Changelog
Monthly project milestones of the BUDOVA corpus and platform. The first public data release is v1.0, planned for Q4 2026.
- v0.10Oct 2026Current
Eighth building code annotated and added to the platform on 12 October: ДБН В.1.1-25:2009 (engineering protection of territories and structures from flooding and inundation), 328 segments. Ninth on 18 October: ДБН В.1.2-7:2021 (basic requirements for buildings and structures — fire safety), 107 segments — 2,927 segments on the platform across nine codes (viewable in the corpus browser). Annotation guidelines v0.22 (118 rules). On 13 October, read-aloud speech collection opened: contributors read short construction texts aloud in the browser, can edit a word to say it the way people do where they come from, and fill in a speaker profile. A new public Voices page shows recordings selected by moderators.
- v0.9Sep 2026
Seventh building code annotated and added to the platform: ДБН В.2.6-220:2017 (roofs of buildings and structures), 559 segments — 2,492 segments on the platform across seven codes (viewable in the corpus browser). Annotation guidelines v0.20 (114 rules). The working text collection passes 49 million raw tokens (not yet annotated or released) from building codes, construction legislation, estimating norms, open-access journals and other open sources; the first public sample release is planned for Q4 2026.
- v0.8Aug 2026
Three more building codes annotated and added to the platform: ДБН В.1.1-45:2017 (buildings and structures in difficult ground conditions, 367 segments), ДБН В.1.2-12:2026 (infill development, 133) and ДБН Б.2.2-1:2008 (cemeteries, crematoria and columbaria, 241). Guidelines v0.15–v0.19; an audit of unanimous records added to quality control.
- v0.7Jul 2026
Text-corpus pipeline: a catalog of construction normative sources, document ingestion, a fragment quality standard with an automated linter, pre-annotation suggestions and a notation glossary in the annotator. First building codes annotated and added to the platform: ДБН В.2.6-98:2009 (concrete and reinforced-concrete structures, 904 segments), ДБН Б.2.2-6:2013 (sanitary cleaning of settlements, 199) and ДБН В.1.2-8:2021 (hygiene, health and environmental protection, 89). Ninth entity label
load_effect; guidelines v0.3–v0.14. - v0.6Jun 2026
Quality and compliance. Contributor consent with e-signature and an admin consent register (GDPR); inter-annotator agreement v2 with entity-window κ; annotation guidelines v0.2; keyboard-accessible span creation and a contrast audit gate; revocable sessions; deterministic, adjudication-resolved corpus export; structured logging, tracing and service alerts.
- v0.5May 2026
Speech collection readiness: an audio-quality pipeline with signal-to-noise checks, a live recording level meter, recording templates and a documented field recording protocol.
- v0.4Apr 2026
Platform hardening and a public-site expansion. Password reset, inter-annotator agreement aggregation endpoint, coverage/provenance APIs, redesigned auth pages, email delivery via ACS; public guidelines rulebook, coverage map, sitemap and home-page teasers.
- v0.3Apr 2026
Azure deployment complete: custom domain
budov.org, nightly Postgres backups, Application Insights, Playwright smoke tests in CI/CD, PostHog product analytics. - v0.2Mar 2026
Annotation platform v1: task creation, NER annotator with span editing, skipped-task unskip flow, admin panel, speech upload + in-browser recording, lexicon editor.
- v0.1Feb 2026
Project bootstrap. Hero landing, docs structure, LINGUA grant received through Microsoft AI for Good Lab. Initial seed data.
- v1.0Q4 2026Planned
First public release: a text-corpus sample with a datasheet and the annotation guidelines on Hugging Face and Zenodo. Speech data and the baseline NER model follow in later releases.
