Skip to content

Improve coordinate table gating and export - #65

Merged
adelavega merged 1 commit into
masterfrom
no_miss
Jul 31, 2026
Merged

Improve coordinate table gating and export#65
adelavega merged 1 commit into
masterfrom
no_miss

Conversation

@adelavega

Copy link
Copy Markdown
Contributor

Summary

Improve ACE’s coordinate-table gating and expose a reusable extraction API for Autonima.

Changes

  • Add public ace.ingest.extract_and_export(...) API using a temporary SQLite database.
  • Separate coordinate-table gating from deterministic coordinate parsing:
    • retain candidate tables in tables.csv even when ACE extracts zero coordinates;
    • keep coordinates.csv limited to successfully parsed coordinates.
  • Improve candidate detection for:
    • hierarchical x/y/z headers;
    • combined coordinate columns;
    • Unicode minus signs;
    • malformed rows;
    • captions and footnotes.
  • Isolate innermost tables and deduplicate equivalent inline and linked tables.
  • Improve handling for PMC, ScienceDirect, Springer, and Journal of Cognitive Neuroscience layouts.
  • Recover legacy JCN tables exposed through JavaScript placeholders using DOI-linked browser retrieval.
  • Correct exported table-footnote metadata.
  • Expose ACE’s version for downstream cache invalidation.

Validation

  • Full ACE test suite: 44 passed
  • Added regression coverage for:
    • candidate tables with zero deterministic coordinates;
    • combined and Unicode coordinates;
    • nested and duplicate tables;
    • public extraction/export;
    • legacy JCN linked-table expansion.

@adelavega
adelavega merged commit 9f2af57 into master Jul 31, 2026
1 check failed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant