Skip to content

Agent

Data Structures for LLM Agent.

Task

Bases: BaseModel

Represents a single task with an instruction.

Attributes:

Name Type Description
id_ str

Identifier for task.
instruction str

The instruction for the task.
Source code in src/llm_agents_from_scratch/data_structures/agent.py 
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
class Task(BaseModel):
    """Represents a single task with an instruction.

    Attributes:
        id_: Identifier for task.
        instruction: The instruction for the task.
    """

    id_: str = Field(default_factory=lambda: str(uuid.uuid4()))
    instruction: str

    def __str__(self) -> str:
        """String representation of Task."""
        return self.instruction

__str__ 

__str__()

String representation of Task.

Source code in src/llm_agents_from_scratch/data_structures/agent.py 
20
21
22
def __str__(self) -> str:
    """String representation of Task."""
    return self.instruction

TaskStep

Bases: BaseModel

Represents a step within a task and its own instruction.

Attributes:

Name Type Description
id_ str

Identifier for task step.
task_id str

ID of associated task.
instruction str

The instruction for the task.
Source code in src/llm_agents_from_scratch/data_structures/agent.py 
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
class TaskStep(BaseModel):
    """Represents a step within a task and its own instruction.

    Attributes:
        id_: Identifier for task step.
        task_id: ID of associated task.
        instruction: The instruction for the task.
    """

    id_: str = Field(default_factory=lambda: str(uuid.uuid4()))
    task_id: str
    instruction: str = Field(
        description="The instruction for this step in the task.",
    )

    def __str__(self) -> str:
        """String representation of TaskStep."""
        return self.instruction

__str__ 

__str__()

String representation of TaskStep.

Source code in src/llm_agents_from_scratch/data_structures/agent.py 
40
41
42
def __str__(self) -> str:
    """String representation of TaskStep."""
    return self.instruction

TaskStepResult

Bases: BaseModel

The result of a task step execution.

Attributes:

Name Type Description
task_step_id str

The ID of the TaskStep that was executed.
content str

The content results of the execution.
Source code in src/llm_agents_from_scratch/data_structures/agent.py 
45
46
47
48
49
50
51
52
53
54
55
56
57
58
class TaskStepResult(BaseModel):
    """The result of a task step execution.

    Attributes:
        task_step_id: The ID of the `TaskStep` that was executed.
        content: The content results of the execution.
    """

    task_step_id: str
    content: str

    def __str__(self) -> str:
        """String representation of TaskStepResult."""
        return self.content

__str__ 

__str__()

String representation of TaskStepResult.

Source code in src/llm_agents_from_scratch/data_structures/agent.py 
56
57
58
def __str__(self) -> str:
    """String representation of TaskStepResult."""
    return self.content

TaskResult

Bases: BaseModel

The result of the task execution.

Attributes:

Name Type Description
task_id str

The ID Task that was executed.
content str

The content results of the task execution.
Source code in src/llm_agents_from_scratch/data_structures/agent.py 
61
62
63
64
65
66
67
68
69
70
71
72
73
74
class TaskResult(BaseModel):
    """The result of the task execution.

    Attributes:
        task_id: The ID `Task` that was executed.
        content: The content results of the task execution.
    """

    task_id: str
    content: str

    def __str__(self) -> str:
        """String representation of TaskResult."""
        return self.content

__str__ 

__str__()

String representation of TaskResult.

Source code in src/llm_agents_from_scratch/data_structures/agent.py 
72
73
74
def __str__(self) -> str:
    """String representation of TaskResult."""
    return self.content

NextStepDecision

Bases: BaseModel

Structured output used within TaskHandler.get_next_step().

Source code in src/llm_agents_from_scratch/data_structures/agent.py 
77
78
79
80
81
82
83
84
85
86
87
class NextStepDecision(BaseModel):
    """Structured output used within TaskHandler.get_next_step()."""

    kind: Literal["next_step", "final_result"]
    content: str = Field(
        description=(
            "If kind='next_step': describe the next tool call to make. "
            "If kind='final_result': the final answer (ONLY when task "
            "objective is fully achieved)."
        ),
    )

RejectedTaskResult

Bases: BaseModel

Human rejection feedback from the approval gate.

Added in Chapter 8. Returned by _process_loop when the operator rejects a proposed TaskResult; passed to get_next_step so it can route deterministically to a new TaskStep without consulting the LLM.

Attributes:

Name Type Description
failed_result_content str

Content of the rejected TaskResult.
feedback str

The operator's raw correction rationale.
Source code in src/llm_agents_from_scratch/data_structures/agent.py 
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
class RejectedTaskResult(BaseModel):
    """Human rejection feedback from the approval gate.

    Added in Chapter 8. Returned by ``_process_loop`` when the operator
    rejects a proposed ``TaskResult``; passed to ``get_next_step`` so it
    can route deterministically to a new ``TaskStep`` without consulting
    the LLM.

    Attributes:
        failed_result_content: Content of the rejected ``TaskResult``.
        feedback: The operator's raw correction rationale.
    """

    failed_result_content: str
    feedback: str

ApprovalResult

Bases: BaseModel

The outcome of the end-of-loop human approval gate.

Added in Chapter 8.

Attributes:

Name Type Description
approved bool

Whether the human accepted the proposed task result.
feedback str

The human's correction rationale on rejection; empty on approval.
Source code in src/llm_agents_from_scratch/data_structures/agent.py 
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
class ApprovalResult(BaseModel):
    """The outcome of the end-of-loop human approval gate.

    Added in Chapter 8.

    Attributes:
        approved: Whether the human accepted the proposed task result.
        feedback: The human's correction rationale on rejection; empty
            on approval.
    """

    approved: bool
    feedback: str = ""

    def __str__(self) -> str:
        """String representation of ApprovalResult."""
        if self.approved:
            return "approved"
        return f"rejected: {self.feedback}"

__str__ 

__str__()

String representation of ApprovalResult.

Source code in src/llm_agents_from_scratch/data_structures/agent.py 
121
122
123
124
125
def __str__(self) -> str:
    """String representation of ApprovalResult."""
    if self.approved:
        return "approved"
    return f"rejected: {self.feedback}"