Schema: optional artifacts + DL-HARD dataset entry#20
Merged
Conversation
`schema.json`: `artifacts.run_file` and `artifacts.reformulated_queries` become optional in schema v1; their filename regex patterns still apply when the fields are present. A run is now valid whether it carries both artifacts, one, or neither. `emit.py`: `build_run_summary` accepts a new `artifacts_present` parameter selecting which artifact fields to include. Default is both, preserving the existing contract for runs produced by the standard pipeline. `submit_run.py`: artifact files are required only when the payload's `artifacts` block references them, and silently skipped when omitted. Tests cover the optional path (empty `artifacts`, only `run_file`, only `reformulated_queries`) and assert the existing required path still works. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
DL-HARD is the hard subset of MS MARCO passage queries. Uses the prebuilt `msmarco-v1-passage` Pyserini index with default BM25 weights (k1=0.9, b=0.4). Eval metrics: `ndcg_cut.10`, `recall.1000`. Co-Authored-By: Claude Opus 4.7 (1M context) <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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Two small additions:
Optional artifacts in schema v1.
artifacts.run_fileandartifacts.reformulated_queriesbecome optional. Their filename regex patterns still apply when the fields are present, so a run remains valid whether it carries both, one, or neither artifact. This lets a run be published from just its scores when the per-run TREC ranking or queries TSV isn't attached.reproducibility/schema.json—artifacts.requiredremoved.reproducibility/lib/emit.py—build_run_summarytakes a newartifacts_presentparameter to control which artifact fields are emitted (defaults to both, unchanged for the standard pipeline).reproducibility/scripts/submit_run.py— artifact files are required only when referenced by the payload, and silently skipped when omitted.msmarco-v1-passage.dlharddataset entry. Hard subset of MS MARCO passage queries. Uses the prebuiltmsmarco-v1-passageindex with default BM25 weights (k1=0.9, b=0.4); eval metricsndcg_cut.10,recall.1000.Test Plan
make repro-test— 44 passed (36 prior + 8 new covering the optional artifact paths and the DL-HARD entry).make repro-check— clean.reproducibility-check.ymlgreen on PR.🤖 Generated with Claude Code