Vend PolicyEngine bundle 4.16.0: populace-us as the certified US default#399
Open
MaxGhenis wants to merge 2 commits into
Open
Vend PolicyEngine bundle 4.16.0: populace-us as the certified US default#399MaxGhenis wants to merge 2 commits into
MaxGhenis wants to merge 2 commits into
Conversation
… release The eCPS-free populace dataset (populace-us-2024-5da5a95-20260611, populace-data 0.1.0) becomes the certified default for country 'us' — it beats the enhanced CPS on train (0.176 vs 1.089), holdout (0.037 vs 0.317), and full-surface (0.213 vs 1.406) loss in the matched symmetric-refit comparison, with parity 0 and bounded weights.
Artifact path references now carry optional repo_id/repo_type pins so inherited datasets (long-term, state, district) keep resolving from their original repos while populace-data is the data package; dataset and region resolution prefer those pins. Release manifest URLs are repo-type-aware (datasets/ prefix for HF dataset repos). The vendored bundle is re-imported with build provenance, region dataset templates, and policyengine-core 3.27.1 (3.27.0's computation-mode check crashed update_variable reforms; fixed upstream in core#501). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Vends bundle 4.16.0 from policyengine-bundles: populace-us becomes the certified US data release (
populace-data 0.1.0, default datasetpopulace_us_2024, pinned by immutable tag + sha256, 571 artifacts including inherited long-term/sub-national datasets).The dataset is built entirely from primary sources (CPS ASEC, IRS PUF, Fed SCF 2022, SIPP, CPS-ORG, MEPS-IC, ACS 2022 — the enhanced CPS is benchmark-only) and beats the enhanced CPS on the sound comparison at matched 41,314 households: train 0.176 vs 1.089, holdout 0.037 vs 0.317, full 0.213 vs 1.406, with parity 0 gaps and zero super-weighted records. Full evidence: https://populace.dev
Supporting changes:
ArtifactPathReferencegains optionalrepo_id/repo_typeand the bundle importer + dataset resolver honor them, so inherited artifacts (e.g. long-term CRFB datasets pinned topolicyengine-us-data@crfb-longrun-20260517) keep resolving from their original repos whilepopulace-datais the data package.https_release_manifest_uriis repo-type-aware (datasets/URL prefix for HF dataset repos).data_build_id,built_with_model_version) flows from the release manifest through bundle compatibility metadata into the vendored country manifest (bundles #31).Generated with the bundles repo's own packaging + import scripts (validate exit 0 end to end).
🤖 Generated with Claude Code