Result shapes¶

Every ortidy solver produces one of exactly three result shapes. Designing around these keeps the library a coherent whole rather than a bag of functions — once you know a solver’s shape, you know what its output looks like.

selection¶

The input table gains a selection / assignment column marking what was chosen — an item, a bin, or an edge in a tidy edge list.

Solver	Selection column
`knapsack`	`isIncluded` (boolean)
`multi_knapsack`	`binId` (assigned bin, or null)
`bin_packing`	`binId`
`assignment`	`selected` (boolean, on `(agent, task)` edges)
`generalized_assignment`	`selected` (on `(task, agent)` edges)
`facility_location`	`selected` (on `(customer, facility)` edges)
`set_cover`	`isSelected` (boolean, per subset)
`blend`	`quantity` (continuous, per item)

edge-flow¶

A numeric flow on an edge list. The solver returns the edges with a flow column.

Solver	Output
`max_flow`	edges + `flow`
`min_cost_flow`	edges + `flow`
`shortest_path`	edges + `onPath`
`transportation`	`(source, sink, cost)` edges + `quantity`
`solve_routing`	edge list of trips with route features

interval-schedule¶

Intervals placed on a timeline.

Solver	Output
`shift_scheduling`	one row per assigned `(workerId, day, shift)`
`job_shop`	tasks + `start` / `end`

Warning

If a new feature doesn’t fit one of these three shapes, that’s a design discussion before it’s an implementation.

The result object¶

Every solver returns a SolveResult:

result.frame       # native dataframe, same backend as the input (None if no solution)
result.status      # OPTIMAL / FEASIBLE / INFEASIBLE / UNBOUNDED / MODEL_INVALID
result.objective   # objective value (None when no solution)
result.metadata    # solver name, wall time, bounds, …
bool(result)       # True when OPTIMAL or FEASIBLE

A FEASIBLE solution is a success, not a failure — bool(result) and result.status.is_success both treat it as one. See the API reference for the full SolveResult and SolveStatus.