ai-station/.venv/lib/python3.12/site-packages/literalai/api/base.py

1064 lines
34 KiB
Python

import os
from abc import ABC, abstractmethod
from typing import Any, Dict, List, Optional, Union
from typing_extensions import deprecated
from literalai.api.helpers.attachment_helpers import AttachmentUpload
from literalai.api.helpers.prompt_helpers import PromptRollout
from literalai.api.helpers.score_helpers import ScoreUpdate
from literalai.cache.shared_cache import SharedCache
from literalai.evaluation.dataset import DatasetType
from literalai.evaluation.dataset_experiment import DatasetExperimentItem
from literalai.my_types import Environment
from literalai.observability.filter import (
generations_filters,
generations_order_by,
scores_filters,
scores_order_by,
steps_filters,
steps_order_by,
threads_filters,
threads_order_by,
users_filters,
)
from literalai.observability.generation import (
ChatGeneration,
CompletionGeneration,
GenerationMessage,
)
from literalai.observability.step import ScoreDict, ScoreType, Step, StepDict, StepType
from literalai.prompt_engineering.prompt import ProviderSettings
def prepare_variables(variables: Dict[str, Any]) -> Dict[str, Any]:
"""
Recursively checks and converts bytes objects in variables.
"""
def handle_bytes(item):
if isinstance(item, bytes):
return "STRIPPED_BINARY_DATA"
elif isinstance(item, dict):
return {k: handle_bytes(v) for k, v in item.items()}
elif isinstance(item, list):
return [handle_bytes(i) for i in item]
elif isinstance(item, tuple):
return tuple(handle_bytes(i) for i in item)
return item
return handle_bytes(variables)
class BaseLiteralAPI(ABC):
def __init__(
self,
api_key: Optional[str] = None,
url: Optional[str] = None,
environment: Optional[Environment] = None,
):
if url and url[-1] == "/":
url = url[:-1]
if api_key is None:
raise Exception("LITERAL_API_KEY not set")
if url is None:
raise Exception("LITERAL_API_URL not set")
self.api_key = api_key
self.url = url
if environment:
os.environ["LITERAL_ENV"] = environment
self.graphql_endpoint = self.url + "/api/graphql"
self.rest_endpoint = self.url + "/api"
self.cache = SharedCache()
@property
def headers(self):
from literalai.version import __version__
h = {
"Content-Type": "application/json",
"x-api-key": self.api_key,
"x-client-name": "py-literal-client",
"x-client-version": __version__,
}
if env := os.getenv("LITERAL_ENV"):
h["x-env"] = env
return h
@abstractmethod
def get_users(
self,
first: Optional[int] = None,
after: Optional[str] = None,
before: Optional[str] = None,
filters: Optional[users_filters] = None,
):
"""
Retrieves a list of users based on pagination and optional filters.
Args:
first (Optional[int]): The number of users to retrieve.
after (Optional[str]): A cursor for use in pagination, fetching records after this cursor.
before (Optional[str]): A cursor for use in pagination, fetching records before this cursor.
filters (Optional[users_filters]): Filters to apply to the user query.
Returns:
`PaginatedResponse[User]`: A paginated response containing the queried user data.
"""
pass
@abstractmethod
def get_user(self, id: Optional[str] = None, identifier: Optional[str] = None):
"""
Retrieves a user based on the provided ID or identifier.
Args:
id (Optional[str]): The unique ID of the user.
identifier (Optional[str]): A unique identifier for the user, such as a username or email.
Returns:
`User`: The user with requested id or identifier.
"""
pass
@abstractmethod
def create_user(self, identifier: str, metadata: Optional[Dict] = None):
"""
Creates a new user with the specified identifier and optional metadata.
Args:
identifier (str): A unique identifier for the user, such as a username or email.
metadata (Optional[Dict]): Additional data associated with the user.
Returns:
`User`: The created user object.
