Change to event handling

[TommasoGhilardi]

@FrancescPoli
I’d like to propose we adopt a DeToX-style workflow for handling triggers and data collection with Tobii Research. It’ll involve a shift from our current global-based pipeline, but I think it’ll simplify the code and give us more flexible export options.

We could use two in-memory lists

• Data samples (raw eye-tracking records)
• Trigger events

Then we can:

• Convert each list into its own pandas DataFrame
• Perform a nearest-timestamp merge so triggers align with samples

Why this helps

• Cleaner code: removes that hard-to-manage global, reducing hidden side effects and making testing easier.
• Dual export modes:
  • Merged DataFrame—each trigger matched to its closest sample (ideal for most analyses)
  • Separate DataFrames—raw samples and raw triggers preserved for anyone needing millisecond-level accuracy (we can document this under “Advanced Usage”)

This does alter our existing pipeline, but it keeps things modular and lets users choose between convenience (one merged table) and precision (two raw tables). I’d love to hear your thoughts or any concerns before I build out a prototype!

[FrancescPoli]

I like this - I think the dual export mode is key, because many people want to have everything in one dataframe. One option here might be whether the triggers are on each row or only at the moment they are sent.

[TommasoGhilardi]

I think this is ready here branch.

However given our conversation about Psychopy i will put this on hold as I am not sure if this is worth changing if we will change again in the near future.

[TommasoGhilardi]

i actually rewrote it. I believe now is better and more reliable. However it may be a little more complicated.. @FrancescPoli what do you think?? is it readable to you here the branch???