Skip to main content
Solved

Custom Transformer - Input attribute values can't be exposed nor used on the prompt


javierml
Contributor
Forum|alt.badge.img+1

Hello everyone

 

I've been designing a new custom transformer to sum up part of my workflow (in FME Form 2025.0). The issue is that I can't add any of my attributes to the text editor even though all attributes are selected to be exposed in the input parametres.

Moreso, If I edit it individually, no attributes appear (as obvious), but if I edit them inside a workbench session and I choose specific attributes to expose, they are still not availaible in the text editor, but they  get instead automatically added as user parametres for the custom transformer. Can anyone explain me how this works? I don't get a clear answer from looking at https://docs.safe.com/fme/2025.0/html/FME-Form-Documentation/FME-Form/Workbench/Using_Custom_Transformers.htm#Editing_Input_Output_Ports

It also irritates me that it’s theoretically a copy of another transformer with a couple of changes (only adding transformers or changing their parametres, none with the inputs) and that transformer allows to add feature attributes values by any prompt.

Best answer by desiree_at_safe

Hello ​@javierml!

 

I've taken a look at your FMX file, and I’m thinking it might be the AttributeManager causing your issue. I would suggest replacing your current AttributeManager with the original version from GoogleGeminiConnector as a workaround.

 

However, your existing AttributeManager needs to be deleted completely before pasting in the original version from GoogleGeminiConnector. Disabling and/or removing connections will not resolve the issue.

Issue Workaround

We're still investigating why you experienced this, as the differences between the two transformers aren't immediately obvious. May I ask if you made any modifications to the AttributeManager transformer while you were modifying the GoogleGeminiConnector, and in which versions of FME you were testing this

 

Please let me know if you have any luck! 🙂

 

View original
Did this help you find an answer to your question?

4 replies

desiree_at_safe
Safer
Forum|alt.badge.img+9

Hello! Based on your description, I suspect this might be related to how the custom transformer is configured for that specific parameter (“Prompt” in your example). Sometimes, when selecting “Attribute Value” doesn’t appear as an option, it could indicate that the parameter (“Prompt”) in the custom transformer might be expecting something else (i.e. if it’s set to a user parameter vs. attribute in the custom transformer).

If you can share your workspace, I could take a look to see how the “Prompt” parameter is set up in your custom transformer. It could give us a better understanding of what you are experiencing.

Looking forward to helping you solve this! 🙂

 


javierml
Contributor
Forum|alt.badge.img+1
  • Author
  • Contributor
  • April 14, 2025

Hi ​@d_mars,

 

I couldn’t exactly find what’s wrong with the prompt set up so I have attached the custom workbench.

 

As you can see, it’s just a workaround from the GoogleGeminiConnector, but instead of using a language-model, I’m trying to use an image generation model. That’s what I find strange because GoogleGeminiConnector doesn’t have this issue at all. 


desiree_at_safe
Safer
Forum|alt.badge.img+9

Hello ​@javierml!

 

I've taken a look at your FMX file, and I’m thinking it might be the AttributeManager causing your issue. I would suggest replacing your current AttributeManager with the original version from GoogleGeminiConnector as a workaround.

 

However, your existing AttributeManager needs to be deleted completely before pasting in the original version from GoogleGeminiConnector. Disabling and/or removing connections will not resolve the issue.

Issue Workaround

We're still investigating why you experienced this, as the differences between the two transformers aren't immediately obvious. May I ask if you made any modifications to the AttributeManager transformer while you were modifying the GoogleGeminiConnector, and in which versions of FME you were testing this

 

Please let me know if you have any luck! 🙂

 


javierml
Contributor
Forum|alt.badge.img+1
  • Author
  • Contributor
  • April 17, 2025

Many thanks ​@desiree_at_safe ! Now it’s solved!

 

I wouldn’t have thought of that in years. I know transformers can have different versions but I didn’t expect them to affect the custom transformer like this. I can’t really remember if I modified the attribute manager other than updating it to the current  version because I wasn’t sure if it’d have some compatibility issues in FME 2025.0 (which they happened, but the other way around).


Cookie policy

We use cookies to enhance and personalize your experience. If you accept you agree to our full cookie policy. Learn more about our cookies.

 
Cookie settings