"""
pass
@abstractmethod
def update_user(
self, id: str, identifier: Optional[str] = None, metadata: Optional[Dict] = None
):
"""
Updates an existing user identified by the given ID, with optional new identifier and metadata.
Args:
id (str): The unique ID of the user to update.
identifier (Optional[str]): A new identifier for the user, such as a username or email.
metadata (Optional[Dict]): New or updated metadata for the user.
Returns:
`User`: The updated user object.
"""
pass
@abstractmethod
def delete_user(self, id: str):
"""
Deletes a user identified by the given ID.
Args:
id (str): The unique ID of the user to delete.
Returns:
Dict: The deleted user as a dict.
"""
pass
@abstractmethod
def get_or_create_user(self, identifier: str, metadata: Optional[Dict] = None):
"""
Retrieves a user by their identifier, or creates a new user if it does not exist.
Args:
identifier (str): The identifier of the user to retrieve or create.
metadata (Optional[Dict]): Metadata to associate with the user if they are created.
Returns:
`User`: The existing or newly created user.
"""
pass
@abstractmethod
def get_threads(
self,
first: Optional[int] = None,
after: Optional[str] = None,
before: Optional[str] = None,
filters: Optional[threads_filters] = None,
order_by: Optional[threads_order_by] = None,
step_types_to_keep: Optional[List[StepType]] = None,
):
"""
Fetches a list of threads based on pagination and optional filters.
Args:
first (Optional[int]): Number of threads to fetch.
after (Optional[str]): Cursor for pagination, fetch threads after this cursor.
before (Optional[str]): Cursor for pagination, fetch threads before this cursor.
filters (Optional[threads_filters]): Filters to apply on the threads query.
order_by (Optional[threads_order_by]): Order by clause for threads.
step_types_to_keep (Optional[List[StepType]]) : If set, only steps of the corresponding types
will be returned
Returns:
`PaginatedResponse[Thread]`: A paginated response containing the queried thread data.
"""
pass
@abstractmethod
def list_threads(
self,
first: Optional[int] = None,
after: Optional[str] = None,
before: Optional[str] = None,
filters: Optional[threads_filters] = None,
order_by: Optional[threads_order_by] = None,
):
"""
Lists threads based on pagination and optional filters, similar to get_threads but may include additional processing.
Args:
first (Optional[int]): Number of threads to list.
after (Optional[str]): Cursor for pagination, list threads after this cursor.
before (Optional[str]): Cursor for pagination, list threads before this cursor.
filters (Optional[threads_filters]): Filters to apply on the threads listing.
order_by (Optional[threads_order_by]): Order by clause for threads.
Returns:
`PaginatedResponse[Thread]`: A paginated response containing the queried thread data.
"""
pass
@abstractmethod
def get_thread(self, id: str):
"""
Retrieves a single thread by its ID.
Args:
id (str): The unique identifier of the thread.
Returns:
`Thread`: The thread corresponding to the provided ID.
"""
pass
@abstractmethod
def create_thread(
self,
name: Optional[str] = None,
metadata: Optional[Dict] = None,
participant_id: Optional[str] = None,
tags: Optional[List[str]] = None,
):
"""
Creates a `Thread` with the specified details.
Args:
name (Optional[str]): Name of the thread.
metadata (Optional[Dict]): Metadata associated with the thread.
participant_id (Optional[str]): Identifier for the participant.
tags (Optional[List[str]]): List of tags associated with the thread.
Returns:
`Thread`: The created thread.
"""
pass
@abstractmethod
def upsert_thread(
self,
id: str,
name: Optional[str] = None,
metadata: Optional[Dict] = None,
participant_id: Optional[str] = None,
tags: Optional[List[str]] = None,
):
"""
Updates an existing thread or creates a new one if it does not exist.
