Sorry for the topic spam, but the old one closed automatically while I was on vacation.
Coming back on the issue, I have now used the chromium image you recommended in my config.
But when trying to use the undetected_chromedriver package to run selenium I still run into an issue I cannot recreate on my own pc. This is the log of the error:
I’m not sure if I could find the path to the chromedriver that is installed through the docker image.
It would be great if we could run undetected_chromedriver as the driver as it allows for easy headless execution. If not, could you maybe provide an example of how you would use selenium on the image?
IT would be more helpfull if you showed the vital parts of your code
No, the driver is not located there, but that is irrelevant, if you are using selenium it will detect the browsers and driver it self. And if you make sure to set chromium to required in package.json it will spawn in a desktop session to chrome is working
The file location is not one I provided, it is auto generated by the undetected_chromedriver package.
This is my import:
This is the code I use to execute the program, I imagine it probably has something to do with the way the package looks for the executable / downloads the custom executable.
I think it is failing due to the chromedriver not being in the default location, as the Patcher.py automatically creates a file path.
From what I understand it automatically downloads a chromedriver that is patched in a way so its harder for websites to detect that it is headless
This seems to break when ran in the docker image however. It won’t download this when giving a custom path, but then I would have to clarify the path of the chromedriver.
Thanks in advance. Please do note that its a package that wraps seleniums driver and thus the behaviour might be a bit different. I think just providing the path to the image chromedriver should resolve the issue, if this is not possible please let me know and I’ll try to work with just selenium instead. It would greatly simplify things for me if this would work though.
I cannot help you with that.
I offer an image that is tested to work with multiple rpa frameworks, but you can look at my dockerfile and see if you can create an image that works for your usecase
You can test your image with something like
docker run -it --rm -e agentid=id-of-agent-that-use-your-test-package -e apiurl=grpc://username:password@grpc.app.openiap.io:443 dockerhub-username/image-name:latest
