Create a LLM skeleton — skellm

This class handles the creation of LLM skeletons

Usage

skellm_skeleton(
  url = class_missing,
  model = class_missing,
  history = class_missing,
  stream = class_missing,
  api_key = class_missing,
  extra = class_missing
)

Arguments

url: The host's URL.
model: Identifier of the model to be used.
history: Messages to be included in the request. By default, is an empty skellm_message_history S7 object.
stream: Whether or not to stream the incomming response. FALSE by default.
api_key: (Optional) The API key of the service.

Value

An object with class skellm_skeleton.

Examples

# For Ollama
skellm_skeleton(
  url = "http://localhost:11434/api/chat", 
  model = "mistral"
)
#> <skellm::skellm_skeleton>
#>  @ url    : chr "http://localhost:11434/api/chat"
#>  @ model  : chr "mistral"
#>  @ history: <skellm_message_history>  list()
#>  @ stream : logi FALSE
#>  @ api_key: NULL
#>  @ extra  : list()

# For OpenAI
skellm_skeleton(
  url = "https://api.openai.com/v1/chat/completions", 
  model = "gpt-3.5-turbo",
  history = skellm_message_history(
    skellm_message(role = "system", content = "You are a helpful assistant"),
    skellm_message(role = "user", content = "What is 2 + 2")
  ),
  stream = TRUE,
  api_key = "<your_api_key>"
)
#> <skellm::skellm_skeleton>
#>  @ url    : chr "https://api.openai.com/v1/chat/completions"
#>  @ model  : chr "gpt-3.5-turbo"
#>  @ history: <skellm_message_history> List of 2
#>  .. $ : <skellm_message>
#>  ..  ..@ role   : chr "system"
#>  ..  ..@ content: chr "You are a helpful assistant"
#>  ..  ..@ name   : NULL
#>  .. $ : <skellm_message>
#>  ..  ..@ role   : chr "user"
#>  ..  ..@ content: chr "What is 2 + 2"
#>  ..  ..@ name   : NULL
#>  @ stream : logi TRUE
#>  @ api_key: chr "<your_api_key>"
#>  @ extra  : list()