Args:
id (str): The unique identifier of the thread.
name (Optional[str]): Name of the thread.
metadata (Optional[Dict]): Metadata associated with the thread.
participant_id (Optional[str]): Identifier for the participant.
tags (Optional[List[str]]): List of tags associated with the thread.
Returns:
`Thread`: The updated or newly created thread.
"""
pass
@abstractmethod
def update_thread(
self,
id: str,
name: Optional[str] = None,
metadata: Optional[Dict] = None,
participant_id: Optional[str] = None,
tags: Optional[List[str]] = None,
):
"""
Updates the specified details of an existing thread.
Args:
id (str): The unique identifier of the thread to update.
name (Optional[str]): New name of the thread.
metadata (Optional[Dict]): New metadata for the thread.
participant_id (Optional[str]): New identifier for the participant.
tags (Optional[List[str]]): New list of tags for the thread.
Returns:
`Thread`: The updated thread.
"""
pass
@abstractmethod
def delete_thread(self, id: str):
"""
Deletes a thread identified by its ID.
Args:
id (str): The unique identifier of the thread to delete.
Returns:
`bool`: True if the thread was deleted, False otherwise.
"""
pass
@abstractmethod
def get_scores(
self,
first: Optional[int] = None,
after: Optional[str] = None,
before: Optional[str] = None,
filters: Optional[scores_filters] = None,
order_by: Optional[scores_order_by] = None,
):
"""
Fetches scores based on pagination and optional filters.
Args:
first (Optional[int]): The number of scores to retrieve.
after (Optional[str]): A cursor for use in pagination, fetching records after this cursor.
before (Optional[str]): A cursor for use in pagination, fetching records before this cursor.
filters (Optional[scores_filters]): Filters to apply to the scores query.
order_by (Optional[scores_order_by]): Ordering options for the scores.
Returns:
`PaginatedResponse[Score]`: A paginated response containing the queried scores.
"""
pass
@abstractmethod
def create_scores(self, scores: List[ScoreDict]):
"""
Creates multiple scores.
Args:
scores (List[ScoreDict]): A list of dictionaries representing the scores to be created.
Returns:
List[ScoreDict]: The created scores as a list of dictionaries.
"""
pass
@abstractmethod
def create_score(
self,
name: str,
value: float,
type: ScoreType,
step_id: Optional[str] = None,
generation_id: Optional[str] = None,
dataset_experiment_item_id: Optional[str] = None,
comment: Optional[str] = None,
tags: Optional[List[str]] = None,
):
"""
Creates a single score in the database.
Args:
name (str): The name of the score.
value (float): The numerical value of the score.
type (ScoreType): The type of the score.
step_id (Optional[str]): The ID of the step associated with the score.
generation_id (Optional[str]): The ID of the generation associated with the score.
dataset_experiment_item_id (Optional[str]): The ID of the dataset experiment item associated with the score.
comment (Optional[str]): An optional comment about the score.
tags (Optional[List[str]]): Optional tags associated with the score.
Returns:
`Score`: The created score.
"""
pass
@abstractmethod
def update_score(
self,
id: str,
update_params: ScoreUpdate,
):
"""
Updates a score identified by its ID with new parameters.
Args:
id (str): The unique identifier of the score to update.
update_params (ScoreUpdate): A dictionary of parameters to update in the score.
Returns:
`Score`: The updated score.
"""
pass
@abstractmethod
def delete_score(self, id: str):
"""
Deletes a score by ID.
Args:
id (str): ID of score to delete.
Returns:
Dict: The deleted `Score` as a dict.
"""
pass
@abstractmethod
def upload_file(
self,
content: Union[bytes, str],
thread_id: Optional[str] = None,
mime: Optional[str] = "application/octet-stream",
):
"""
Uploads a file to the server.
Args:
content (Union[bytes, str]): The content of the file to upload.
thread_id (Optional[str]): The ID of the thread associated with the file.
mime (Optional[str]): The MIME type of the file. Defaults to 'application/octet-stream'.
