crab.core.models package

Submodules

crab.core.models.action module

class crab.core.models.action.Action(*, name: str, entry: Callable, parameters: type[BaseModel], returns: type[BaseModel], description: str | None = None, kept_params: list[str] = [], env_name: str | None = None, local: bool = False)

Bases: BaseModel

The core operational unit within the Crab system.

This class stores parameters and return type definitions and can be easily converted into a JSON schema. It supports argument verification and includes a feature for retaining specific parameters.

name

The name of the action.

Type:

str

entry

The actual entry function of the action.

Type:

Callable

parameters

Definition of input parameters.

Type:

type[BaseModel]

returns

Definition of the return type. Note: The actual return type is specified by the returns attribute in this model.

Type:

type[BaseModel]

description

A clear and concise description of the function’s purpose and behavior. Defaults to None.

Type:

str | None

kept_params

Parameters retained for internal use by the Crab system, such as ‘env’ for storing the current environment. These parameters do not appear in the parameters field and are automatically injected at runtime. Defaults to an empty dictionary.

Type:

dict[str, Any]

env_name

Specify the environment the action is associated with. Defualts to None.

Type:

Optinal[str]

description: str | None

entry: Callable

env_name: str | None

classmethod from_function(func: Callable) → Self

Generates an action from functions annotated by @action.

get_required_params() → dict[str, FieldInfo]

kept_params: list[str]

local: bool

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'description': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'entry': FieldInfo(annotation=Callable, required=True), 'env_name': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'kept_params': FieldInfo(annotation=list[str], required=False, default=[]), 'local': FieldInfo(annotation=bool, required=False, default=False), 'name': FieldInfo(annotation=str, required=True), 'parameters': FieldInfo(annotation=type[BaseModel], required=True), 'returns': FieldInfo(annotation=type[BaseModel], required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

name: str

parameters: type[BaseModel]

returns: type[BaseModel]

run(**kwargs) → Any

Varifies the action parameters then runes the action.

set_kept_param(**params) → Self

to_openai_json_schema() → dict

Gets openai json schema from an action

to_raw_action() → dict[str, Any]

Gets serialized action for remote execution

crab.core.models.action.ClosedAction

The action type alias with no reuqired parameters

alias of Annotated[Action, AfterValidator(func=_check_no_param)]

crab.core.models.action.EMPTY_MODEL

alias of Empty

crab.core.models.benchmark_interface module

class crab.core.models.benchmark_interface.StepResult(*, truncated: bool, terminated: bool, action_returns: Any, evaluation_results: dict[str, Any], info: dict[str, Any])

Bases: BaseModel

action_returns: Any

evaluation_results: dict[str, Any]

info: dict[str, Any]

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'action_returns': FieldInfo(annotation=Any, required=True), 'evaluation_results': FieldInfo(annotation=dict[str, Any], required=True), 'info': FieldInfo(annotation=dict[str, Any], required=True), 'terminated': FieldInfo(annotation=bool, required=True), 'truncated': FieldInfo(annotation=bool, required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

terminated: bool

truncated: bool

crab.core.models.config module

class crab.core.models.config.BenchmarkConfig(*, name: str, tasks: list[Task], environments: list[EnvironmentConfig], default_env: str | None = None, multienv: bool = False, prompting_tools: dict[str, dict[str, Action]] = {}, root_action_space: list[Action] = [], step_limit: int = 30, common_setup: list[Annotated[Action, AfterValidator(func=_check_no_param)]] = [])

Bases: BaseModel

common_setup: list[Annotated[Action, AfterValidator(func=_check_no_param)]]

default_env: str | None

environments: list[EnvironmentConfig]

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'common_setup': FieldInfo(annotation=list[Annotated[Action, AfterValidator]], required=False, default=[]), 'default_env': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'environments': FieldInfo(annotation=list[EnvironmentConfig], required=True), 'multienv': FieldInfo(annotation=bool, required=False, default=False), 'name': FieldInfo(annotation=str, required=True), 'prompting_tools': FieldInfo(annotation=dict[str, dict[str, Action]], required=False, default={}), 'root_action_space': FieldInfo(annotation=list[Action], required=False, default=[]), 'step_limit': FieldInfo(annotation=int, required=False, default=30), 'tasks': FieldInfo(annotation=list[Task], required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

multienv: bool

name: str

prompting_tools: dict[str, dict[str, Action]]

root_action_space: list[Action]

step_limit: int

tasks: list[Task]

class crab.core.models.config.EnvironmentConfig(*, name: str, action_space: list[Action], observation_space: list[Annotated[Action, AfterValidator(func=_check_no_param)]], description: str = '', reset: Action | None = None, remote_url: str | None = None, extra_attributes: dict[str, Any] = {})

