Add message branching

Updated the behaviour of commands:

- `lmcli edit`
  - by default create a new branch/message branch with the edited contents
  - add --in-place to avoid creating a branch
  - no longer delete messages after the edited message
  - only do the edit, don't fetch a new response
- `lmcli retry`
  - create a new branch rather than replacing old messages
  - add --offset to change where to retry from

pkg/cmd/prompt.go

@@ -15,22 +15,27 @@ func PromptCmd(ctx *lmcli.Context) *cobra.Command {
 		Short: "Do a one-shot prompt",
 		Long:  `Prompt the Large Language Model and get a response.`,
 		RunE: func(cmd *cobra.Command, args []string) error {
-			message := inputFromArgsOrEditor(args, "# What would you like to say?\n", "")
-			if message == "" {
+			input := inputFromArgsOrEditor(args, "# Write your prompt below\n", "")
+			if input == "" {
 				return fmt.Errorf("No message was provided.")
 			}
 
-			messages := []model.Message{
-				{
+			var messages []model.Message
+
+			// TODO: stop supplying system prompt as a message
+			system := ctx.GetSystemPrompt()
+			if system != "" {
+				messages = append(messages, model.Message{
 					Role:    model.MessageRoleSystem,
-					Content: ctx.GetSystemPrompt(),
-				},
-				{
-					Role:    model.MessageRoleUser,
-					Content: message,
-				},
+					Content: system,
+				})
 			}
 
+			messages = append(messages, model.Message{
+				Role:    model.MessageRoleUser,
+				Content: input,
+			})
+
 			_, err := cmdutil.Prompt(ctx, messages, nil)
 			if err != nil {
 				return fmt.Errorf("Error fetching LLM response: %v", err)