feat(video): implement OpenAI video creation support

Add video creation functionality using OpenAI SDK. Update types to match OpenAI's video API and implement the actual creation method in the OpenAI client.

src/renderer/src/aiCore/legacy/clients/openai/OpenAIResponseAPIClient.ts

@@ -153,9 +153,9 @@ export class OpenAIResponseAPIClient extends OpenAIBaseClient<
     return await sdk.responses.create(payload, options)
   }
 
-  public async createVideo(params: CreateVideoParams) {
-    // TODO: implement it
-    throw new Error('Not implemented')
+  public async createVideo(params: CreateVideoParams): Promise<OpenAI.Videos.Video> {
+    const sdk = await this.getSdkInstance()
+    return await sdk.videos.create(params.params, params.options)
   }
 
   private async handlePdfFile(file: FileMetadata): Promise<OpenAI.Responses.ResponseInputFile | undefined> {

src/renderer/src/aiCore/legacy/index.ts

@@ -181,12 +181,11 @@ export default class AiProvider {
   }
 
   public async createVideo(params: CreateVideoParams): Promise<CreateVideoResult> {
-    throw new Error('Not implemented')
-    // if (this.apiClient instanceof OpenAIResponseAPIClient) {
-    //   return this.apiClient.createVideo(params)
-    // } else {
-    //   throw new Error('Video generation is not supported by this provider')
-    // }
+    if (this.apiClient instanceof OpenAIResponseAPIClient) {
+      return this.apiClient.createVideo(params)
+    } else {
+      throw new Error('Video generation is not supported by this provider')
+    }
   }
 
   public getBaseURL(): string {

src/renderer/src/types/video.ts

@@ -1,10 +1,6 @@
-export type CreateVideoRequest = {
-  type: 'openai'
-}
+import OpenAI from '@cherrystudio/openai'
 
-export type CreateVideoParams = {
-  providerId: string
-  modelId: string
-}
+// Only OpenAI for now.
+export type CreateVideoParams = { params: OpenAI.VideoCreateParams; options?: OpenAI.RequestOptions }
 
-export type CreateVideoResult = {}
+export type CreateVideoResult = OpenAI.Videos.Video