如何实现对 Array<File> 类型的数据进行 chunk 切分

Discussion
1

My scenario is that in a flow, I use a set of images for the model to understand, but the model’s provider has a limit on the number of images per request—say, 5. So I want to chunk this set of images into groups of 5 and call the model in parallel for each group.

I’d like to ask the community if anyone has any ideas for this scenario.

  • The List Operator doesn’t support chunk operations.

  • The Code node can operate on arrays, but the output type of the node cannot be set to File, so the type mismatch prevents its use in subsequent nodes.

2

Just had a sudden idea to implement it using a Loop nested within an Iteration:

Manually calculate the number of chunks, start a Loop, inside the Loop use an index variable to compute a range, then perform an iteration over the original list, but with an if condition currentIndex in range to filter. (This effectively acts like list.filter((_, idx) => idx in range).)

But didn’t expect that a Loop can’t contain an Iteration inside it :joy:

3

Well, although Loop can’t wrap Iteration, it can wrap a workflow, and the workflow can then use iteration.