Supporting LABBench2

[jamesbraza]

See PR title

[Copilot]

Pull request overview

This PR migrates the LABBench2 datasets (FIG_QA2 and TABLE_QA2) from the "futurehouse/litqa3" Hugging Face Hub to a new "futurehouse/labbench2" hub. The migration includes updating the dataset loading mechanism to use hf_name="all" instead of None, and updating test expectations to match the new dataset contents.

Changes:

• Renamed dataset constant from DEFAULT_LITQA3_HF_HUB_NAME to DEFAULT_LABBENCH2_HF_HUB_NAME and updated the hub path
• Modified dataset loading logic to use hf_name="all" for FIG_QA2 and TABLE_QA2 datasets, with appropriate type annotation updates
• Updated test questions and expected dataset size to match the new LABBench2 dataset content

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
packages/labbench/src/aviary/envs/labbench/task.py	Renames the Hugging Face Hub constant, updates the dataset loading logic to use hf_name="all" for LABBench2 datasets, and adds str to the type annotation
packages/labbench/tests/test_tasks.py	Updates test expectations with new question content from LABBench2 and adjusts FIG_QA2 dataset size from 100 to 101

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.