Structured UAP Case Database and Full-Text Corpus: PURSUE Release 1 (776 cases, 117 documents)

This dataset accompanies the paper: Pavón, M. (2026). Systematic Analysis of Declassified U.S. Government Records on Unidentified Aerial Phenomena: FBI Case File 62-HQ-83894, USAF Project Blue Book, and the Full PURSUE Release 1 Corpus. It contains a structured SQLite case database (776 incident cases, 29 fields per case), machine-readable full-text files for all 117 source PDFs (≈4,156 pages), and the Python scripts used for extraction and analysis. Source documents were released by the U.S. Department of War on May 8, 2026, through the PURSUE declassification portal (war.gov/UFO) and are not redistributed here. Corpus breakdown by agency: FBI 62-HQ-83894 (422 cases, 1944–1967), US Navy RG 38 (217 cases, 1945–1949), DOD/DOW mission reports (44 cases, 2016–2025), NASA transcripts (29 cases, 1954–1995), USAF/Blue Book additional (29 cases), US Army/USAAF (17 cases), DOD/AARO and DOS (7 cases), other (11 cases). Extraction pipeline: three-tier approach combining pre-extracted text, pypdf (digitally-native PDFs), and Reducto AI document parsing (scanned PDFs). Structured case extraction via Reducto AI extract_data with a 17-field JSON schema. The complete prompts and schema are reproduced in Supplementary Material A of the paper. Contents: pursue_reducto_cases.db (SQLite, queryable directly), full_text.zip (117 UTF-8 text files), scripts/ (4 Python scripts for extraction, integration, and figure generation).