Many additive output fields (no existing keys renamed or removed):
- Discovery/chaining:
city_id, place_id.
- Timing:
start_time_formatted, current_start_timestamp (next occurrence), end_time (often null), days_until_event (derived), is_happening_now.
- Classification:
category, event_kind, rsvp_style.
- Tickets/price:
ticket_source, price, price_info, price_range_text.
- Attendance:
attendance_facepile_count, attendance_tier (small/medium/large/mega, derived), interest_to_going_ratio (derived).
- Hosts:
host_context_text, host_names, creator, and description_links.
Skipped on purpose: is_recurring (no reliable signal in the API), country_code (no source field), end_utc (no end timestamp). Attendee identities are never returned.
Additive output fields (no existing keys renamed or removed):
start_utc - event start as an ISO 8601 UTC string, derived from start_timestamp. A standard join key for combining datasets.
responded_count - total responded (interested + going), for parity with other event scrapers.
Also in this release:
- City-based event discovery via
cityUrls (a city's Facebook Events URL) with an optional cityTime filter (today / this week / next week). The most reliable way to discover events.
- Auto-detect inputs: fill any combination of city URLs, page/organizer URLs, event URLs, or keywords - the Actor scrapes everything provided. The
mode and fetchFullDetails inputs were retired (full details are always fetched); old task inputs that still include them keep working.
maximumResults default is now 0 = all events.
- Cleaner run logs (one summary line per source).
Note: Facebook's keyword event search is currently unreliable upstream and may return no results; use cityUrls for dependable discovery.