feat(backend): New "update microagent prompt" API

[amanape]

• This change is worth documenting at https://docs.all-hands.dev/
• Include this change in the Release Notes. If checked, you must provide an end-user friendly description for your change below

End-user friendly description of the problem this fixes or functionality this introduces.

---

Summarize what the PR does, explaining any non-trivial design decisions.
Steps:

1. Get conversation ID and event ID of event we want the agent to learn
2. Get context surrounding given event ID
3. Generate a prompt with the help of an LLM to use as initial query

Blocks #8592

---

Link of any specific issues this addresses:

---

To run this PR locally, use the following command:

docker run -it --rm   -p 3000:3000   -v /var/run/docker.sock:/var/run/docker.sock   --add-host host.docker.internal:host-gateway   -e SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:edcbba1-nikolaik   --name openhands-app-edcbba1   docker.all-hands.dev/all-hands-ai/openhands:edcbba1

[enyst]

I see the point... Still, we shouldn't need the index in history, only event ids... We could apply a filter like this on get_events so that it gives us only the kind of events that can be part of an agent's history. Not enough, because we counted events, not agent history events...

OK, better, we can move to Tim's new interface to the event stream, search_events instead of get_events, because it also has a limit argument (and a nicer filter). So we can do something like

search_events (
  event_id=...,
  filter=EventFilter(...),
  limit=context_size
)

and then get those before event_id similarly, with reverse=True 🤔

[enyst]

Please let me know if you want me to tweak the events and the prompt. Probably on Monday if that's okay with you.

[amanape]

I've changed it to search events with the start and end indices, though I'm not sure if this is entirely correct given your comment on filter, limits, and reverse.

If you think there is a better way to do this, feel free! As for the prompt (if you mean the jinja file), yes, any improvements there would be appreciated as well 😄

[openhands-ai]

Looks like there are a few issues preventing this PR from being merged!

• GitHub Actions are failing:
  • Docker
  • Run Python Unit Tests

If you'd like me to help, just leave a comment, like

@OpenHands please fix the failing actions on PR #8357

Feel free to include any additional details that might help me get this PR into a better state.

_{^{You can manage your notification settings}}