| # OpenGov.gr Diaboyleuseis Dataset | |
| This dataset contains the AI4Deliberation SQLite database with public consultation data from the Greek government's opengov.gr platform. The primary file is `deliberation_data_gr_MIGRATED_FRESH_20250602170747.db` (Git LFS). | |
| Last update: 2025-06-02. | |
| ## Database Schema | |
| The database contains the following tables (current record counts): | |
| ### Core Tables | |
| - **ministries**: Government ministries (37 records) | |
| - **consultations**: Public consultations (1,104 records) | |
| - **articles**: Consultation articles/documents (28,942 records) | |
| - **comments**: Public comments on consultations (125,488 records) | |
| - **documents**: Supporting documents (2,164 records) | |
| ### Legal Reference Tables (currently empty) | |
| - **nomoi**: Greek laws | |
| - **ypourgikes_apofaseis**: Ministerial decisions | |
| - **proedrika_diatagmata**: Presidential decrees | |
| - **eu_regulations**: EU regulations | |
| - **eu_directives**: EU directives | |
| ## Usage | |
| Use the SQLite database directly: | |
| ```bash | |
| sqlite3 deliberation_data_gr_MIGRATED_FRESH_20250602170747.db "SELECT COUNT(*) FROM consultations;" | |
| ``` | |
| If you need PostgreSQL, export from SQLite using your preferred tool (e.g., `sqlite3 .dump` → `psql`). | |
| ## Data Fields | |
| ### Key fields include: | |
| - **Text content**: Raw HTML, cleaned content, extraction metadata | |
| - **Language detection**: Greek/English percentage scores | |
| - **Quality metrics**: Badness scores for content quality | |
| - **Timestamps**: Creation and update times | |
| - **Relationships**: Foreign keys linking consultations, articles, comments | |
| ## Updating the dataset | |
| The scraping and update pipeline lives in the companion repository: https://github.com/eellak/AI4Deliberation. | |
| 1. Clone the repo and create a Python venv. | |
| 2. Install requirements relevant to the scraper: `pip install -r ai4deliberation_pipeline/requirements.txt` (at minimum: `sqlalchemy`, `requests`, `beautifulsoup4`, `markdownify`, `PyYAML`). | |
| 3. Update the database in place: | |
| ```bash | |
| python -m ai4deliberation_pipeline.scraper.run_update \ | |
| --db-path sqlite:////full/path/to/deliberation_data_gr_MIGRATED_FRESH_20250602170747.db \ | |
| --batch-size 5 | |
| ``` | |
| Finished consultations are left untouched; new consultations and new comments/documents on unfinished ones are added. | |
| 4. Commit and push the updated `.db` file to this dataset repo (Git LFS handles the binary). | |
| ## Source | |
| Generated from opengov.gr using the AI4Deliberation pipeline. | |
| ## License | |
| This dataset contains public government data from opengov.gr platform. | |