ReGKG-2025

ReGKG-2025 is an archive-derived reconstruction sample for the missing GDELT Global Knowledge Graph (GKG) interval from 14 June 2025, 18:00 UTC, through 2 July 2025, 02:00 UTC, inclusive. Article discovery and text retrieval used existing Internet Archive Wayback Machine and Common Crawl captures. Records follow the 27-column GKG 2.1 layout, with text-derived fields generated using gdelt-regkg package. The package includes record and archive provenance, source-level summaries, source/day calibration weights, a manual-review journal, processing configurations, documentation, and integrity checksums.