Returns:
Dict: A dictionary containing the object key and URL of the uploaded file, or None values if the upload fails.
"""
pass
@abstractmethod
def create_attachment(
self,
thread_id: Optional[str] = None,
step_id: Optional[str] = None,
id: Optional[str] = None,
metadata: Optional[Dict] = None,
mime: Optional[str] = None,
name: Optional[str] = None,
object_key: Optional[str] = None,
url: Optional[str] = None,
content: Optional[Union[bytes, str]] = None,
path: Optional[str] = None,
):
"""
Creates an attachment associated with a thread and step, potentially uploading file content.
Args:
thread_id (str): The ID of the thread to which the attachment is linked.
step_id (str): The ID of the step to which the attachment is linked.
id (Optional[str]): The ID of the attachment, if updating an existing one.
metadata (Optional[Dict]): Metadata associated with the attachment.
mime (Optional[str]): MIME type of the file, if content is provided.
name (Optional[str]): Name of the attachment.
object_key (Optional[str]): Object key of the uploaded file, if already known.
url (Optional[str]): URL of the uploaded file, if already known.
content (Optional[Union[bytes, str]]): File content to upload.
path (Optional[str]): Path where the file should be stored.
Returns:
`Attachment`: The created attachment.
"""
pass
@abstractmethod
def update_attachment(self, id: str, update_params: AttachmentUpload):
"""
Updates an existing attachment with new parameters.
Args:
id (str): The unique identifier of the attachment to update.
update_params (AttachmentUpload): The parameters to update in the attachment.
Returns:
`Attachment`: The updated attachment.
"""
pass
@abstractmethod
def get_attachment(self, id: str):
"""
Retrieves an attachment by ID.
Args:
id (str): ID of the attachment to retrieve.
Returns:
`Attachment`: The attachment object with requested ID.
"""
pass
@abstractmethod
def delete_attachment(self, id: str):
"""
Deletes an attachment identified by ID.
Args:
id (str): The unique identifier of the attachment to delete.
Returns:
`Attachment`: The deleted attachment.
"""
pass
@abstractmethod
def create_step(
self,
thread_id: Optional[str] = None,
type: Optional[StepType] = "undefined",
start_time: Optional[str] = None,
end_time: Optional[str] = None,
input: Optional[Dict] = None,
output: Optional[Dict] = None,
metadata: Optional[Dict] = None,
parent_id: Optional[str] = None,
name: Optional[str] = None,
tags: Optional[List[str]] = None,
root_run_id: Optional[str] = None,
):
"""
Creates a new step with the specified parameters.
Args:
thread_id (Optional[str]): The ID of the thread this step is associated with.
type (Optional[StepType]): The type of the step, defaults to "undefined".
start_time (Optional[str]): The start time of the step.
end_time (Optional[str]): The end time of the step.
input (Optional[Dict]): Input data for the step.
output (Optional[Dict]): Output data from the step.
metadata (Optional[Dict]): Metadata associated with the step.
parent_id (Optional[str]): The ID of the parent step, if any.
name (Optional[str]): The name of the step.
tags (Optional[List[str]]): Tags associated with the step.
root_run_id (Optional[str]): The ID of the root run, if any.
Returns:
`Step`: The created step.
"""
pass
@abstractmethod
def update_step(
self,
id: str,
type: Optional[StepType] = None,
input: Optional[str] = None,
output: Optional[str] = None,
metadata: Optional[Dict] = None,
name: Optional[str] = None,
tags: Optional[List[str]] = None,
start_time: Optional[str] = None,
end_time: Optional[str] = None,
parent_id: Optional[str] = None,
):
"""
Updates an existing step identified by its ID with new parameters.
