My workflow involves the user uploading images (Media Uploader), then performing a prediction on those images. First, I need to process those images, ideally through Pipeline Builder. The AI suggests “Configuring the action to call a function or service that initiates the pipeline run with the uploaded media as input.” I’m not sure what this refers to, and clarification would be much appreciated. What is a function in Workshop? Can they help me?
Assuming you have the media uploader widget configured to save the media item in a media set, you can import that media set into pipeline builder immediately. This thread from the community should give you a comprehensive answer to what you can do with a media set
https://community.palantir.com/t/how-do-i-access-attachments-from-objects-in-a-pipeline/2152/4
To directly answer the what is a function question, it’s a mechanism for writing and executing arbitrary code in the ontology context. You can learn more about them here https://www.palantir.com/docs/foundry/functions/overview
A function in Workshop is essentially a reusable piece of logic or code that can be executed to perform specific tasks (like George said), like image processing or triggering pipelines. When the AI suggests configuring an action to call a function, it means setting up a workflow step that automatically triggers this pre-defined logic once an image is uploaded. Functions can help streamline your process by automating tasks like initiating pipeline runs with uploaded media as input, reducing manual effort and ensuring consistency.
If you’re not sure how to create or configure a function, Workshop provides a low-code environment where you can define these functions visually or through simple scripting. So, this makes it easier to build and adjust workflows without deep programming knowledge.
Hope this helps!
Thanks for the reply! Just to clarify, my intended workflow does not include manually feeding images into a Pipeline; this is done behind the scenes, automatically, as soon as the user uploads the image. Is this what you had in mind?
Got it, yes, with the Media Uploader widget configured to write the files into a media set, it automatically uploads the media item into the media set, and then your pipeline can be configured to run on a schedule.
Alright, got it. Here’s the game plan:
- Upload images through Media Uploader
- Media Uploader connects to Media Set
- Media Set connects to Streaming Pipeline (I assume, since that’s the only way to get my changes updated in real-time)
With this, I hope to process my data for ML modeling. I don’t know if I’m doing something wrong, there are no built-in functions for Media Sets in a Streaming Pipeline, as shown below:
My guess is that I’m meant to use a code repository for the processing, and call that repository via custom functions in Pipeline Builder. I’m unable to find adequate sources for a workflow like this, so I’d just like to confirm with you.