Bases: BaseModel

action_space: list[Action]

description: str

extra_attributes: dict[str, Any]

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'action_space': FieldInfo(annotation=list[Action], required=True), 'description': FieldInfo(annotation=str, required=False, default=''), 'extra_attributes': FieldInfo(annotation=dict[str, Any], required=False, default={}), 'name': FieldInfo(annotation=str, required=True), 'observation_space': FieldInfo(annotation=list[Annotated[Action, AfterValidator]], required=True), 'remote_url': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'reset': FieldInfo(annotation=Union[Action, NoneType], required=False, default=None)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

name: str

observation_space: list[Annotated[Action, AfterValidator(func=_check_no_param)]]

remote_url: str | None

reset: Action | None

class crab.core.models.config.VMEnvironmentConfig(*, inside_environment: EnvironmentConfig, remote_url: str = 'http://192.168.0.0:8000')

Bases: BaseModel

inside_environment: EnvironmentConfig

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'inside_environment': FieldInfo(annotation=EnvironmentConfig, required=True), 'remote_url': FieldInfo(annotation=str, required=False, default='http://192.168.0.0:8000')}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

remote_url: str

crab.core.models.evaluator module

class crab.core.models.evaluator.Evaluator(*, name: str, entry: Callable, parameters: type[BaseModel], returns: type[BaseModel], description: str | None = None, kept_params: list[str] = [], env_name: str | None = None, local: bool = False, require_submit: bool = False)

Bases: Action

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'description': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'entry': FieldInfo(annotation=Callable, required=True), 'env_name': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'kept_params': FieldInfo(annotation=list[str], required=False, default=[]), 'local': FieldInfo(annotation=bool, required=False, default=False), 'name': FieldInfo(annotation=str, required=True), 'parameters': FieldInfo(annotation=type[BaseModel], required=True), 'require_submit': FieldInfo(annotation=bool, required=False, default=False), 'returns': FieldInfo(annotation=type[BaseModel], required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

classmethod must_return_bool(v: type[BaseModel]) → type[BaseModel]

require_submit: bool

crab.core.models.task module

class crab.core.models.task.GeneratedTask(*, description: str, tasks: list[SubTaskInstance], adjlist: str, id: str = None)

Bases: BaseModel

adjlist: str

description: str

id: str

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'adjlist': FieldInfo(annotation=str, required=True), 'description': FieldInfo(annotation=str, required=True), 'id': FieldInfo(annotation=str, required=False, default_factory=uuid4), 'tasks': FieldInfo(annotation=list[SubTaskInstance], required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

tasks: list[SubTaskInstance]

class crab.core.models.task.SubTask(*, id: str, description: str, attribute_dict: dict[str, list[str] | str], output_type: str, output_generator: Callable[[Any], str] | Literal['manual'] | None = None, evaluator_generator: Callable[[Any], DiGraph] | None = None, setup: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)] = [], teardown: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)] = [], extra_action: list[Action] = [])

Bases: BaseModel

attribute_dict: dict[str, list[str] | str]

description: str

evaluator_generator: Callable[[Any], DiGraph] | None

classmethod expand_attribute_type(attribute_dict: dict[str, list[str] | str]) → dict[str, list[str]]

extra_action: list[Action]

id: str

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'attribute_dict': FieldInfo(annotation=dict[str, Union[list[str], str]], required=True), 'description': FieldInfo(annotation=str, required=True), 'evaluator_generator': FieldInfo(annotation=Union[Callable[list, DiGraph], NoneType], required=False, default=None), 'extra_action': FieldInfo(annotation=list[Action], required=False, default=[]), 'id': FieldInfo(annotation=str, required=True), 'output_generator': FieldInfo(annotation=Union[Callable[list, str], Literal['manual'], NoneType], required=False, default=None), 'output_type': FieldInfo(annotation=str, required=True), 'setup': FieldInfo(annotation=Union[list[Annotated[Action, AfterValidator]], Annotated[Action, AfterValidator]], required=False, default=[]), 'teardown': FieldInfo(annotation=Union[list[Annotated[Action, AfterValidator]], Annotated[Action, AfterValidator]], required=False, default=[])}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

output_generator: Callable[[Any], str] | Literal['manual'] | None

output_type: str

setup: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)]

teardown: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)]

class crab.core.models.task.SubTaskInstance(*, task: SubTask, attribute: dict[str, Any], output: str | None = None, id: str = None)

