feat: Add US state and Congressional district dataset support

[anth-volk]

Fixes #190

Summary

This PR adds support for US state-level and Congressional district-level simulations by:

• Adding utilities to construct GCS paths for state and district datasets (get_us_state_dataset_path, get_us_congressional_district_dataset_path)
• Integrating parse_gs_url() from policyengine-core for consistent GCS URL parsing with version support
• Creating a centralized .datasets/ directory for all downloaded datasets, supporting nested paths (e.g., .datasets/states/RI.h5, .datasets/districts/CA-01.h5)
• Refactoring _set_data() and _apply_region_to_simulation() for cleaner US region handling
• Replacing the deprecated SimplifiedGoogleStorageClient with VersionAwareStorageClient supporting both generation-based and metadata-based versioning
• Adding CI/CD workflow support for the 0.x maintenance branch

Important Note

Only the Rhode Island state-level dataset (states/RI.h5) currently exists in the proper GCS location. While this code correctly implements the feature, it won't run correctly in production for other states or Congressional districts until those datasets are uploaded to gs://policyengine-us-data/states/ and gs://policyengine-us-data/districts/.

Changes

New Files

• policyengine/utils/data/version_aware_storage_client.py - Storage client with dual versioning support
• tests/utils/data/test_version_aware_storage_client.py - Tests for the new client
• tests/utils/data/test_datasets.py - Tests for dataset path utilities

Modified Files

• policyengine/simulation.py - Refactored _set_data(), _set_data_from_gs(), and _apply_region_to_simulation() for US regions
• policyengine/utils/data/datasets.py - Added state/district path utilities, integrated parse_gs_url()
• policyengine/utils/data/caching_google_storage_client.py - Updated to use VersionAwareStorageClient
• policyengine/utils/google_cloud_bucket.py - Added .datasets/ directory support, renamed parameters for clarity (filepath → gcs_key, returns local_path)
• policyengine/utils/data_download.py - Updated parameter naming for clarity
• policyengine/utils/maps.py - Fixed download calls (was using invalid repo= parameter)
• policyengine/outputs/macro/comparison/calculate_economy_comparison.py - Updated download calls

Deleted Files

• policyengine/utils/data/simplified_google_storage_client.py - Replaced by VersionAwareStorageClient
• tests/utils/data/test_simplified_google_storage_client.py - Removed with deprecated client

Test plan

• Verify Rhode Island state simulation runs correctly with existing dataset
• Verify downloaded files are stored in .datasets/ directory
• Verify unit tests pass for new storage client and dataset utilities
• Verify other states/districts fail gracefully until datasets are uploaded

🤖 Generated with Claude Code