gpt_engineer.core.ai.ClipboardAI

class gpt_engineer.core.ai.ClipboardAI(**_)[source]

Bases: AI

Initialize the AI class.

Parameters:

  • model_name (str, optional) – The name of the model to use, by default “gpt-4”.

  • temperature (float, optional) – The temperature to use for the model, by default 0.1.

backoff_inference(messages)

Perform inference using the language model while implementing an exponential backoff strategy.

This function will retry the inference in case of a rate limit error from the OpenAI API. It uses an exponential backoff strategy, meaning the wait time between retries increases exponentially. The function will attempt to retry up to 7 times within a span of 45 seconds.

Parameters:

  • messages (List[Message]) – A list of chat messages which will be passed to the language model for processing.

  • callbacks (List[Callable]) – A list of callback functions that are triggered after each inference. These functions can be used for logging, monitoring, or other auxiliary tasks.

Returns:

The output from the language model after processing the provided messages.

Return type:

Any

Raises:

openai.error.RateLimitError – If the number of retries exceeds the maximum or if the rate limit persists beyond the allotted time, the function will ultimately raise a RateLimitError.

Example

>>> messages = [SystemMessage(content="Hello"), HumanMessage(content="How's the weather?")]
>>> response = backoff_inference(messages)
static deserialize_messages(jsondictstr: str) → List[AIMessage | HumanMessage | SystemMessage]

Deserialize a JSON string to a list of messages.

Parameters:

jsondictstr (str) – The JSON string to deserialize.

Returns:

The deserialized list of messages.

Return type:

List[Message]

static multiline_input()[source]
next(messages: List[AIMessage | HumanMessage | SystemMessage], prompt: str | None = None, *, step_name: str) → List[AIMessage | HumanMessage | SystemMessage][source]

Not yet fully supported

static serialize_messages(messages: List[AIMessage | HumanMessage | SystemMessage]) → str[source]

Serialize a list of messages to a JSON string.

Parameters:

messages (List[Message]) – The list of messages to serialize.

Returns:

The serialized messages as a JSON string.

Return type:

str

start(system: str, user: Any, *, step_name: str) → List[AIMessage | HumanMessage | SystemMessage]

Start the conversation with a system message and a user message.

Parameters:

  • system (str) – The content of the system message.

  • user (str) – The content of the user message.

  • step_name (str) – The name of the step.

Returns:

The list of messages in the conversation.

Return type:

List[Message]