Please help me troubleshoot Beanscrape- not exporting transactions properly

Bugs
1

Hi,

I’m a new user to Surebeans and am trying to set up Beanscrape to scrape transactions from my bank account, and can’t get it to export transactions properly. As a heads up, I do not have a background in coding or computer science.

I’m using the desktop application for Surebeans and having been following the User Guide. I was able to make a new journal, and begin the process of making a .csx script for my bank, since it does not seem like one exists for it already. I went through the part where I teach Beanscrape what to do. However, I noticed that when I try to export my transaction, although I can see the exported file, I’m not actually able to open it.

I then moved on to copy/pasting the instructions into an LLM and testing the output. After troubleshooting a few issues, I got the script to consistently get me to my transactions page. However, the script then stops and seems unable to figure out how to export transactions. Even if I help it by doing it myself during the test, I still cannot open the exported file, and Beanscrape doesn’t scrap any transactions.

Can somebody please help me figure out what the issue is? I’m happy to provide more information.

2

Hey @ngills ! Thanks for posting and for trying Surebeans. I’d be glad to help you. So sorry it took so long for me to see this. I’ll have to adjust my notifications.

Thanks for the detailed write up. You’ve been following the process perfectly, so there’s nothing wrong with the steps you took. There’s certainly no need to be a coder to set up Beanscrape. That’s what the LLM is supposed to take care of.

It seems to me that Beanscrape is struggling to find or click the export button, or if it does, to parse the downloaded file. One thing you can try right away is to re-teach Beanscrape and make sure you clicked the export button during teaching.

Other than that, would help me is if you could provide the detailed log that Surebeans shows when Beanscrape is running. I can provide a screenshot of where that is when I get to a computer later today. The csx file would also be helpful. Note that both of these can contain personal information, so please feel free to either DM me or email [email protected].

Thanks again for trying Surebeans! We’ll get you set up.

1 Like
3

Doug,

Thank you for the reply.

I tried re-teaching and ran into the same issues as before. I’ll DM you with more information.

4

ngills and I backchanneled and got him going. Circling back here for the benefit of everyone else.

The problem was threefold:

  1. The script name was e.g. bank2.csx but BankId was “bank”. Surebeans assumed the file name would match the BankId. Fixed in 0.80.1.

  2. On Windows, a bug in Beanscrape prevented saving bank credentials to secure storage. Fixed in 0.19.2.

  3. The generated script wasn’t stuck. It was just very slow, waiting 30s after each button click. By default, Playwright, the tool Beanscrape uses to automate a browser, waits for a page navigation after any button click. This particular bank uses buttons that don’t cause a navigation. Rather, it’s a single page app (SPA) and button clicks cause document updates on the same page.

    After waiting Playwright’s default click timeout of 30s, it did proceed, but that’s too long to be practical since there were 5+ button clicks. The fix is to tell Playwright to not wait for a page navigation.

    Before: await ClickAsync(loadMoreBtn.First);
    After: await loadMoreBtn.ClickAsync(new() { NoWaitAfter = true });

1 Like
5

Also following up on this:

since it does not seem like one exists for it already

Correct. As a legal precaution, Surebeans does not distribute scripts. But just by coincidence, I also have an account at ngills’ bank and already had a working script for it. That’s how I deduced the NoWaitAfter solution above.

1 Like
6

Thank you for all the help Doug!

1 Like