Managing Your Data

Database Structure

datannur uses a client-side relational database powered by jsonjsdb. Your metadata must be structured as a relational database with specific requirements:

• Database location: By default in /data/db/ folder (see path for customization options)
• Tables: Each table is stored in two files (.json and .json.js)
• Table registry: __table__.json file lists all available tables
• Primary keys: Must be a column named id
• Foreign keys: Columns named after the foreign table with _id suffix (e.g., dataset_id)
• Many-to-many relationships: Two approaches available:
  • Array of IDs (recommended): Use _ids suffix with comma-separated values (e.g., tag_ids: "1,3,7")
  • Junction tables: Use underscore notation (e.g., dataset_tag table)

File Format Specifications

Each table is stored in two formats:

.json files - Standard JSON format:

[
  {
    "id": 1,
    "name": "Example item",
    "description": "Item description"
  },
  {
    "id": 2,
    "name": "Another item",
    "description": "Another description"
  }
]

.json.js files - Compact format (auto-generated, optimized for browser):

jsonjs.data['dataset'] = [
  ['id', 'name', 'description'],
  [1, 'Example item', 'Item description'],
  [2, 'Another item', 'Another description'],
]

💡 Note: The demo database is generated from /data/db-source/ with python3 datannur.py build-db-source. Edit the source files there, then rebuild /data/db/; the .json.js files are derived for optimal browser performance.

Table Registry

The __table__.json file serves as a registry of all available tables in your database:

[
  {
    "name": "dataset",
    "last_modif": 1753608552
  },
  {
    "name": "folder",
    "last_modif": 1757018090
  },
  {
    "name": "__table__",
    "last_modif": 1757018100
  }
]

Key features:

• name: Table name (must match the corresponding .json and .json.js filenames)
• last_modif: Unix timestamp of last modification (in seconds), used for caching optimization
• Special entry: The "__table__" entry tracks the overall metadata update time, displayed in the catalog interface

Data Schema Overview

The catalog supports several entities with flexible relationships. All tables are optional - use only what you need:

📋 Schema Reference: For complete schema details, see the metadata page at index.html#/meta in your catalog, which displays all tables and variables based on the current data structure. The "localisation" column indicates whether each table/variable exists only in schema, only in data, or both (when empty).

🔗 Entity Structure: For information about entities and their relationships, see the about page at index.html#/about?tab=aboutStructure in your catalog.

Geographic metadata

Datasets can carry optional geographic metadata:

• bbox: bounding box as an array of four numbers [west, south, east, north] in WGS84 (lon/lat)
• crs: coordinate reference system, e.g. "EPSG:2056"
• geometry_type: point, linestring, polygon, … (vector datasets)
• spatial_resolution: spatial resolution in metres (raster datasets)

When present, the catalog shows a "Geo" column in lists and a coverage map on the dataset (and folder) page, and these fields drive the geospatial exports (DCAT/GeoDCAT-AP, STAC, ISO 19139). A folder's coverage is the union of its datasets' bounding boxes.

For variables, two type values support geodata: geometry (the geometry column of a vector dataset) and band (one variable per band of a raster dataset, with its pixel statistics).

Configuration Options

The config.json file allows you to customize various application settings:

[
  {
    "id": "contact_email",
    "value": "contact@yourdomain.com"
  },
  {
    "id": "banner",
    "value": "![main-banner no_caption](data/img/main-banner.png)"
  }
]

Available options:

• contact_email: Contact email displayed in the catalog interface

Global Filter Rules

The configFilter.json file defines global database filters displayed in the application header. Each rule can target any table and field:

[
  {
    "id": "open_data",
    "name": "Open Data",
    "entity": "dataset",
    "field": "type",
    "value": "open_data",
    "is_active_default": true
  }
]

Use one row per matched value. When a filter is disabled by the user, matching rows are removed from the in-memory database and related rows are removed through jsonjsdb relations.

About page/tab customization:

The "About" content (both homepage tab and dedicated page) is composed of three sections: banner + body + more_info. Each can be customized independently using Markdown.

• banner: Custom main banner image
  • Add no_caption to hide image caption
  • Add {darkMode} in the filename (main-banner{darkMode}). This will show main-banner.png in light mode and main-banner-dark.png in dark mode
• body: Custom main content
• more_info: Custom additional information

Source files to jsonjsdb format

• Maintain editable source files in /data/db-source/:
  • top-level *.json files for metadata tables
  • md/*.md files for Markdown documents
  • dataset/*.csv files for dataset previews
• Use python3 datannur.py build-db-source to compile source files to both .json and .json.js formats in /data/db/
• Run from the application folder with:

  python3 datannur.py build-db-source

• Re-run the script after changing source files to update the generated database.