Total failure needs Help!

I love the idea of crewAI, but as of yet, have been totally unsuccessful in getting a crew to work.
Debugging this is incredibly difficult. I’ve been using Gemini CLI to help, but its run out of ideas.
I’m doing this all locally, with an ollama model running in another DOS tab on a Windows 11 PC.
I’m willing to pay, some, if someone could look over my crew and tell me what I’m doing wrong.
I’m getting a nonsensical error: “An unexpected error occurred for Lewis, Missouri: ‘s’”, where ‘s’ is the {e} in my try:, except Exception as e: structure.

Can anyone help? Pls contact me at gaiagrower77@gmail.com for faster response. Thank you.

1 Like

Can you please share the code with us?

Hello Rok!

Thank you for asking!! Attached is a zip file of all of the files in my crew.

I’ve got the “classical” crewAI structure. Here’s an what the crew’s file structure looks like:

cboCrew
±- pyproject.toml
±- src/
.env
.venv
±-src
±- cboCrew
±- crew.py
±- main.py
±- lgr.py
±- config
±- agents.yaml
±- tasks.yaml
±- data
±- urls.txt
±- keywords.txt
±- badUrls.txt
±- counties.csv

I’d like to do my first crews locally, rather than crewAI’s AMP(?), at least to start. Gotta be able to demo crews to businesses on my laptop and not count on having web connection.

I have many specific crews I’d like to build, but need to get one working first.

I’m trying to run through all the lines in data\counties.csv, and for each state;county, kick off a new crew to:

  1. Search for all community based organizations in this state and county and grab their organization name and website link, using SerperDevTool() (see instantiation in crew.py).
  2. Crawl each website returned, looking for, in this case, emails, which are then logged to a file: Email;Organization, using ScrapeWebsiteTool() (crew.py).

I’m on a Windows 11 computer. In one DOS window I start the olloma server. In another, I run the appropriate model. In the 3rd, I run the crew.
I’ve run crewai install, activate the .venv, and then: crewai run. For every state;county line in the source csv file, i.e., for every time a crew is kicked off, I’m getting an error like this:
“An unexpected error occurred for Douglas, Missouri: ‘s’”. Douglas and Missouri are the current values of the state and county variables and filled in as part of the Exception clause.
The ‘s’ is the error message, the “as e” in that clause. How in the heck can I debug something like that?

Maybe there’s something obvious that you can see. Gemini CLI has not been of help.
I’m in Oregon, PST. It’s a few minutes after 5pm here now. I’ll be able to see any of your emails, and will reply w/short messages during the day.
Usually working with the computer in late afternoon until late evening.

I’m looking forward to your thoughts. And, thank you for offering!
Baruch Bashan

(Attachment cboCrew.zip is missing)

Hello Rok!

Thank you for asking!! I’m not “allowed” to send you a zip file of all of the files in my crew. Nor am I allowed to share access to my google drive’s folder.
Here’s the link. Please let me know if you can get the zip file or not. https://drive.google.com/drive/folders/17qe-gE-tQghyXoaWRkA-yaEjAz4A26Wg?usp=sharing

I’ve got the “classical” crewAI structure. Here’s an what the crew’s file structure looks like:

cboCrew
±- pyproject.toml
±- src/
.env
.venv
±-src
±- cboCrew
±- crew.py
±- main.py
±- lgr.py
±- config
±- agents.yaml
±- tasks.yaml
±- data
±- urls.txt
±- keywords.txt
±- badUrls.txt
±- counties.csv

I’d like to do my first crews locally, rather than crewAI’s AMP(?), at least to start. Gotta be able to demo crews to businesses on my laptop and not count on having web connection.

I have many specific crews I’d like to build, but need to get one working first.

I’m trying to run through all the lines in data\counties.csv, and for each state;county, kick off a new crew to:

  1. Search for all community based organizations in this state and county and grab their organization name and website link, using SerperDevTool() (see instantiation in crew.py).
  2. Crawl each website returned, looking for, in this case, emails, which are then logged to a file: Email;Organization, using ScrapeWebsiteTool() (crew.py).

I’m on a Windows 11 computer. In one DOS window I start the olloma server. In another, I run the appropriate model. In the 3rd, I run the crew.
I’ve run crewai install, activate the .venv, and then: crewai run. For every state;county line in the source csv file, i.e., for every time a crew is kicked off, I’m getting an error like this:
“An unexpected error occurred for Douglas, Missouri: ‘s’”. Douglas and Missouri are the current values of the state and county variables and filled in as part of the Exception clause.
The ‘s’ is the error message, the “as e” in that clause. How in the heck can I debug something like that?

Maybe there’s something obvious that you can see. Gemini CLI has not been of help.
I’m in Oregon, PST. It’s a few minutes after 5pm here now. I’ll be able to see any of your emails, and will reply w/short messages during the day.
Usually working with the computer in late afternoon until late evening.

I’m looking forward to your thoughts. And, thank you for offering!
Baruch Bashan