Add support for multiple images and videos in response methods (#393)

* Add support for multiple images and videos in response methods

- Updated `respond` and `streamResponse` methods to accept `[UserInput.Image]?` and `[UserInput.Video]?` arrays.
- Added overloads to maintain backward compatibility with existing single `image` and `video` parameters.
- Simplified internal logic using default empty arrays for nil values, improving code readability and flexibility.

* Fix mismatched input types in `respond` method

- Changed parameters from `UserInput.Image?` and `UserInput.Video?` to `[UserInput.Image]?` and `[UserInput.Video]?`.
- Previously, the method expected arrays internally but accepted single optional inputs, leading to a mismatch.
- This fix ensures consistency and prepares for cleaner multi-input support in follow-up refactors.

* Fix missing `try await` in `respond` overload

- Corrected the call to the main async `respond` method by adding `try await`.
- Prevents async runtime errors and ensures the method correctly propagates asynchronous execution and error handling.

* Enforce non-optional arrays for media inputs

- Changed `images` and `videos` parameters in `respond` and `streamResponse` methods to be non-optional.
- Removed internal fallback to empty arrays (`?? []`), shifting responsibility to the caller.
- Updated doc comments to reflect that these inputs are now required as arrays.
- Clarifies the method contract and simplifies internal logic by eliminating conditional handling.

Author: Tavernari

Libraries/MLXLMCommon/Streamlined.swift

@@ -179,23 +179,26 @@ public class ChatSession {
         self.generator = .init(
             model: .context(model), instructions: instructions, processing: processing,
             generateParameters: generateParameters)
-    }
-
+    }     
+    
     /// Produces a response to a prompt.
     ///
     /// - Parameters:
     ///   - prompt: the prompt
-    ///   - image: optional image (for use with VLMs)
-    ///   - video: optional video (for use with VLMs)
+    ///   - images: list of image (for use with VLMs)
+    ///   - videos: list of video (for use with VLMs)
     /// - Returns: response from the model
     public func respond(
-        to prompt: String, image: UserInput.Image? = nil, video: UserInput.Video? = nil
+        to prompt: String, 
+        images: [UserInput.Image], 
+        videos: [UserInput.Video]
     ) async throws -> String {
         generator.messages = [
             .user(
                 prompt,
-                images: image.flatMap { [$0] } ?? [],
-                videos: video.flatMap { [$0] } ?? [])
+                images: images,
+                videos: videos
+            )
         ]
         return try await generator.generate()
     }
@@ -204,18 +207,59 @@ public class ChatSession {
     ///
     /// - Parameters:
     ///   - prompt: the prompt
-    ///   - image: optional image (for use with VLMs)
-    ///   - video: optional video (for use with VLMs)
+    ///   - images: optional image (for use with VLMs)
+    ///   - videos: optional video (for use with VLMs)
+    /// - Returns: response from the model
+    public func respond(
+        to prompt: String, 
+        image: UserInput.Image? = nil, 
+        video: UserInput.Video? = nil
+    ) async throws -> String {
+        try await respond(
+            to: prompt,
+            images: image.flatMap { [$0] } ?? [],
+            videos: video.flatMap { [$0] } ?? []
+        )
+    }
+
+    /// Produces a response to a prompt.
+    ///
+    /// - Parameters:
+    ///   - prompt: the prompt
+    ///   - images: list of image (for use with VLMs)
+    ///   - videos: list of video (for use with VLMs)
     /// - Returns: a stream of tokens (as Strings) from the model
     public func streamResponse(
-        to prompt: String, image: UserInput.Image? = nil, video: UserInput.Video? = nil
+        to prompt: String, 
+        images: [UserInput.Image], 
+        videos: [UserInput.Video]
     ) -> AsyncThrowingStream<String, Error> {
         generator.messages = [
             .user(
                 prompt,
-                images: image.flatMap { [$0] } ?? [],
-                videos: video.flatMap { [$0] } ?? [])
+                images: images,
+                videos: videos
+            )
         ]
         return generator.stream()
     }
+
+    /// Produces a response to a prompt.
+    ///
+    /// - Parameters:
+    ///   - prompt: the prompt
+    ///   - image: optional image (for use with VLMs)
+    ///   - video: optional video (for use with VLMs)
+    /// - Returns: a stream of tokens (as Strings) from the model
+    public func streamResponse(
+        to prompt: String, 
+        image: UserInput.Image? = nil, 
+        video: UserInput.Video? = nil
+    ) -> AsyncThrowingStream<String, Error> {
+        streamResponse(
+            to: prompt, 
+            images: image.flatMap { [$0] } ?? [],
+            videos: video.flatMap { [$0] } ?? []
+        )
+    }
 }