Args:
id (str): The unique identifier of the step to update.
type (Optional[StepType]): The type of the step.
input (Optional[str]): Input data for the step.
output (Optional[str]): Output data from the step.
metadata (Optional[Dict]): Metadata associated with the step.
name (Optional[str]): The name of the step.
tags (Optional[List[str]]): Tags associated with the step.
start_time (Optional[str]): The start time of the step.
end_time (Optional[str]): The end time of the step.
parent_id (Optional[str]): The ID of the parent step, if any.
Returns:
`Step`: The updated step.
"""
pass
@abstractmethod
def get_steps(
self,
first: Optional[int] = None,
after: Optional[str] = None,
before: Optional[str] = None,
filters: Optional[steps_filters] = None,
order_by: Optional[steps_order_by] = None,
):
"""
Fetches a list of steps based on pagination and optional filters.
Args:
first (Optional[int]): Number of steps to fetch.
after (Optional[str]): Cursor for pagination, fetch steps after this cursor.
before (Optional[str]): Cursor for pagination, fetch steps before this cursor.
filters (Optional[steps_filters]): Filters to apply on the steps query.
order_by (Optional[steps_order_by]): Order by clause for steps.
Returns:
`PaginatedResponse[Step]`: The list of steps matching the criteria.
"""
pass
@abstractmethod
def get_step(
self,
id: str,
):
"""
Retrieves a step by its ID.
Args:
id (str): The unique identifier of the step to retrieve.
Returns:
`Step`: The step with requested ID.
"""
pass
@abstractmethod
def delete_step(
self,
id: str,
):
"""
Deletes a step identified by its ID.
Args:
id (str): The unique identifier of the step to delete.
Returns:
`bool`: True if the step was deleted successfully, False otherwise.
"""
pass
@abstractmethod
def send_steps(self, steps: List[Union[StepDict, "Step"]]):
"""
Sends a list of steps to process.
Step ingestion happens asynchronously if you configured a cache. See [Cache Configuration](https://docs.literalai.com/self-hosting/deployment#4-cache-configuration-optional).
Args:
steps (List[Union[StepDict, Step]]): A list of steps or step dictionaries to send.
Returns:
`Dict`: Dictionary with keys "ok" (boolean) and "message" (string).
"""
pass
@abstractmethod
def get_generations(
self,
first: Optional[int] = None,
after: Optional[str] = None,
before: Optional[str] = None,
filters: Optional[generations_filters] = None,
order_by: Optional[generations_order_by] = None,
):
"""
Fetches a list of generations based on pagination and optional filters.
Args:
first (Optional[int]): The number of generations to retrieve.
after (Optional[str]): A cursor for use in pagination, fetching records after this cursor.
before (Optional[str]): A cursor for use in pagination, fetching records before this cursor.
filters (Optional[generations_filters]): Filters to apply to the generations query.
order_by (Optional[generations_order_by]): Order by clause for generations.
Returns:
A list of generations that match the criteria.
"""
pass
@abstractmethod
def create_generation(
self, generation: Union[ChatGeneration, CompletionGeneration]
):
"""
Creates a new generation, either a chat or completion type.
```py
from literalai.observability.generation import ChatGeneration
from literalai import LiteralClient
literalai_client = LiteralClient(api_key="lsk-***")
example_generation = ChatGeneration(
messages=[
{
"role": "user",
"content": "Hello, how can I help you today?"
},
],
message_completion={
"role": "assistant",
"content": "Sure, I can help with that. What do you need to know?"
},
model="gpt-4o-mini",
provider="OpenAI"
)
literalai_client.api.create_generation(example_generation)
```
Args:
generation (Union[ChatGeneration, CompletionGeneration]): The generation data to create.
Returns:
`Union[ChatGeneration, CompletionGeneration]`: The created generation, either a chat or completion type.
"""
pass
@abstractmethod
def create_dataset(
self,
name: str,
description: Optional[str] = None,
metadata: Optional[Dict] = None,
type: DatasetType = "key_value",
):
"""
Creates a new dataset with the specified properties.
