com.openai.unity 7.1.0 (#146)

- Refactor ImagesEndpoint - Changed all endpoint signatures to `IReadOnlyList<ImageResult>` - Fix Text To Speech generation for webgl - Updated Chat sample with image generation capabilities
RageAgainstThePixel · Dec 14, 2023 · 562bf5c · 562bf5c
1 parent a446783
commit 562bf5c
Show file tree

Hide file tree

Showing 11 changed files with 277 additions and 316 deletions.
diff --git a/Documentation~/README.md b/Documentation~/README.md
@@ -1107,12 +1107,10 @@ var api = new OpenAIClient();
 var request = new ImageGenerationRequest("A house riding a velociraptor", Models.Model.DallE_3);
 var imageResults = await api.ImagesEndPoint.GenerateImageAsync(request);
 
-foreach (var (path, texture) in results)
+foreach (var result in imageResults)
 {
-    Debug.Log(path);
-    // path == file://path/to/image.png
-    Assert.IsNotNull(texture);
-    // texture == The preloaded Texture2D
+    Debug.Log(result.ToString());
+    Assert.IsNotNull(result.Texture);
 }
 ```
 
@@ -1125,12 +1123,10 @@ var api = new OpenAIClient();
 var request = new ImageEditRequest(Path.GetFullPath(imageAssetPath), Path.GetFullPath(maskAssetPath), "A sunlit indoor lounge area with a pool containing a flamingo", size: ImageSize.Small);
 var imageResults = await api.ImagesEndPoint.CreateImageEditAsync(request);
 
-foreach (var (path, texture) in imageResults)
+foreach (var result in imageResults)
 {
-    Debug.Log(path);
-    // path == file://path/to/image.png
-    Assert.IsNotNull(texture);
-    // texture == The preloaded Texture2D
+    Debug.Log(result.ToString());
+    Assert.IsNotNull(result.Texture);
 }
 ```
 
@@ -1143,12 +1139,10 @@ var api = new OpenAIClient();
 var request = new ImageVariationRequest(imageTexture, size: ImageSize.Small);
 var imageResults = await api.ImagesEndPoint.CreateImageVariationAsync(request);
 
