ReportsTask

The ReportsTask orchestrates the generation of all final reports for consumption by end users and analytics systems. It coordinates three report subtasks that produce different views of the processed data.

Overview

Orchestrates 3 report generation subtasks
Writes results to database tables for consumption
Provides flexibility to skip or select specific reports
Ensures consistent reporting across all chains

Flow Diagram

Report Types

Constructor Parameters

Parameter	Type	Description
`name`	`Optional[str]`	Custom name for the task
`job_context`	`JobContext`	Context with Spark session and configuration
`database_sink`	`DatabaseSink`	Database connection for writing reports
`skip_reports`	`Optional[List[str]]`	Names of reports to skip (without “Task” suffix)
`only_run_reports`	`Optional[List[str]]`	Names of only reports to run (without “Task” suffix)
`write_to_catalog`	`Optional[bool]`	If true, write output tables using `DatabaseSink` (default: False)
`is_incremental`	`Optional[bool]`	If true, tasks that support incremental logic will run in incremental mode (default: None)

Note: ReportsTask forwards write_to_catalog and is_incremental to all subtasks when they are created, so you can control persistence and incremental mode centrally.

Subtasks

ReportsTask runs the following subtasks in order:

DailyReportTask - One row per stay date per booking
BookingRoomsReportTask - Aggregated data per reservation
GuestLoyaltyReportTask - Guest KPIs and loyalty flags
PickupReportTask - Pickup and pace analysis (not included in the default orchestrator; run separately if required)

Models

Requires

ProcessedGuestModel - Processed guest data
ProcessedReservationModel - Processed reservation data
ProcessedRoomModel - Processed room data

Provides

None (writes directly to database tables)

Incremental

The ReportsTask and a subset of its subtasks support incremental execution to avoid full regeneration of the output tables. The task-level requires_incremental is [ProcessedAddedRoomModel]. In the current implementation, both DailyReportTask and BookingRoomsReportTask allow incremental updates (_ALLOWS_INCREMENTAL = True) and will upsert only affected rows when is_incremental=True and write_to_catalog=True.

Report Output Tables

Report	Table Name	Granularity	Primary Use
Daily	`daily`	One row per room-stay-date	Detailed daily analysis, occupancy tracking
Booking Rooms	`booking_rooms`	One row per reservation	Revenue analysis, booking patterns
Guest loyalty	`guest_loyalty` (curated)	One row per guest	Guest segmentation and CRM use

Catalog / Database writing

Reports write to chain-scoped database tables via DatabaseSink. Whether a task writes to the DB depends on write_to_catalog. Use write_to_catalog=True to persist reports and False to run local-only transformations.

Database Sink

Reports use DatabaseSink to write results.

Tables are namespaced by chain ID:

athenaeum_daily
athenaeum_booking_rooms
mews_daily
mews_booking_rooms

DailyReportTask - Most detailed report
BookingRoomsReportTask - Reservation-level report
GuestLoyaltyReportTask - Guest-focused report
PickupReportTask - Snapshot-based analysis
Task - Base task class

Best Practices

Use only_run_reports for selective regeneration to save time
Monitor database sizes as reports can grow large
Consider partitioning large report tables by date
Test with skip_reports to identify slow reports
Document report dependencies for downstream consumers