Probably because Firefox is run in non-headless mode: https://stackoverflow.com/questions/46809135/webdriver-exceptionprocess-unexpectedly-closed-with-status-1

Now, technically it runs in headless mode on my system but the resulting image is incorrect (only gives correct image output when not run in headless mode).

[rich-iannone]

I suppose we could instead run with chrome using: https://github.com/browser-actions/setup-chrome.

[machow]

Ah, dang--what does the image look like? It looks there are options to run FF headless in that SO post and this one, so I'm curious if there's a way to get it to work?!

[rich-iannone]

The firefox screenshot looks like this in headless mode:

and otherwise just fine if we run w/o headless:

I'll try out some of the things in that SO post to see if there's a way to work around this!

[rich-iannone]

Seems like "---headless=new" solved the problem! See https://www.selenium.dev/blog/2023/headless-is-going-away/. I'll test this with the rest of the web drivers to ensure everything works with that.

[rich-iannone]

With "---headless=new" all four browsers now work well in headless mode. Pushed the change.

[machow]

• [ ]