Hey guys, I can’t figure out how it’s possible that the output from one module has 18 bindles, and the input of the following module only gets 1?!
Please check the video, and I will appreciate any help you can provide
Hey guys, I can’t figure out how it’s possible that the output from one module has 18 bindles, and the input of the following module only gets 1?!
Please check the video, and I will appreciate any help you can provide
Hey Carlos,
1st - you don’t need an iterator after a List module. The List module is already returning a separate bundle for each item found and not an array.
2nd - you have the entire output bundle mapped in the iterator and not an array. So even if it was needed, its still wrong.
3rd - again you are mapping the entire output bundle from the iterator and not the specific item you need. So even if the iterator was doing something, the next module will still not work.
Delete your iterators and map the items you need directly, not the module output bundles.
Thanks for your response @Stoyan_Vatov , it makes sense, but it doesn’t work. I need all the folders to be sent, so the system can get all the lists and then retrieve all the tasks. Your solution retreives only the first folder, please check the video.
Any ideas?
Ok the field you are trying to map inside expects a singular folder ID. What you are trying to map is the entire output of the module, not folder IDs. This includes headers and background information and api keys and everything.
Yes the List all Lists module will do only one folder, because it works one folder at a time. Then the List all Tasks will list the first list. When its done go to the second and so on, then when the lists are done it will go to the next folder and keep repeating until everything is done.
What you need is to go back to a single bundle after the tasks are listed. What are you trying to send to chatGPT? Most likely you will need a Text Aggregator module with the List all Folders as the source module. And inside you will build the text string you need to send to chatGPT for analysis.
Hey @Stoyan_Vatov thanks for your response, I figured a better way that this initial way suggested by ChatGPT. I still seem to face a similar issue, and that is, that even though there are 9 bundles going out from module one, chat GPT only processes the first one…
You still need to aggregate the response. Most likely with a Text Aggregator so you can send a string to chatGPT. And if you want all the tasks, then you still need the list all lists and list all tasks modules.
The List all Folders module only returns folder level information, it doesnt go deeper.
Thanks brother, will try and report
Are you sending only the task names to chatGPT? Or does it need more? Descriptions, dates?
I agregated the answers and the scenario seems to work properly, but chatGPT sent me an empty response, so I guess the issue now is in the configuration of ChatHPT or what I am feeding from the aggregator, is it the ‘array” that I need to feed chatGPT?
Can you show a screenshot of what you are sending now and what the response looks like?
Here’s a breakdown of all the setup:
Why is the Role for Message 1 set as “System” instead of “User”? You probably don’t even need more than 1 message to ChatGPT if you’re not simulating a previous conversation for context.
Also, as @Stoyan_Vatov asked, do you really need all fields from Clickup in order to generate a summary of Tasks? This seems to be overkill and might make you pay for more credits than you need.
Lastly, make sure the Clickup module you chose is giving you the correct response. You stated that you don’t know why the result was reduced from 18 to 9 bundles.
Great points @damato you are right, I now filtered only the active tasks to come out from the ClickUp module, I deleted the user in GPT and shifted system to user and added only a set of instructions there.
You are also right. GPT is using an enormous amount of tokens, and it won’t be sustainable, even though I simplified the prompt. So I hope that by sending less data into GPT and removing one of the messages, it will consume fewer tokens…
GPT only needs to receive active tasks and then write an email organizing the tasks per user/assignee.
The result was reduced from 18 to 9 bundles, because before I was getting all the customer folders (18) and then I realized I could do ok with just getting the total of active tasks (9), so I adjusted the scenario to this:
Great to know!
Is it working as expected now?
yes apparently so! Thank you ![]()