Args:
name (str): The name of the dataset.
description (Optional[str]): A description of the dataset.
metadata (Optional[Dict]): Additional metadata for the dataset.
type (DatasetType): The type of the dataset, defaults to "key_value".
Returns:
`Dataset`: The created dataset, empty initially.
"""
pass
@abstractmethod
def get_dataset(self, id: Optional[str] = None, name: Optional[str] = None):
"""
Retrieves a dataset by its ID or name.
Args:
id (Optional[str]): The unique identifier of the dataset.
name (Optional[str]): The name of the dataset.
Returns:
The dataset data as returned by the REST helper function.
"""
pass
@abstractmethod
def update_dataset(
self,
id: str,
name: Optional[str] = None,
description: Optional[str] = None,
metadata: Optional[Dict] = None,
):
"""
Updates an existing dataset identified by its ID with new properties.
Args:
id (str): The unique identifier of the dataset to update.
name (Optional[str]): A new name for the dataset.
description (Optional[str]): A new description for the dataset.
metadata (Optional[Dict]): New or updated metadata for the dataset.
Returns:
`Dataset`: The dataset with updated properties.
"""
pass
@abstractmethod
def delete_dataset(self, id: str):
"""
Deletes a dataset identified by its ID.
Args:
id (str): The unique identifier of the dataset to delete.
Returns:
`Dataset`: The deleted dataset.
"""
pass
@abstractmethod
def create_experiment(
self,
name: str,
dataset_id: Optional[str] = None,
prompt_variant_id: Optional[str] = None,
params: Optional[Dict] = None,
):
"""
Creates a new experiment associated with a specific dataset.
Args:
name (str): The name of the experiment.
dataset_id (Optional[str]): The unique identifier of the dataset.
prompt_variant_id (Optional[str]): The identifier of the prompt variant to associate to the experiment.
params (Optional[Dict]): Additional parameters for the experiment.
Returns:
DatasetExperiment: The newly created experiment object.
"""
pass
@abstractmethod
def create_experiment_item(self, experiment_item: DatasetExperimentItem):
"""
Creates an experiment item within an existing experiment.
Args:
experiment_item (DatasetExperimentItem): The experiment item to be created, containing all necessary data.
Returns:
DatasetExperimentItem: The newly created experiment item with scores attached.
"""
pass
@abstractmethod
def create_dataset_item(
self,
dataset_id: str,
input: Dict,
expected_output: Optional[Dict] = None,
metadata: Optional[Dict] = None,
):
"""
Creates an item in a dataset with the specified properties.
Args:
dataset_id (str): The unique identifier of the dataset.
input (Dict): The input data for the dataset item.
expected_output (Optional[Dict]): The expected output data for the dataset item.
metadata (Optional[Dict]): Additional metadata for the dataset item.
Returns:
DatasetItem: The created dataset item.
"""
pass
@abstractmethod
def get_dataset_item(self, id: str):
"""
Retrieves a dataset item by ID.
Args:
id (str): ID of the `DatasetItem` to retrieve.
Returns:
`DatasetItem`: The dataset item.
"""
pass
@abstractmethod
def delete_dataset_item(self, id: str):
"""
Deletes a dataset item by ID.
Args:
id (str): ID of the dataset item to delete.
Returns:
`DatasetItem`: The deleted item.
"""
pass
@abstractmethod
def add_step_to_dataset(
self, dataset_id: str, step_id: str, metadata: Optional[Dict] = None
):
"""
Adds a step to a dataset.
Args:
dataset_id (str): The unique identifier of the dataset.
step_id (str): The unique identifier of the step to add.
metadata (Optional[Dict]): Additional metadata related to the step to add.
Returns:
Dict: The created `DatasetItem`.
"""
pass
@abstractmethod
def add_generation_to_dataset(
self, dataset_id: str, generation_id: str, metadata: Optional[Dict] = None
):
"""
Adds a generation to a dataset.
