Capture and pass random seed (#28)

Defaults to set seed for testing to minimise spurious differences. Also fleshes out a bunch of other API parameters, which I _think_ will be useful for thinking about the API more broadly.

Fixes #27

R/chat.R

@@ -24,6 +24,10 @@ NULL
 #' @param model The model to use for the chat; set to `NULL` (the default) to
 #'   use a reasonable model, currently `gpt-4o-mini`. We strongly recommend
 #'   explicitly choosing a model for all but the most casual use.
+#' @param seed Optional integer seed that ChatGPT uses to try and make output
+#'   more reproducible.
+#' @param api_args Named list of arbitrary extra arguments passed to every
+#'   chat API call.
 #' @param echo If `TRUE`, the `chat()` method streams the response to stdout by
 #'   default. (Note that this has no effect on the `stream()`, `chat_async()`,
 #'   and `stream_async()` methods.)
@@ -64,25 +68,32 @@ new_chat_openai <- function(system_prompt = NULL,
                             base_url = "https://api.openai.com/v1",
                             api_key = openai_key(),
                             model = NULL,
+                            seed = NULL,
+                            api_args = list(),
                             echo = FALSE) {
   check_string(system_prompt, allow_null = TRUE)
   check_openai_conversation(messages, allow_null = TRUE)
-    check_string(base_url)
+  check_string(base_url)
   check_string(api_key)
   check_string(model, allow_null = TRUE, allow_na = TRUE)
+  check_number_decimal(seed, allow_null = TRUE)
   check_bool(echo)
 
   model <- model %||% "gpt-4o-mini"
+  if (is_testing() && is.null(seed)) {
+    seed <- 1014
+  }
 
-  messages <- apply_system_prompt_openai(system_prompt, messages)
-
-  ChatOpenAI$new(
+  model <- openai_model(
     base_url = base_url,
     model = model,
-    messages = messages,
-    api_key = api_key,
-    echo = echo
+    seed = seed,
+    extra_args = api_args,
+    api_key = api_key
   )
+
+  messages <- apply_system_prompt_openai(system_prompt, messages)
+  ChatOpenAI$new(model = model, messages = messages, echo = echo)
 }
 
 apply_system_prompt_openai <- function(system_prompt, messages) {
@@ -138,23 +149,17 @@ check_openai_conversation <- function(messages, allow_null = FALSE) {
 #' @rdname new_chat_openai
 ChatOpenAI <- R6::R6Class("ChatOpenAI",
   public = list(
-    #' @param base_url The base URL to the endpoint; the default uses ChatGPT.
-    #' @param api_key The API key to use for authentication. You generally should
-    #'   not supply this directly, but instead set the `OPENAI_API_KEY` environment
-    #'   variable.
-    #' @param model The model to use for the chat; defaults to GPT-4o mini.
+    #' @param model Model object.
     #' @param messages An unnamed list of messages to start the chat with (i.e.,
     #'   continuing a previous conversation). If `NULL` or zero-length list, the
     #'   conversation begins from scratch.
+    #' @param seed Optional integer seed that ChatGPT uses to try and make output
+    #'   more reproducible.
     #' @param echo If `TRUE`, the `chat()` method streams the response to stdout
     #'   (while also returning the final response). Note that this has no effect
     #'   on the `stream()`, `chat_async()`, and `stream_async()` methods.
-    initialize = function(base_url, model, messages, api_key, echo = FALSE) {
-      private$model <- openai_model(
-        base_url = base_url,
-        model = model,
-        api_key = api_key
-      )
+    initialize = function(model, messages, seed = NULL, echo = FALSE) {
+      private$model <- model
       private$msgs <- messages %||% list()
       private$echo <- echo
     },

R/http.R

@@ -1,4 +1,7 @@
 chat_stream <- function(is_done, parse_data) {
+  # silence R CMD check note
+  yield <- NULL
+
   force(is_done)
   force(parse_data)
 
@@ -26,6 +29,8 @@ chat_stream <- function(is_done, parse_data) {
 }
 
 chat_stream_async <- function(is_done, parse_data, polling_interval_secs = 0.1) {
+  # silence R CMD check note
+  yield <- await <- NULL
   force(is_done)
   force(parse_data)
   force(polling_interval_secs)

R/open-ai.R

@@ -1,11 +1,15 @@
 openai_model <- function(base_url = "https://api.openai.com/v1",
-                            model = "gpt-4o-mini",
-                            api_key = openai_key()) {
+                         model = "gpt-4o-mini",
+                         seed = NULL,
+                         extra_args = list(),
+                         api_key = openai_key()) {
   structure(
     list(
       base_url = base_url,
       model = model,
-      api_key = api_key
+      seed = seed,
+      api_key = api_key,
+      extra_args = list()
     ),
     class = "elmer::openai_model"
   )
@@ -23,8 +27,11 @@ openai_chat <- function(mode = c("value", "stream", "async-stream", "async-value
   req <- openai_chat_req(
     messages = messages,
     tools = tools,
-    model = model,
-    stream = stream
+    model = model$model,
+    seed = model$seed,
+    stream = stream,
+    base_url = model$base_url,
+    api_key = model$api_key
   )
 
   handle_response <- switch(mode,
@@ -78,23 +85,30 @@ openai_key <- function() {
   key
 }
 
-openai_chat_req <- function(model,
-                            messages,
+# https://platform.openai.com/docs/api-reference/chat/create
+openai_chat_req <- function(messages,
                             tools = list(),
-                            stream = FALSE) {
-  if (length(tools) == 0) {
-    # OpenAI rejects tools=[]
-    tools <- NULL
-  }
+                            model = "gpt-4o-mini",
+                            seed = NULL,
+                            stream = TRUE,
+                            extra_args = list(),
+                            base_url = "https://api.openai.com/v1",
+                            api_key = openai_key()) {
 
-  data <- list(
-    model = model$model,
-    stream = stream,
+  check_string(model)
+  check_number_whole(seed, allow_null = TRUE)
+  check_bool(stream)
+
+  data <- compact(list2(
     messages = messages,
-    tools = tools
-  )
+    model = model,
+    seed = seed,
+    stream = stream,
+    tools = tools,
+    !!!extra_args
+  ))
 
-  req <- openai_request(base_url = model$base_url, key = model$api_key)
+  req <- openai_request(base_url = base_url, key = api_key)
   req <- req_url_path_append(req, "/chat/completions")
   req <- req_body_json(req, data)
   req