Bases: BaseModel

attribute: dict[str, Any]

dump_model() → dict[str, Any]

id: str

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'attribute': FieldInfo(annotation=dict[str, Any], required=True), 'id': FieldInfo(annotation=str, required=False, default_factory=uuid4), 'output': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'task': FieldInfo(annotation=SubTask, required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

output: str | None

task: SubTask

class crab.core.models.task.Task(*, id: str, description: str, evaluator: DiGraph | Evaluator, setup: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)] = [], teardown: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)] = [], extra_action: list[Action] = [])

Bases: BaseModel

classmethod change_evaluator_to_graph(evaluator: DiGraph | Evaluator) → str

description: str

evaluator: DiGraph | Evaluator

extra_action: list[Action]

id: str

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {'arbitrary_types_allowed': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'description': FieldInfo(annotation=str, required=True), 'evaluator': FieldInfo(annotation=Union[DiGraph, Evaluator], required=True), 'extra_action': FieldInfo(annotation=list[Action], required=False, default=[]), 'id': FieldInfo(annotation=str, required=True), 'setup': FieldInfo(annotation=Union[list[Annotated[Action, AfterValidator]], Annotated[Action, AfterValidator]], required=False, default=[]), 'teardown': FieldInfo(annotation=Union[list[Annotated[Action, AfterValidator]], Annotated[Action, AfterValidator]], required=False, default=[])}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

setup: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)]

teardown: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)]

classmethod to_list(action: Action | list[Action]) → list[Action]

Module contents

class crab.core.models.Action(*, name: str, entry: Callable, parameters: type[BaseModel], returns: type[BaseModel], description: str | None = None, kept_params: list[str] = [], env_name: str | None = None, local: bool = False)

Bases: BaseModel

The core operational unit within the Crab system.

This class stores parameters and return type definitions and can be easily converted into a JSON schema. It supports argument verification and includes a feature for retaining specific parameters.

name

The name of the action.

Type:

str

entry

The actual entry function of the action.

Type:

Callable

parameters

Definition of input parameters.

Type:

type[BaseModel]

returns

Definition of the return type. Note: The actual return type is specified by the returns attribute in this model.

Type:

type[BaseModel]

description

A clear and concise description of the function’s purpose and behavior. Defaults to None.

Type:

str | None

kept_params

Parameters retained for internal use by the Crab system, such as ‘env’ for storing the current environment. These parameters do not appear in the parameters field and are automatically injected at runtime. Defaults to an empty dictionary.

Type:

dict[str, Any]

env_name

Specify the environment the action is associated with. Defualts to None.

Type:

Optinal[str]

description: str | None

entry: Callable

env_name: str | None

classmethod from_function(func: Callable) → Self

Generates an action from functions annotated by @action.

get_required_params() → dict[str, FieldInfo]

kept_params: list[str]

local: bool

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'description': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'entry': FieldInfo(annotation=Callable, required=True), 'env_name': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'kept_params': FieldInfo(annotation=list[str], required=False, default=[]), 'local': FieldInfo(annotation=bool, required=False, default=False), 'name': FieldInfo(annotation=str, required=True), 'parameters': FieldInfo(annotation=type[BaseModel], required=True), 'returns': FieldInfo(annotation=type[BaseModel], required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

name: str

parameters: type[BaseModel]

returns: type[BaseModel]

run(**kwargs) → Any

Varifies the action parameters then runes the action.

set_kept_param(**params) → Self

to_openai_json_schema() → dict

Gets openai json schema from an action

to_raw_action() → dict[str, Any]

Gets serialized action for remote execution

class crab.core.models.ActionOutput(*, name: str, arguments: dict[str, Any], env: str | None = None)

Bases: BaseModel

arguments: dict[str, Any]

env: str | None

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'arguments': FieldInfo(annotation=dict[str, Any], required=True), 'env': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'name': FieldInfo(annotation=str, required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

name: str

class crab.core.models.BackendOutput(*, message: str | None, action_list: list[ActionOutput] | None)

Bases: BaseModel

action_list: list[ActionOutput] | None

message: str | None

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'action_list': FieldInfo(annotation=Union[list[ActionOutput], NoneType], required=True), 'message': FieldInfo(annotation=Union[str, NoneType], required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

class crab.core.models.BenchmarkConfig(*, name: str, tasks: list[Task], environments: list[EnvironmentConfig], default_env: str | None = None, multienv: bool = False, prompting_tools: dict[str, dict[str, Action]] = {}, root_action_space: list[Action] = [], step_limit: int = 30, common_setup: list[Annotated[Action, AfterValidator(func=_check_no_param)]] = [])