Args:
dataset_id (str): The unique identifier of the dataset.
generation_id (str): The unique identifier of the generation to add.
metadata (Optional[Dict]): Additional metadata related to the generation to add.
Returns:
Dict: The created `DatasetItem`.
"""
pass
@abstractmethod
@deprecated("Use get_or_create_prompt_lineage instead")
def create_prompt_lineage(self, name: str, description: Optional[str] = None):
"""
Deprecated. Please use **get_or_create_prompt_lineage** instead.
"""
pass
@abstractmethod
def get_or_create_prompt_lineage(
self, name: str, description: Optional[str] = None
):
"""
Creates a prompt lineage with the specified name and optional description.
If the prompt lineage with that name already exists, it is returned.
Args:
name (str): The name of the prompt lineage.
description (Optional[str]): An optional description of the prompt lineage.
Returns:
`Dict`: The prompt lineage data as a dictionary.
"""
pass
@abstractmethod
@deprecated("Please use `get_or_create_prompt` instead.")
def create_prompt(
self,
name: str,
template_messages: List[GenerationMessage],
settings: Optional[ProviderSettings] = None,
):
"""
Deprecated. Please use `get_or_create_prompt` instead.
"""
pass
@abstractmethod
def get_or_create_prompt(
self,
name: str,
template_messages: List[GenerationMessage],
settings: Optional[ProviderSettings] = None,
tools: Optional[List[Dict]] = None,
):
"""
A `Prompt` is fully defined by its `name`, `template_messages`, `settings` and tools.
If a prompt already exists for the given arguments, it is returned.
Otherwise, a new prompt is created.
Args:
name (str): The name of the prompt to retrieve or create.
template_messages (List[GenerationMessage]): A list of template messages for the prompt.
settings (Optional[Dict]): Optional settings for the prompt.
tools (Optional[List[Dict]]): Optional tool options for the model
Returns:
Prompt: The prompt that was retrieved or created.
"""
pass
@abstractmethod
def get_prompt(
self,
id: Optional[str] = None,
name: Optional[str] = None,
version: Optional[int] = None,
):
"""
Gets a prompt either by:
- `id`
- `name` and (optional) `version`
At least the `id` or the `name` must be passed to the function.
If both are provided, the `id` is used.
Args:
id (str): The unique identifier of the prompt to retrieve.
name (str): The name of the prompt to retrieve.
version (Optional[int]): The version number of the prompt to retrieve.
Returns:
Prompt: The prompt with the given identifier or name.
"""
pass
@abstractmethod
def create_prompt_variant(
self,
name: str,
template_messages: List[GenerationMessage],
settings: Optional[ProviderSettings] = None,
tools: Optional[List[Dict]] = None,
):
"""
Creates a prompt variant to use as part of an experiment.
This variant is not an official Prompt version until manually saved.
Args:
name (str): Name of the variant to create.
template_messages (List[GenerationMessage]): A list of template messages for the prompt.
settings (Optional[Dict]): Optional settings for the prompt.
tools (Optional[List[Dict]]): Optional tools for the model.
Returns:
prompt_variant_id: The ID of the created prompt variant id, which you can link to an experiment.
"""
pass
@abstractmethod
def get_prompt_ab_testing(self, name: str):
"""
Get the A/B testing configuration for a prompt lineage.
Args:
name (str): The name of the prompt lineage.
Returns:
List[PromptRollout]
"""
pass
@abstractmethod
def update_prompt_ab_testing(self, name: str, rollouts: List[PromptRollout]):
"""
Update the A/B testing configuration for a prompt lineage.
Args:
name (str): The name of the prompt lineage.
rollouts (List[PromptRollout]): The percentage rollout for each prompt version.
Returns:
Dict
"""
pass
@abstractmethod
def get_my_project_id(self):
"""
Retrieves the project ID associated with the API key.
Returns:
`str`: The project ID for current API key.
"""
pass