Chat Models¶

aana.core.models.chat ¶

Role ¶

Role = Literal['system', 'user', 'assistant']

The role of a participant in a conversation.

"system": Used for instructions or context provided to the model.
"user": Represents messages from the user.
"assistant": Represents LLM responses.

Prompt ¶

Prompt = str

The prompt for the LLM.

Question ¶

Question = str

The question.

ChatMessage ¶

Bases: BaseModel

A chat message.

ATTRIBUTE	DESCRIPTION
`content`	the text of the message TYPE: `str`
`role`	the role of the message TYPE: `Role`

ChatDialog ¶

Bases: BaseModel

A chat dialog.

ATTRIBUTE	DESCRIPTION
`messages`	the messages in the dialog TYPE: `list[ChatMessage]`

from_list ¶

from_list(messages)

Create a ChatDialog from a list of dictionaries.

PARAMETER	DESCRIPTION
`messages`	the list of messages TYPE: `list[dict[str, str]]`

RETURNS	DESCRIPTION
`ChatDialog`	the chat dialog TYPE: `ChatDialog`

Source code in aana/core/models/chat.py

@classmethod
def from_list(cls, messages: list[dict[str, str]]) -> "ChatDialog":
    """Create a ChatDialog from a list of dictionaries.

    Args:
        messages (list[dict[str, str]]): the list of messages

    Returns:
        ChatDialog: the chat dialog
    """
    return ChatDialog(messages=[ChatMessage(**message) for message in messages])

from_prompt ¶

from_prompt(prompt)

Create a ChatDialog from a prompt.

PARAMETER	DESCRIPTION
`prompt`	the prompt TYPE: `str`

RETURNS	DESCRIPTION
`ChatDialog`	the chat dialog TYPE: `ChatDialog`

Source code in aana/core/models/chat.py

@classmethod
def from_prompt(cls, prompt: str) -> "ChatDialog":
    """Create a ChatDialog from a prompt.

    Args:
        prompt (str): the prompt

    Returns:
        ChatDialog: the chat dialog
    """
    return ChatDialog(messages=[ChatMessage(content=prompt, role="user")])

ChatCompletionRequest ¶

Bases: BaseModel

A chat completion request for OpenAI compatible API.

ATTRIBUTE	DESCRIPTION
`model`	the model name (name of the LLM deployment) TYPE: `str`
`messages`	a list of messages comprising the conversation so far TYPE: `list[ChatMessage]`
`temperature`	float that controls the randomness of the sampling TYPE: `float`
`top_p`	float that controls the cumulative probability of the top tokens to consider TYPE: `float`
`max_tokens`	the maximum number of tokens to generate TYPE: `int`
`repetition_penalty`	float that penalizes new tokens based on whether they appear in the prompt and the generated text so far TYPE: `float`
`stream`	if set, partial message deltas will be sent TYPE: `bool`

ChatCompletionChoice ¶

Bases: BaseModel

A chat completion choice for OpenAI compatible API.

ATTRIBUTE	DESCRIPTION
`index`	the index of the choice in the list of choices TYPE: `int`
`message`	a chat completion message generated by the model TYPE: `ChatMessage`

ChatCompletion ¶

Bases: BaseModel

A chat completion for OpenAI compatible API.

ATTRIBUTE	DESCRIPTION
`id`	a unique identifier for the chat completion TYPE: `str`
`model`	the model used for the chat completion TYPE: `str`
`created`	the Unix timestamp (in seconds) of when the chat completion was created TYPE: `int`
`choices`	a list of chat completion choices TYPE: `list[ChatCompletionChoice]`
`object`	the object type, which is always `chat.completion` TYPE: `Literal['chat.completion']`