Bases: BaseModel

common_setup: list[Annotated[Action, AfterValidator(func=_check_no_param)]]

default_env: str | None

environments: list[EnvironmentConfig]

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'common_setup': FieldInfo(annotation=list[Annotated[Action, AfterValidator]], required=False, default=[]), 'default_env': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'environments': FieldInfo(annotation=list[EnvironmentConfig], required=True), 'multienv': FieldInfo(annotation=bool, required=False, default=False), 'name': FieldInfo(annotation=str, required=True), 'prompting_tools': FieldInfo(annotation=dict[str, dict[str, Action]], required=False, default={}), 'root_action_space': FieldInfo(annotation=list[Action], required=False, default=[]), 'step_limit': FieldInfo(annotation=int, required=False, default=30), 'tasks': FieldInfo(annotation=list[Task], required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

multienv: bool

name: str

prompting_tools: dict[str, dict[str, Action]]

root_action_space: list[Action]

step_limit: int

tasks: list[Task]

class crab.core.models.EnvironmentConfig(*, name: str, action_space: list[Action], observation_space: list[Annotated[Action, AfterValidator(func=_check_no_param)]], description: str = '', reset: Action | None = None, remote_url: str | None = None, extra_attributes: dict[str, Any] = {})

Bases: BaseModel

action_space: list[Action]

description: str

extra_attributes: dict[str, Any]

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'action_space': FieldInfo(annotation=list[Action], required=True), 'description': FieldInfo(annotation=str, required=False, default=''), 'extra_attributes': FieldInfo(annotation=dict[str, Any], required=False, default={}), 'name': FieldInfo(annotation=str, required=True), 'observation_space': FieldInfo(annotation=list[Annotated[Action, AfterValidator]], required=True), 'remote_url': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'reset': FieldInfo(annotation=Union[Action, NoneType], required=False, default=None)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

name: str

observation_space: list[Annotated[Action, AfterValidator(func=_check_no_param)]]

remote_url: str | None

reset: Action | None

class crab.core.models.Evaluator(*, name: str, entry: Callable, parameters: type[BaseModel], returns: type[BaseModel], description: str | None = None, kept_params: list[str] = [], env_name: str | None = None, local: bool = False, require_submit: bool = False)

Bases: Action

description: str | None

entry: Callable

env_name: str | None

kept_params: list[str]

local: bool

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'description': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'entry': FieldInfo(annotation=Callable, required=True), 'env_name': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'kept_params': FieldInfo(annotation=list[str], required=False, default=[]), 'local': FieldInfo(annotation=bool, required=False, default=False), 'name': FieldInfo(annotation=str, required=True), 'parameters': FieldInfo(annotation=type[BaseModel], required=True), 'require_submit': FieldInfo(annotation=bool, required=False, default=False), 'returns': FieldInfo(annotation=type[BaseModel], required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

classmethod must_return_bool(v: type[BaseModel]) → type[BaseModel]

name: str

parameters: type[BaseModel]

require_submit: bool

returns: type[BaseModel]

class crab.core.models.GeneratedTask(*, description: str, tasks: list[SubTaskInstance], adjlist: str, id: str = None)

Bases: BaseModel

adjlist: str

description: str

id: str

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'adjlist': FieldInfo(annotation=str, required=True), 'description': FieldInfo(annotation=str, required=True), 'id': FieldInfo(annotation=str, required=False, default_factory=uuid4), 'tasks': FieldInfo(annotation=list[SubTaskInstance], required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

tasks: list[SubTaskInstance]

crab.core.models.Message

alias of tuple[str, MessageType]

class crab.core.models.MessageType(value)

Bases: IntEnum

An enumeration.

