org.gov2.ro

Explorer: https://org.gov2.ro (publicat cu sh scripts/publish_pages.sh).

Fă o listă a instituților publice din România. Adunăm cât mai multe detalii.

Va fi și input data (cu site-urile publice) pentru gov2-ro/stiri

Roadmap

MVP

• scrape roepas
• look at alternate sources
• consolidate data
• integrate ANI "Declarații integritate" (declaration counts, state-owned companies)
• re-export the two capped ANI county groups (full tail of rare names)
• resolve ANI-only records (state companies) to a CUI via the ONRC open data

Pipeline

uv run scripts/consolidate.py          # -> data/consolidated/ (institutii.csv/json, metadata.json, rows.json)
python3 -m http.server                 # then open http://localhost:8000/explorer/

Sources it reads: the ANAF list (backbone, docs/reference/), the ROePAS scrape (data/roepas/, scripts/scrape_roepas.py), the Registrul PDF extraction (docs/reference/) and the ANI exports (data/declaratii-integritate/).

ANI exports (data/declaratii-integritate/)

Kibana Lens → "Top values of institutie_id.keyword" + "Count of records", filtered on a group of counties, exported as CSV (Lens Visualization [...].csv). Kibana keeps at most 10,000 values per export, so:

• keep each export under 10,000 rows: split a county group until it is (8 counties at a time is a safe start); the script warns about any file that hit the cap;
• re-exporting a group in parts is fine: a value present in several files is counted once, so the old capped file can stay until the new ones are in.

Alt sources

• have a look at Lista_EP_portal_01102026 - sursa
• gov.ro
• more lists in gDrive to consult later

---

Copyright (c) 2026 gov2.ro Date reutilizabile conform EUPL v1.2 (the "Licence")