rtac.ac_functionalities.ta_runner

In this module classes to run the target algorithm and observe its output through the wrapper are implemented.

Functions

gb_record(self, ta_output)

Records runtime output of the target algorithm if there was any new.

non_block_read(ta_output[, logs])

Function for reading subprocess.PIPE output without blocking the application until there is output.

ta_runner_factory(scenario, logs, core)

Class factory to return the initialized target algorithm runner class appropriate to the RTAC method specified in scenario.ac.

Classes

AbstractTARunner(scenario, logs, core)

Abstract TARunner.

BaseTARunner(scenario, logs, core)

Target algorithm runner for ReACTR implementation.

TARunnerpp(scenario, logs, core)

Target algorithm runner for ReACTR++ implementation.
class rtac.ac_functionalities.ta_runner.AbstractTARunner(scenario: Namespace, logs: RTACLogs, core: int)[source]

Bases: ABC

Abstract TARunner.

Parameters:

  • scenario (argparse.Namespace) – Namespace containing all settings for the RTAC.

  • logs (RTACLogs) – Object containing loggers and logging functions.

  • core (int) – Number of the parallel run started on a core.

abstract check_output(ta_output: bytes) → None[source]

Checks the output, if there was any, and declares the instance as solved by the contender if the corresponding marker is present.

Parameters:

ta_output (bytes) – Output from subprocess.PIPE.

Return type:

None

abstract check_result() → None[source]

If this contender solved the problem instance the rtac data is populated by the resulting information according to the RTAC method scenario.ac.

Return type:

None

abstract kill_run() → None[source]

Terminates this process/ target algorithm run, as well as the other contenders. Several layers of termination are included to ensure termination on different platforms.

Return type:

None

abstract run(instance: str, config: Configuration, rtac_data: RTACData | RTACDatapp) → None[source]

Manages the target algorithm runner functions depending on the state of the run, according to the RTAC method specified in scenario.ac.

Parameters:

  • instance (str) – Path to the problem instance to solve.

  • config (Configuration) – Configuration of parameter values to run the problem instance with.

  • rtac_data (RTACData | RTACDatapp) – Object containing data and objects necessary throughout the RTAC modules.

Return type:

None

abstract start_run(instance: str, config: Configuration, rtac_data: RTACData | RTACDatapp) → None[source]

Starts the target algorithm run and populates data necessary for coronation of tournament members.

Parameters:

  • instance (str) – Path to the problem instance to solve.

  • config (Configuration) – Representation of the configuration.

  • rtac_data (RTACData | RTACDatapp) – Object containing data and objects necessary throughout the RTAC modules.

Return type:

None

abstract translate_config(config: Configuration) → Any[source]

Convert dictionary representation of the configuration to the format needed by the wrapper to pass to the target algorithm.

Parameters:

config (Configuration) – Configuration of parameter values to run the problem instance with.

Returns:

New representation of the configuration suitable for the target algorithm.

Return type:

Any

class rtac.ac_functionalities.ta_runner.BaseTARunner(scenario: Namespace, logs: RTACLogs, core: int)[source]

Bases: AbstractTARunner

Target algorithm runner for ReACTR implementation.

Parameters:

  • scenario (argparse.Namespace) – Namespace containing all settings for the RTAC.

  • logs (RTACLogs) – Object containing loggers and logging functions.

  • core (int) – Identifier for the parallel run on a specific core.

check_output(ta_output: bytes) → None[source]

Checks the output, if there was any, and declares the instance as solved by the contender if the corresponding marker is present.

Parameters:

ta_output (bytes) – Output from subprocess.PIPE.

Return type:

None

check_result() → None[source]

If this contender solved the problem instance the rtac data is populated by the resulting information.

Return type:

None

kill_run() → None[source]

Terminates this process/ target algorithm run, as well as the other contenders. Several layers of termination are included to ensure termination on different platforms.

Return type:

None

run(instance: str, config: Configuration, rtac_data: RTACData | RTACDatapp, sync_event: Event) → None[source]

Manages the target algorithm runner functions depending on the state of the run.

Parameters:

  • instance (str) – Path to the problem instance to solve.

  • config (Configuration) – Configuration of parameter values to run the problem instance with.

  • rtac_data (RTACData | RTACDatapp) – Object containing data and objects necessary throughout the RTAC modules.

Return type:

None

start_run(instance: str, config: Any, rtac_data: RTACData | RTACDatapp) → None[source]

Starts the target algorithm run and populates data necessary for coronation of tournament members.

Parameters:

  • instance (str) – Path to the problem instance to solve.

  • config (Any) – Representation of the configuration.

  • rtac_data (RTACData | RTACDatapp) – Object containing data and objects necessary throughout the RTAC modules.

Return type:

None

translate_config(config: Configuration) → Any[source]

Convert dictionary representation of the configuration to the format needed by the wrapper to pass to the target algorithm.

Parameters:

config (Configuration) – Configuration of parameter values to run the problem instance with.

Returns:

New representation of the configuration.

Return type:

Any

class rtac.ac_functionalities.ta_runner.TARunnerpp(scenario: Namespace, logs: RTACLogs, core: int)[source]

Bases: BaseTARunner

Target algorithm runner for ReACTR++ implementation.

Parameters:

  • scenario (argparse.Namespace) – Namespace containing all settings for the RTAC.

  • logs (RTACLogs) – Object containing loggers and logging functions.

  • core (int) – Identifier for the parallel run on a specific core.

check_output(ta_output: bytes) → None[source]

Checks the output, if there was any, and either declares the instance as solved by the contender (if the corresponding marker is present) or outputs intermediate target algorithm output.

Parameters:

ta_output (bytes) – Output from subprocess.PIPE.

Return type:

None

rtac.ac_functionalities.ta_runner.gb_record(self, ta_output: bytes) → None[source]

Records runtime output of the target algorithm if there was any new.

Parameters:

ta_output (bytes) – Output from subprocess.PIPE.

Return type:

None

rtac.ac_functionalities.ta_runner.non_block_read(ta_output: bytes, logs: RTACLogs | None = None) → str[source]

Function for reading subprocess.PIPE output without blocking the application until there is output. It checks for output and exits immediately if none is available.

Parameters:

  • ta_output (bytes) – subprocess.PIPE output to be read.

  • logs (RTACLogs) – Object containing loggers and logging functions. Defaults to None.

Returns:

Either the output as a string or an empty string if no output was available.

Return type:

str

rtac.ac_functionalities.ta_runner.ta_runner_factory(scenario: Namespace, logs: RTACLogs, core: int) → BaseTARunner | TARunnerpp[source]

Class factory to return the initialized target algorithm runner class appropriate to the RTAC method specified in scenario.ac.

Parameters:

  • scenario (argparse.Namespace) – Namespace containing all settings for the RTAC.

  • logs (RTACLogs) – Object containing loggers and logging functions.

  • core (int) – Number of the parallel run started on a core.

Returns:

Initialized target algorithm runner object matching the RTAC method of the scenario.

Return type:

BaseTARunner or TARunnerpp