defmodule LangChain.Function do @moduledoc """ Defines a "function" that can be provided to an LLM for the LLM to optionally execute and pass argument data to. A function is defined using a schema. * `name` - The name of the function given to the LLM. * `description` - A description of the function provided to the LLM. This should describe what the function is used for or what it returns. This information is used by the LLM to decide which function to call and for what purpose. * ` parameters` - A list of `Function.FunctionParam` structs that are converted to a JSONSchema format. (Use in place of `parameters_schema`) * ` parameters_schema` - A [JSONSchema structure](https://json-schema.org/learn/getting-started-step-by-step.html) that describes the required data structure format for how arguments are passed to the function. (Use if greater control or unsupported features are needed.) * `function` - An Elixir function to execute when an LLM requests to execute the function. The function can return a string, a tuple, or a `%ToolResult{}` struct for advanced control. Returning a ToolResult allows for multi-modal responses (list of ContentParts), cache control, and processed_content. * `async` - Boolean value that flags if this can function can be executed asynchronously, potentially concurrently with other calls to the same function. Defaults to `false`. * `options` - A Keyword list of options that can be passed to the LLM. For example, this can be used for passing caching config to Anthropic. When passing arguments from an LLM to a function, they go through a single `map` argument. This allows for multiple keys or named parameters. ## Example This example defines a function that an LLM can execute for performing basic math calculations. **NOTE:** This is a partial implementation of the `LangChain.Tools.Calculator`. Function.new(%{ name: "calculator", description: "Perform basic math calculations", parameters_schema: %{ type: "object", properties: %{ expression: %{type: "string", description: "A simple mathematical expression."} }, required: ["expression"] }, function: fn(%{"expression" => expr} = _args, _context) -> {:ok, "42?"} end) }) The `function` attribute is an Elixir function that can be executed when the function is "called" by the LLM. The `args` argument is the JSON data passed by the LLM after being parsed to a map. The `context` argument is passed through as the `context` on a `LangChain.Chains.LLMChain`. This is whatever context data is needed for the function to do it's work. Context examples could be data like user_id, account_id, account struct, billing level, etc. ## Function Parameters The `parameters` field is a list of `LangChain.FunctionParam` structs. This is a convenience for defining the parameters to the function. If it does not work for more complex use-cases, then use the `parameters_schema` to declare it as needed. The `parameters_schema` is an Elixir map that follows a [JSONSchema](https://json-schema.org/learn/getting-started-step-by-step.html) structure. It is used to define the required data structure format for receiving data to the function from the LLM. NOTE: Only use `parameters` or `parameters_schema`, not both. ## Expanded Parameter Examples Function with no arguments: alias LangChain.Function Function.new!(%{name: "get_current_user_info"}) Function that takes a simple required argument: alias LangChain.FunctionParam Function.new!(%{name: "set_user_name", parameters: [ FunctionParam.new!(%{name: "user_name", type: :string, required: true}) ]}) Function that takes an array of strings: Function.new!(%{name: "set_tags", parameters: [ FunctionParam.new!(%{name: "tags", type: :array, item_type: "string"}) ]}) Function that takes two arguments and one is an object/map: Function.new!(%{name: "update_preferences", parameters: [ FunctionParam.new!(%{name: "unique_code", type: :string, required: true}) FunctionParam.new!(%{name: "data", type: :object, object_properties: [ FunctionParam.new!(%{name: "auto_complete_email", type: :boolean}), FunctionParam.new!(%{name: "items_per_page", type: :integer}), ]}) ]}) The `LangChain.FunctionParam` is nestable allowing for arrays of object and objects with nested objects. ## Example that also stores the Elixir result Sometimes we want to process a `ToolCall` from the LLM and keep the processed Elixir data for ourselves. This is particularly useful when using an LLM to perform structured data extraction. Our Elixir function may even process that data into a newly created Ecto Schema database entry. The result of the `ToolCall` that goes back to the LLM must be in a string form. That typically means returning a JSON string of the result data. To make it easier to process the data, return a string response to the LLM, but **keep** the original Elixir data as well, our Elixir function can return a 3-tuple result. Function.new!(%{name: "create_invoice", parameters: [ FunctionParam.new!(%{name: "vendor_name", type: :string, required: true}) FunctionParam.new!