Core app & interfaces

HOTS – application main process.

class hots.core.app.App(config: AppConfig)

Bases: object

Application entry point for HOTS.

pre_loop()

Build and solve optimization models before performing streaming loop.

run()

Run the initial evaluation and streaming update loop.

shutdown(signum=None, frame=None)

Handle shutdown signals gracefully.

HOTS core interfaces: plugin base classes.

class hots.core.interfaces.ClusteringPlugin

Bases: ABC

Interface for clustering plugins.

abstract fit(df: DataFrame) → Series

Perform clustering and return labels.

class hots.core.interfaces.ConnectorPlugin

Bases: ABC

Interface for connector plugins.

abstract apply_moves(moves: Any) → None

Apply relocation moves to the target environment.

class hots.core.interfaces.OptimizationPlugin

Bases: ABC

Interface for optimization backends (Pyomo, OR-Tools, …).

abstract build(*, u_mat=None, w_mat=None, v_mat=None, dv_mat=None)

Create and return a model handle (e.g., a Pyomo model wrapper).

abstract solve(*, solver=None)

Solve the problem using current data.

class hots.core.interfaces.ProblemPlugin

Bases: ABC

Interface for domain‑specific problem plugins (placement, allocation, etc.).

abstract adjust(solution: Any, **kwargs) → Any

Given an optimization solution (and any auxiliary data), return an adjusted solution.

abstract initial(labels: Series, df_indiv: DataFrame, df_host: DataFrame) → list[dict[str, Any]]

Produce the initial problem solution if needed.

Parameters:

• labels – cluster labels from initial clustering

• df_indiv – individual‐level data

• df_host – host‐level data

Returns:

list of move dicts to apply