Skip to content

feat: GenAI Client(evals): Update SDK class to implement multi-turn agent scraping in evals.run_inference function #6345

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
copybara-service wants to merge 1 commit into
base: main
Choose a base branch
from
Open

feat: GenAI Client(evals): Update SDK class to implement multi-turn agent scraping in evals.run_inference function #6345

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 12 additions & 0 deletions vertexai/_genai/types/__init__.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1042,6 +1042,12 @@
from .common import UpdateMultimodalDatasetConfig
from .common import UpdateMultimodalDatasetConfigDict
from .common import UpdateMultimodalDatasetConfigOrDict
from .common import UserScenario
from .common import UserScenarioDict
from .common import UserScenarioOrDict
from .common import UserSimulatorConfig
from .common import UserSimulatorConfigDict
from .common import UserSimulatorConfigOrDict
from .common import VertexBaseConfig
from .common import VertexBaseConfigDict
from .common import VertexBaseConfigOrDict
Expand All @@ -1059,6 +1065,9 @@
"PromptTemplateData",
"PromptTemplateDataDict",
"PromptTemplateDataOrDict",
"UserScenario",
"UserScenarioDict",
"UserScenarioOrDict",
"EvaluationPrompt",
"EvaluationPromptDict",
"EvaluationPromptOrDict",
Expand Down Expand Up @@ -1155,6 +1164,9 @@
"EvaluationRunAgentConfig",
"EvaluationRunAgentConfigDict",
"EvaluationRunAgentConfigOrDict",
"UserSimulatorConfig",
"UserSimulatorConfigDict",
"UserSimulatorConfigOrDict",
"EvaluationRunInferenceConfig",
"EvaluationRunInferenceConfigDict",
"EvaluationRunInferenceConfigOrDict",
Expand Down
81 changes: 81 additions & 0 deletions vertexai/_genai/types/common.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -490,6 +490,32 @@ class PromptTemplateDataDict(TypedDict, total=False):
PromptTemplateDataOrDict = Union[PromptTemplateData, PromptTemplateDataDict]


class UserScenario(_common.BaseModel):
"""User scenario to help simulate multi-turn agent running results."""

starting_prompt: Optional[str] = Field(
default=None,
description="""The prompt that starts the conversation between the simulated user and the agent under test.""",
)
conversation_plan: Optional[str] = Field(
default=None,
description="""The plan for the conversation, used to drive the multi-turn agent run and generate the simulated agent evaluation dataset.""",
)


class UserScenarioDict(TypedDict, total=False):
"""User scenario to help simulate multi-turn agent running results."""

starting_prompt: Optional[str]
"""The prompt that starts the conversation between the simulated user and the agent under test."""

conversation_plan: Optional[str]
"""The plan for the conversation, used to drive the multi-turn agent run and generate the simulated agent evaluation dataset."""


UserScenarioOrDict = Union[UserScenario, UserScenarioDict]


class EvaluationPrompt(_common.BaseModel):
"""Represents the prompt to be evaluated."""

Expand All @@ -501,6 +527,10 @@ class EvaluationPrompt(_common.BaseModel):
prompt_template_data: Optional[PromptTemplateData] = Field(
default=None, description="""Prompt template data."""
)
user_scenario: Optional[UserScenario] = Field(
default=None,
description="""User scenario to help simulate multi-turn agent running results.""",
)


class EvaluationPromptDict(TypedDict, total=False):
Expand All @@ -515,6 +545,9 @@ class EvaluationPromptDict(TypedDict, total=False):
prompt_template_data: Optional[PromptTemplateDataDict]
"""Prompt template data."""

user_scenario: Optional[UserScenarioDict]
"""User scenario to help simulate multi-turn agent running results."""


EvaluationPromptOrDict = Union[EvaluationPrompt, EvaluationPromptDict]

Expand Down Expand Up @@ -1830,6 +1863,38 @@ class EvaluationRunAgentConfigDict(TypedDict, total=False):
]


class UserSimulatorConfig(_common.BaseModel):
"""Configuration for a user simulator that uses an LLM to generate messages."""

model_name: Optional[str] = Field(
default=None,
description="""The model name to use for multi-turn agent scraping.""",
)
model_configuration: Optional[genai_types.GenerateContentConfig] = Field(
default=None, description="""The configuration for the model."""
)
max_turn: Optional[int] = Field(
default=None,
description="""Maximum number of invocations allowed. Stops run-off conversations.""",
)


class UserSimulatorConfigDict(TypedDict, total=False):
"""Configuration for a user simulator that uses an LLM to generate messages."""

model_name: Optional[str]
"""The model name to use for multi-turn agent scraping."""

model_configuration: Optional[genai_types.GenerateContentConfigDict]
"""The configuration for the model."""

max_turn: Optional[int]
"""Maximum number of invocations allowed. Stops run-off conversations."""


UserSimulatorConfigOrDict = Union[UserSimulatorConfig, UserSimulatorConfigDict]


class EvaluationRunInferenceConfig(_common.BaseModel):
"""This field is experimental and may change in future versions.

Expand All @@ -1843,6 +1908,10 @@ class EvaluationRunInferenceConfig(_common.BaseModel):
default=None,
description="""The fully qualified name of the publisher model or endpoint to use for inference.""",
)
user_simulator_config: Optional[UserSimulatorConfig] = Field(
default=None,
description="""Configuration for user simulation in multi-turn agent scraping. If provided, and the dataset contains conversation plans, user simulation will be triggered.""",
)


class EvaluationRunInferenceConfigDict(TypedDict, total=False):
Expand All @@ -1857,6 +1926,9 @@ class EvaluationRunInferenceConfigDict(TypedDict, total=False):
model: Optional[str]
"""The fully qualified name of the publisher model or endpoint to use for inference."""

user_simulator_config: Optional[UserSimulatorConfigDict]
"""Configuration for user simulation in multi-turn agent scraping. If provided, and the dataset contains conversation plans, user simulation will be triggered."""


EvaluationRunInferenceConfigOrDict = Union[
EvaluationRunInferenceConfig, EvaluationRunInferenceConfigDict
Expand Down Expand Up @@ -13546,6 +13618,11 @@ class EvalRunInferenceConfig(_common.BaseModel):
generate_content_config: Optional[genai_types.GenerateContentConfig] = Field(
default=None, description="""The config for the generate content call."""
)
user_simulator_config: Optional[UserSimulatorConfig] = Field(
default=None,
description="""Configuration for user simulation in multi-turn agent scraping. If provided, and the dataset contains
conversation plans, user simulation will be triggered.""",
)


class EvalRunInferenceConfigDict(TypedDict, total=False):
Expand All @@ -13560,6 +13637,10 @@ class EvalRunInferenceConfigDict(TypedDict, total=False):
generate_content_config: Optional[genai_types.GenerateContentConfigDict]
"""The config for the generate content call."""

user_simulator_config: Optional[UserSimulatorConfigDict]
"""Configuration for user simulation in multi-turn agent scraping. If provided, and the dataset contains
conversation plans, user simulation will be triggered."""


EvalRunInferenceConfigOrDict = Union[EvalRunInferenceConfig, EvalRunInferenceConfigDict]

Expand Down
Loading