(%{name: "total_amount", type: :string, required: true}) ], function: &execute_create_invoice/2 }) # ... def execute_create_invoice(args, %{account_id: account_id} = _context) do case MyApp.Invoices.create_invoice(account_id, args) do {:ok, invoice} -> {:ok, "SUCCESS", invoice} {:error, changeset} -> {:error, "ERROR: " <> LangChain.Utils.changeset_error_to_string(changeset)} end end In this example, the `LangChain.Function` is tied to the `MyApp.Invoices.create_invoice/2` function in our application. The Elixir function returns a 3-tuple result. The `"SUCCESS"` is returned to the LLM. In our scenario, we don't care to return a JSON version of the invoice. The important part is we return the actual `%MyApp.Invoices.Invoice{}` struct in the tuple. This is stored on the `LangChain.ToolResult`'s `processed_content` field. This is really helpful when all we want is the final, fully processed Elixir result. This pairs well with the `LLMChain.run(chain, mode: :until_success)`. This is when we want the LLM to perform some data extraction and it should be re-run until it succeeds and we have our final, processed result in the `ToolResult`. Note: The LLM may issue one or more `ToolCall`s in a single assistant message. Each Elixir function's `ToolResult` may contain a `processed_content`. ## Explicit ToolResult Control For advanced use cases where you need explicit control over the `ToolResult` structure or want to set LLM-specific options, your Elixir function can return a fully constructed `%ToolResult{}` struct. The `content` field can be a list of ContentParts for multi-modal responses. This approach is particularly useful when you need to: * Set LLM-specific options (like Anthropic's `cache_control`) * Set custom error states beyond simple string responses * Customize the `display_text` for the ToolResult * Provide detailed metadata in the `options` field The `options` field can contain any LLM-specific configuration that gets passed through to the chat model's API conversion layer. If an LLM does not support it, it will be ignored. """ use Ecto.Schema import Ecto.Changeset require Logger alias __MODULE__ alias LangChain.LangChainError @primary_key false embedded_schema do field :name, :string field :description, :string # Optional text the UI can display for when the function is executed. field :display_text, :string # Optional flag to indicate if the function should be executed in strict mode. # Defaults to `false`. field :strict, :boolean, default: false # flag if the function should be auto-evaluated. Defaults to `false` # requiring an explicit step to perform the evaluation. # field :auto_evaluate, :boolean, default: false field :function, :any, virtual: true # Track if the function can be executed async. Defaults to `false`. field :async, :boolean, default: false # parameters_schema is a map used to express a JSONSchema structure of inputs and what's required field :parameters_schema, :map # parameters is a list of `LangChain.FunctionParam` structs. field :parameters, {:array, :any}, default: [] field :options, :any, virtual: true, default: [] end @type t :: %Function{} @type arguments :: %{String.t() => any()} @type context :: nil | %{atom() => any()} @create_fields [ :name, :description, :display_text, :strict, :parameters_schema, :parameters, :function, :async, :options ] @required_fields [:name] @doc """ Build a new function. """ @spec new(attrs :: map()) :: {:ok, t()} | {:error, Ecto.Changeset.t()} def new(attrs \\ %{}) do %Function{} |> cast(attrs, @create_fields) |> common_validation() |> apply_action(:insert) end @doc """ Build a new function and return it or raise an error if invalid. """ @spec new!(attrs :: map()) :: t() | no_return() def new!(attrs \\ %{}) do case new(attrs) do {:ok, function} -> function {:error, changeset} -> raise LangChainError, changeset end end @spec common_validation(Ecto.Changeset.t()) :: Ecto.Changeset.t() defp common_validation(changeset) do changeset |> validate_required(@required_fields) |> validate_length(:name, max: 64) |> validate_parameter_exclusivity() |> validate_function_arity() end @doc """ Execute the function passing in arguments and additional optional context. This is called by a `LangChain.Chains.LLMChain` when a `Function` execution is requested by the LLM. """ @spec execute(t(), arguments(), context()) :: any() | no_return() def execute(%Function{function: fun} = function, arguments, context) do Logger.debug("Executing function #{inspect(function.name)}") with :ok <- validate_required_params(function, arguments) do execute_with_error_handling(function, fun, arguments, context) end end @doc """ Given a list of functions, return the `display_text` for the named function. If it not found, return the fallback text. """ @spec get_display_text([t()], String.t(), String.t()) :: String.t() def get_display_text(functions, function_name, fallback_text \\ "Perform action") def get_display_text(functions, function_name, fallback_text) do case