-foreach (var (path, texture) in imageResults)
+foreach (var result in imageResults)
 {
-    Debug.Log(path);
-    // path == file://path/to/image.png
-    Assert.IsNotNull(texture);
-    // texture == The preloaded Texture2D
+    Debug.Log(result.ToString());
+    Assert.IsNotNull(result.Texture);
 }
 ```
 
@@ -1157,14 +1151,12 @@ Alternatively, the endpoint can directly take a Texture2D with Read/Write enable
 ```csharp
 var api = new OpenAIClient();
 var request = new ImageVariationRequest(imageTexture, size: ImageSize.Small);
-var results = await api.ImagesEndPoint.CreateImageVariationAsync(request);
-// imageTexture is of type Texture2D
-foreach (var (path, texture) in results)
+var imageResults = await api.ImagesEndPoint.CreateImageVariationAsync(request);
+
+foreach (var result in imageResults)
 {
-    Debug.Log(path);
-    // path == file://path/to/image.png
-    Assert.IsNotNull(texture);
-    // texture == The preloaded Texture2D
+    Debug.Log(result.ToString());
+    Assert.IsNotNull(result.Texture);
 }
 ```
 

diff --git a/Runtime/Audio/AudioEndpoint.cs b/Runtime/Audio/AudioEndpoint.cs
@@ -6,6 +6,7 @@
 using System.Threading;
 using System.Threading.Tasks;
 using UnityEngine;
+using UnityEngine.Networking;
 using UnityEngine.Scripting;
 using Utilities.WebRequestRest;
 
@@ -56,12 +57,17 @@ public async Task<Tuple<string, AudioClip>> CreateSpeechAsync(SpeechRequest requ
                 _ => throw new NotSupportedException(request.ResponseFormat.ToString())
             };
             var payload = JsonConvert.SerializeObject(request, OpenAIClient.JsonSerializationOptions);
-            var response = await Rest.PostAsync(GetUrl("/speech"), payload, new RestParameters(client.DefaultRequestHeaders), cancellationToken);
-            response.Validate(EnableDebug);
-            await Rest.ValidateCacheDirectoryAsync();
-            var cachedPath = Path.Combine(Rest.DownloadCacheDirectory, $"{request.Voice}-{DateTime.UtcNow:yyyyMMddThhmmss}.{ext}");
-            await File.WriteAllBytesAsync(cachedPath, response.Data, cancellationToken).ConfigureAwait(true);
-            var clip = await Rest.DownloadAudioClipAsync($"file://{cachedPath}", audioFormat, cancellationToken: cancellationToken);
+            var clipName = $"{request.Voice}-{DateTime.UtcNow:yyyyMMddThhmmss}.{ext}";
+            var clip = await Rest.DownloadAudioClipAsync(
+                GetUrl("/speech"),
+                audioFormat,
+                UnityWebRequest.kHttpVerbPOST,
+                clipName,
+                payload,
+                parameters: new RestParameters(client.DefaultRequestHeaders),
+                debug: EnableDebug,
+                cancellationToken: cancellationToken);
+            Rest.TryGetDownloadCacheItem(clipName, out var cachedPath);
             return new Tuple<string, AudioClip>(cachedPath, clip);
         }
 

diff --git a/Runtime/Chat/Conversation.cs b/Runtime/Chat/Conversation.cs
@@ -13,9 +13,9 @@ public sealed class Conversation
     {
         [Preserve]
         [JsonConstructor]
-        public Conversation([JsonProperty("messages")] List<Message> messages)
+        public Conversation([JsonProperty("messages")] List<Message> messages = null)
         {
-            this.messages = messages;
+            this.messages = messages ?? new List<Message>();
         }
 
         private readonly List<Message> messages;

diff --git a/Runtime/Files/FilesEndpoint.cs b/Runtime/Files/FilesEndpoint.cs
@@ -146,7 +146,7 @@ public async Task<FileResponse> GetFileInfoAsync(string fileId, CancellationToke
         public async Task<string> DownloadFileAsync(string fileId, IProgress<Progress> progress = null, CancellationToken cancellationToken = default)
         {
             var file = await GetFileInfoAsync(fileId, cancellationToken);
-            return await Rest.DownloadFileAsync(GetUrl($"/{file.Id}/content"), file.FileName, new RestParameters(client.DefaultRequestHeaders, progress), cancellationToken: cancellationToken);
+            return await Rest.DownloadFileAsync(GetUrl($"/{file.Id}/content"), file.FileName, new RestParameters(client.DefaultRequestHeaders, progress), EnableDebug, cancellationToken);
         }
     }
 }
diff --git a/Runtime/Images/ImageGenerationRequest.cs b/Runtime/Images/ImageGenerationRequest.cs
@@ -13,13 +13,6 @@ namespace OpenAI.Images
     [Preserve]
     public sealed class ImageGenerationRequest
     {
-        [Preserve]
-        [Obsolete("Use new constructor")]
-        public ImageGenerationRequest(string prompt, int numberOfResults = 1, ImageSize size = ImageSize.Large, string user = null, ResponseFormat responseFormat = ResponseFormat.Url)
-        {
-            throw new NotSupportedException();
-        }
-
         /// <summary>
         /// Constructor.
         /// </summary>
@@ -59,22 +52,23 @@ public ImageGenerationRequest(string prompt, int numberOfResults = 1, ImageSize
         /// <param name="user">
         /// A unique identifier representing your end-user, which can help OpenAI to monitor and detect abuse.
         /// </param>
+        [JsonConstructor]
         public ImageGenerationRequest(
-            string prompt,
-            Model model = null,
-            int numberOfResults = 1,
-            string quality = null,
-            ResponseFormat responseFormat = ResponseFormat.Url,
-            string size = null,
-            string style = null,
-            string user = null)
+            [JsonProperty("prompt")] string prompt,
+            [JsonProperty("model")] Model model = null,
+            [JsonProperty("n")] int numberOfResults = 1,
+            [JsonProperty("quality")] string quality = null,
+            [JsonProperty("response_format")] ResponseFormat responseFormat = ResponseFormat.Url,
+            [JsonProperty("size")] string size = null,
+            [JsonProperty("style")] string style = null,
+            [JsonProperty("user")] string user = null)
         {
             Prompt = prompt;
             Model = string.IsNullOrWhiteSpace(model?.Id) ? Models.Model.DallE_2 : model;
             Number = numberOfResults;
             Quality = quality;
             ResponseFormat = responseFormat;
-            Size = size;
+            Size = size ?? "1024x1024";
             Style = style;
             User = user;
         }

diff --git a/Runtime/Images/ImageResult.cs b/Runtime/Images/ImageResult.cs
@@ -1,12 +1,13 @@
 // Licensed under the MIT License. See LICENSE in the project root for license information.
 
 using Newtonsoft.Json;
+using UnityEngine;
 using UnityEngine.Scripting;
 
 namespace OpenAI.Images
 {
     [Preserve]
-    internal sealed class ImageResult
+    public sealed class ImageResult
     {
         [Preserve]
         [JsonConstructor]
@@ -31,5 +32,40 @@ public ImageResult(
         [Preserve]
         [JsonProperty("revised_prompt")]
         public string RevisedPrompt { get; private set; }
+
+        [Preserve]
+        [JsonIgnore]
+        public string CachedPath { get; internal set; }
+
+        [Preserve]
+        [JsonIgnore]
+        public Texture2D Texture { get; internal set; }
+
+        [Preserve]
+        public static implicit operator Texture2D(ImageResult imageResult) => imageResult.Texture;
+
+        [Preserve]
+        public static implicit operator string(ImageResult imageResult) => imageResult.ToString();
+
+        [Preserve]
+        public override string ToString()
+        {
+            if (!string.IsNullOrWhiteSpace(CachedPath))
+            {
+                return CachedPath;
+            }
+
+            if (!string.IsNullOrWhiteSpace(Url))
+            {
+                return Url;
+            }
+
+            if (!string.IsNullOrWhiteSpace(B64_Json))
+            {
+                return B64_Json;
+            }
+
+            return string.Empty;
+        }
     }
 }