feat(eval): SWE-Bench stability improvement and add utils

[xingyaoww]

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

• Include this change in the Release Notes. If checked, you must provide an end-user friendly description for your change below

---

Give a summary of what the PR does, explaining any non-trivial design decisions

• move SWE-Bench eval_infer.py runtime creation inside try-except, so we can properly cleanup in case of error
• add combine completion util scripts useful for SWE-Bench evaluation/SWE-Gym rollout
• improve memory efficiency for SWE-Bench update output script to support loading larger files

---

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:ffac44a-nikolaik   --name openhands-app-ffac44a   docker.all-hands.dev/all-hands-ai/openhands:ffac44a

[neubig]

Basically LGTM

[xingyaoww]

@neubig looks like most of these changes are already merged into main in another PR 😅 -- just leaving one line bug fix

[xingyaoww]

We merged changes from other PRs 😓 Let's close this one!