A Change Of Plan

So after searching for a solution to the IMAP trigger not working for my Yahoo email account, even trying it with another of my Yahoo accounts, I've come to the conclusion that what I read about Yahoo and IMAP being unreliable must be true. Well, it suits my ego more for it to be that rather than I'm doing something wrong :)

Feeling a bit defeated (I never like not being able to find a solution), I decided to just sign up to the respective AI newsletters using one of my GMail accounts. That's not something that I wanted to do due to my professional pride, but sometimes you just have to admit defeat and move onto another solution. After all, this is day 3 and I still haven't got a working trigger.

The benefit of using a GMail account is that n8n has nodes specifically for the Google suite of tools, and they're meant to be a whole lot more reliable. So now, I'll just set up a GMail trigger to fire when new emails are received. Within these I can then filter out just the ones that I'm interested in and proceed from there.

Firstly though, I've set up a filter in my GMail to label the emails that I'm interested in. This label can then be used in tje GMail trigger so that only the emails that I'm interested are retrieved. This saves a lot of processing instead of retrieving ALL emails and then filtering out the ones that I'm not interested in.

A quick run of the trigger and I can see that a test email that I forwarded to my GMail account has come through. One small issue though, as the data retrieved does not contain the email body, the part that I'm interested in. A quick guess that this might be down to the "Simple" setting on the GMail trigger being set to true (when I need it to be false) leads me to all of the email data being retrieved, including the email body.

So, I'll leave it there for today. I'm pleased that I've now got the emails being picked up and the data I need available. Tomorrow I can get on with extracting the data, creating a prompt for the LLM (e.g. OpenAI GPT), and passing in the data as part of the prompt to the LLM and see what its outputted.