Enum.find(functions, &(&1.name == function_name)) do nil -> fallback_text %Function{display_text: display_text} -> display_text end end # Validates that the function field contains a function with arity 2 @spec validate_function_arity(Ecto.Changeset.t()) :: Ecto.Changeset.t() defp validate_function_arity(changeset) do changeset |> get_field(:function) |> do_validate_function_arity(changeset) end @spec do_validate_function_arity(any(), Ecto.Changeset.t()) :: Ecto.Changeset.t() defp do_validate_function_arity(function, changeset) when is_function(function, 2) do changeset end defp do_validate_function_arity(function, changeset) when is_function(function) do {:arity, arity} = Elixir.Function.info(function, :arity) add_error(changeset, :function, "expected arity of 2 but has arity #{inspect(arity)}") end defp do_validate_function_arity(_not_a_function, changeset) do add_error(changeset, :function, "is not an Elixir function") end # Validates that only one of parameters or parameters_schema is provided @spec validate_parameter_exclusivity(Ecto.Changeset.t()) :: Ecto.Changeset.t() defp validate_parameter_exclusivity(changeset) do params_list = get_field(changeset, :parameters) schema_map = get_field(changeset, :parameters_schema) do_validate_parameter_exclusivity(changeset, params_list, schema_map) end @spec do_validate_parameter_exclusivity(Ecto.Changeset.t(), list(), map() | nil) :: Ecto.Changeset.t() defp do_validate_parameter_exclusivity(changeset, params, schema) when is_map(schema) and is_list(params) and params != [] do add_error(changeset, :parameters, "Cannot use both parameters and parameters_schema") end defp do_validate_parameter_exclusivity(changeset, _params, _schema), do: changeset @spec execute_with_error_handling(t(), function(), arguments(), context()) :: {:ok, any()} | {:ok, any(), any()} | {:error, String.t()} defp execute_with_error_handling(function, fun, arguments, context) do fun.(arguments, context) |> normalize_execution_result(function) rescue err -> Logger.error( "Function! #{function.name} failed in execution. Exception: #{LangChainError.format_exception(err, __STACKTRACE__)}" ) {:error, "ERROR: #{LangChainError.format_exception(err, __STACKTRACE__, :short)}"} end # Normalizes the various return types from function execution into consistent tagged tuples @spec normalize_execution_result(any(), t()) :: {:ok, any()} | {:ok, any(), any()} | {:error, String.t()} defp normalize_execution_result({:ok, llm_result, processed_content}, _function) do {:ok, llm_result, processed_content} end defp normalize_execution_result({:ok, result}, _function) do {:ok, result} end defp normalize_execution_result({:interrupt, message, data}, _function) when is_binary(message) do {:interrupt, message, data} end defp normalize_execution_result({:error, reason}, _function) when is_binary(reason) do {:error, reason} end defp normalize_execution_result({:error, reason}, _function) do {:error, "#{inspect(reason)}"} end defp normalize_execution_result(text, _function) when is_binary(text) do {:ok, text} end defp normalize_execution_result(parts, _function) when is_list(parts) do {:ok, parts} end defp normalize_execution_result(other, function) do Logger.warning( "Function #{function.name} unexpectedly returned #{inspect(other)}. Expect a string. Unable to present as response to LLM." ) {:error, "An unexpected response was returned from the tool."} end # Validates that all required parameters are present in the arguments @spec validate_required_params(t(), arguments()) :: :ok | {:error, String.t()} defp validate_required_params(%Function{parameters: params}, arguments) when is_list(params) and params != [] do params |> collect_required_param_names() |> find_missing_params(arguments) |> format_missing_params_result() end defp validate_required_params(_function, _arguments), do: :ok # Extracts names of required parameters from the parameter list @spec collect_required_param_names([struct()]) :: [String.t()] defp collect_required_param_names(params) do Enum.reduce(params, [], fn param, acc -> if param.required, do: [param.name | acc], else: acc end) end # Finds parameters that are required but missing from the arguments @spec find_missing_params([String.t()], arguments()) :: [String.t()] defp find_missing_params([], _arguments), do: [] defp find_missing_params(required_names, arguments) do Enum.reject(required_names, &Map.has_key?(arguments, &1)) end # Formats the result based on whether any parameters are missing @spec format_missing_params_result([String.t()]) :: :ok | {:error, String.t()} defp format_missing_params_result([]), do: :ok @missing_params_error_template """ Missing required parameters for this tool. Required parameters: ~s Ensure you're passing the correct parameter names as defined in the tool schema. """ defp format_missing_params_result(missing_params) do expected = Enum.join(missing_params, ", ") {:error, :io_lib.format(@missing_params_error_template, [expected]) |> IO.iodata_to_binary()} end end