Text Load Line From File - Index is sampled when node is run not when prompt is queued.

[kendrick90]

Text Load Line From File - Index is sampled when node is run not when prompt is queued. In other words if I manually click through the indexes and queue the prompts one at a time, by the time the node is ran the index is set to the last one so I get lots of generations with the last index instead of what the index was at when I clicked queue prompt. If anyone has tips for how to queue prompts sequentially like this I would appreciate it.

[andreszs]

Your issue is rather confusing, considering that the index from the Load Line From File node may not auto-set itself. The index is an input, and your problem would be in node that is providing such value.

That being said, here's a sample workflow that loads lines from files sequentially. Each time the workflow is queued, the index receives a value increased by 1 compared to its previous value, allowing to get a prompt from each line of the text file provided.

To reset the counter and restart from the first line, you set the Logic Boolean node value to 1, queue a single workflow, and set it back to 0 for normal operation.