download-resource

name: download-resource description: | Download raw files (CSV, GeoJSON, Shapefile, PDF, GBFS) from Montréal's open data portal. Handles non-DataStore resources that require direct download. / Télécharger des fichiers bruts (CSV, GeoJSON, Shapefile, PDF, GBFS) du portail de données ouvertes de Montréal. triggers: - download, file, GeoJSON, shapefile, CSV, export, raw data - télécharger, fichier, exporter, données brutes

Download Resources / Télécharger des ressources

When to Download vs Query / Quand télécharger vs interroger

*DataStore has a 32,000 row limit per query. For large datasets, downloading the full CSV is often simpler.

Step 1: Find the Download URL

curl -s 'https://donnees.montreal.ca/api/3/action/package_show?id=DATASET_SLUG' | \
  python3 -c "
import json, sys
for r in json.load(sys.stdin)['result']['resources']:
    ds = '✓ DS' if r.get('datastore_active') else '✗'
    sz = r.get('size', 0)
    sz_mb = f'{sz/1048576:.1f}MB' if sz else '?'
    print(f'{r.get(\"format\",\"?\")} | {ds} | {sz_mb} | {r[\"url\"][:100]}')
"

Step 2: Download

# Simple download
curl -LO 'URL_FROM_ABOVE'

# Download with a specific filename
curl -L -o trees.csv 'https://donnees.montreal.ca/dataset/b89fd27d-4b49-461b-8e54-fa2b34a628c4/resource/64e28fe6-ef37-437a-972d-d1d3f1f7d891/download/arbres-publics.csv'

Always use -L (follow redirects). Some URLs redirect.

Format-Specific Handling

CSV

• Encoding: UTF-8
• Separator: comma (standard)
• Load with pandas: pd.read_csv('file.csv')

GeoJSON

• Standard GeoJSON (RFC 7946)
• Load with geopandas: gpd.read_file('file.geojson')
• Or parse with json.load()

Shapefile (ZIP)

• Downloaded as .zip containing .shp, .dbf, .prj, .shx
• Unzip first, then load with geopandas: gpd.read_file('file.shp')
• Coordinate system is usually NAD 83 MTM Zone 8 (EPSG:32188)

GBFS (BIXI)

• Not a downloadable file — it's a live REST API
• Entry point: https://gbfs.velobixi.com/gbfs/gbfs.json
• Returns URLs for station info, station status, system info

# Get station status (live availability)
curl -s 'https://gbfs.velobixi.com/gbfs/en/station_status.json'
# Get station info (locations, capacity)
curl -s 'https://gbfs.velobixi.com/gbfs/en/station_information.json'

GTFS / GTFS-Realtime (STM)

• GTFS static: ZIP file with stops.txt, routes.txt, trips.txt, etc.
• GTFS-Realtime: Protocol Buffer format (requires gtfs-realtime-bindings library)
• STM developer registration may be required for real-time feeds
• See the transit domain skill for details

PDF

• Documentation, data dictionaries, methodology descriptions
• Not machine-queryable — use for reference only

Large File Considerations

Some datasets are substantial:

• Tree inventory CSV: ~135 MB (333,556 rows)
• Road condition data: large GeoJSON files
• Historical BIXI trips: multi-GB across years

For large files:

1. Download once and cache locally
2. Use head or pandas.read_csv(nrows=100) to preview before full load
3. For analysis, consider loading into a local SQLite database

Provenance / Provenance

Note on non-CKAN downloads: BIXI GBFS and STM GTFS feeds are hosted on separate infrastructure. See the transit domain skill for their specific URLs, auth requirements, and publishers.

Related Skills / Compétences connexes

• understand-ckan — API reference and URL patterns
• discover-datasets — Find the right dataset first
• query-dataset — Query without downloading (DataStore API)