Disarmament Corpus: A Structured Full-Text Dataset of 15 International Arms Control and Disarmament Treaties (1925–2017)

A structured, article-level corpus of 15 core international disarmament and arms control treaties spanning 1925–2017, prepared for natural language processing, legal text analysis, and computational social science research. Coverage: 1,220 units / ~204,000 words across 15 treaties including the Geneva Protocol (1925), NPT (1968), BWC (1972), CWC (1993), Ottawa Treaty (1997), and TPNW (2017). Structure: Each unit carries fields for treaty_id, year, category, unit_type (PREAMBLE, ARTICLE, PARAGRAPH, SECTION, PROTOCOL), unit_id, title, and text. All text is in English. Sources: Official treaty texts from UNODA, OPCW, IAEA, OPANAL, and the UN Treaty Collection. Text extracted from official PDFs using pdfplumber and pytesseract (Python).Files: corpus_all_units.json (all units), corpus_index.csv (treaty-level metadata), texts/ (raw .txt files), structured/ (per-treaty JSON).