Update prompt format

Sources/SpeziLLMLocal/LLMLocalSchema+PromptFormatting.swift

@@ -12,6 +12,85 @@ import SpeziLLM
 extension LLMLocalSchema {
     /// Holds default prompt formatting strategies for [Llama2](https://ai.meta.com/llama/) as well as [Phi-2](https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/) models.
     public enum PromptFormattingDefaults {
+        /// Prompt formatting closure for the [Llama3](https://ai.meta.com/llama/) model
+        public static let llama3: (@Sendable (LLMContext) throws -> String) = { chat in // swiftlint:disable:this closure_body_length
+            let BEGINOFTEXT = "<|begin_of_text|>"
+            let SYSTEM = "system"
+            let USER = "user"
+            let ASSISTANT = "assistant"
+            let STARTHEADERID = "<|start_header_id|>"
+            let ENDHEADERID = "<|end_header_id|>"
+            let EOTID = "<|eot_id|>"
+
+            guard chat.first?.role == .system else {
+                throw LLMLocalError.illegalContext
+            }
+
+            var systemPrompts: [String] = []
+            var initialUserPrompt: String = ""
+
+            for contextEntity in chat {
+                if contextEntity.role != .system {
+                    if contextEntity.role == .user {
+                        initialUserPrompt = contextEntity.content
+                        break
+                    } else {
+                        throw LLMLocalError.illegalContext
+                    }
+                }
+
+                systemPrompts.append(contextEntity.content)
+            }
+
+            /// Build the initial Llama3 prompt structure
+            ///
+            /// A template of the prompt structure looks like:
+            /// """
+            /// <s>[INST] <<SYS>>
+            /// {your_system_prompt}
+            /// <</SYS>>
+            ///
+            /// {user_message_1} [/INST]
+            /// """
+            var prompt = """
+            \(BEGINOFTEXT)
+            \(STARTHEADERID)\(SYSTEM)\(ENDHEADERID)
+            \(systemPrompts.joined(separator: " "))
+            \(EOTID)
+            
+            \(STARTHEADERID)\(USER)\(ENDHEADERID)
+            \(initialUserPrompt)
+            \(EOTID)
+            
+            """ + " "   // Add a spacer to the generated output from the model
+
+            for contextEntity in chat.dropFirst(2) {
+                if contextEntity.role == .assistant() {
+                    /// Append response from assistant to the Llama3 prompt structure
+                    ///
+                    /// A template for appending an assistant response to the overall prompt looks like:
+                    /// {user_message_1} [/INST]){model_reply_1}</s>
+                    prompt += """
+                    \(STARTHEADERID)\(ASSISTANT)\(ENDHEADERID)
+                    \(contextEntity.content) 
+                    \(EOTID)
+                    """
+                } else if contextEntity.role == .user {
+                    /// Append response from user to the Llama3 prompt structure
+                    ///
+                    /// A template for appending an assistant response to the overall prompt looks like:
+                    /// <s>[INST] {user_message_2} [/INST]
+                    prompt += """
+                    \(STARTHEADERID)\(USER)\(ENDHEADERID)
+                    \(contextEntity.content) 
+                    \(EOTID)
+                    """ + " "   // Add a spacer to the generated output from the model
+                }
+            }
+
+            return prompt
+        }
+
         /// Prompt formatting closure for the [Llama2](https://ai.meta.com/llama/) model
         public static let llama2: (@Sendable (LLMContext) throws -> String) = { chat in     // swiftlint:disable:this closure_body_length
             /// BOS token of the LLM, used at the start of each prompt passage.

Tests/UITests/TestApp/LLMLocal/LLMLocalChatTestView.swift

@@ -27,7 +27,8 @@ struct LLMLocalChatTestView: View {
                     with: LLMLocalSchema(
                         modelPath: .cachesDirectory.appending(path: "llm.gguf"),
                         parameters: .init(maxOutputLength: 512),
-                        contextParameters: .init(contextWindowSize: 1024)
+                        contextParameters: .init(contextWindowSize: 1024),
+                        formatChat: LLMLocalSchema.PromptFormattingDefaults.llama3
                     )
                 )
             }