IMAGE_JPG_BASE64 = 1

TEXT = 0

class crab.core.models.StepResult(*, truncated: bool, terminated: bool, action_returns: Any, evaluation_results: dict[str, Any], info: dict[str, Any])

Bases: BaseModel

action_returns: Any

evaluation_results: dict[str, Any]

info: dict[str, Any]

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'action_returns': FieldInfo(annotation=Any, required=True), 'evaluation_results': FieldInfo(annotation=dict[str, Any], required=True), 'info': FieldInfo(annotation=dict[str, Any], required=True), 'terminated': FieldInfo(annotation=bool, required=True), 'truncated': FieldInfo(annotation=bool, required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

terminated: bool

truncated: bool

class crab.core.models.SubTask(*, id: str, description: str, attribute_dict: dict[str, list[str] | str], output_type: str, output_generator: Callable[[Any], str] | Literal['manual'] | None = None, evaluator_generator: Callable[[Any], DiGraph] | None = None, setup: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)] = [], teardown: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)] = [], extra_action: list[Action] = [])

Bases: BaseModel

attribute_dict: dict[str, list[str] | str]

description: str

evaluator_generator: Callable[[Any], DiGraph] | None

classmethod expand_attribute_type(attribute_dict: dict[str, list[str] | str]) → dict[str, list[str]]

extra_action: list[Action]

id: str

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'attribute_dict': FieldInfo(annotation=dict[str, Union[list[str], str]], required=True), 'description': FieldInfo(annotation=str, required=True), 'evaluator_generator': FieldInfo(annotation=Union[Callable[list, DiGraph], NoneType], required=False, default=None), 'extra_action': FieldInfo(annotation=list[Action], required=False, default=[]), 'id': FieldInfo(annotation=str, required=True), 'output_generator': FieldInfo(annotation=Union[Callable[list, str], Literal['manual'], NoneType], required=False, default=None), 'output_type': FieldInfo(annotation=str, required=True), 'setup': FieldInfo(annotation=Union[list[Annotated[Action, AfterValidator]], Annotated[Action, AfterValidator]], required=False, default=[]), 'teardown': FieldInfo(annotation=Union[list[Annotated[Action, AfterValidator]], Annotated[Action, AfterValidator]], required=False, default=[])}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

output_generator: Callable[[Any], str] | Literal['manual'] | None

output_type: str

setup: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)]

teardown: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)]

class crab.core.models.SubTaskInstance(*, task: SubTask, attribute: dict[str, Any], output: str | None = None, id: str = None)

Bases: BaseModel

attribute: dict[str, Any]

dump_model() → dict[str, Any]

id: str

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'attribute': FieldInfo(annotation=dict[str, Any], required=True), 'id': FieldInfo(annotation=str, required=False, default_factory=uuid4), 'output': FieldInfo(annotation=Union[str, NoneType], required=False, default=None), 'task': FieldInfo(annotation=SubTask, required=True)}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

output: str | None

task: SubTask

class crab.core.models.Task(*, id: str, description: str, evaluator: DiGraph | Evaluator, setup: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)] = [], teardown: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)] = [], extra_action: list[Action] = [])

Bases: BaseModel

classmethod change_evaluator_to_graph(evaluator: DiGraph | Evaluator) → str

description: str

evaluator: DiGraph | Evaluator

extra_action: list[Action]

id: str

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {'arbitrary_types_allowed': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'description': FieldInfo(annotation=str, required=True), 'evaluator': FieldInfo(annotation=Union[DiGraph, Evaluator], required=True), 'extra_action': FieldInfo(annotation=list[Action], required=False, default=[]), 'id': FieldInfo(annotation=str, required=True), 'setup': FieldInfo(annotation=Union[list[Annotated[Action, AfterValidator]], Annotated[Action, AfterValidator]], required=False, default=[]), 'teardown': FieldInfo(annotation=Union[list[Annotated[Action, AfterValidator]], Annotated[Action, AfterValidator]], required=False, default=[])}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

setup: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)]

teardown: list[Annotated[Action, AfterValidator(func=_check_no_param)]] | Annotated[Action, AfterValidator(func=_check_no_param)]

classmethod to_list(action: Action | list[Action]) → list[Action]

class crab.core.models.VMEnvironmentConfig(*, inside_environment: EnvironmentConfig, remote_url: str = 'http://192.168.0.0:8000')

Bases: BaseModel

inside_environment: EnvironmentConfig

model_computed_fields: ClassVar[Dict[str, ComputedFieldInfo]] = {}

A dictionary of computed field names and their corresponding ComputedFieldInfo objects.

model_config: ClassVar[ConfigDict] = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

model_fields: ClassVar[Dict[str, FieldInfo]] = {'inside_environment': FieldInfo(annotation=EnvironmentConfig, required=True), 'remote_url': FieldInfo(annotation=str, required=False, default='http://192.168.0.0:8000')}

Metadata about the fields defined on the model, mapping of field names to [FieldInfo][pydantic.fields.FieldInfo] objects.

This replaces Model.__fields__ from Pydantic V1.

remote_url: str