feat(openai): add transcription, image generation, and fix video sizes

[caffeinum]

Summary

• Add OpenAI Whisper transcription support (whisper-1, gpt-4o-transcribe, gpt-4o-mini-transcribe)
• Add OpenAI image generation (gpt-image-1, gpt-image-1-mini, gpt-image-1.5, dall-e-2, dall-e-3)
• Fix Sora video size mappings - add 1024x1792, 1792x1024 for 9:21/21:9 aspect ratios
• Add example files for new features

New Features

Transcription

import { experimental_transcribe as transcribe } from "ai";
import { openai } from "vargai/ai";

const result = await transcribe({
  model: openai.transcriptionModel("whisper-1"),
  audio: audioBuffer,
});

Image Generation

import { generateImage } from "ai";
import { openai } from "vargai/ai";

const { images } = await generateImage({
  model: openai.imageModel("gpt-image-1"),
  prompt: "a futuristic city at sunset",
});

// or with dall-e-3
const { images } = await generateImage({
  model: openai.imageModel("dall-e-3"),
  prompt: "a whimsical treehouse",
  providerOptions: {
    openai: { quality: "hd", style: "vivid" },
  },
});

Video (existing, with fixed sizes)

import { generateVideo, openai } from "vargai/ai";

const { video } = await generateVideo({
  model: openai.videoModel("sora-2"),
  prompt: "a cat walking on a beach",
  aspectRatio: "21:9", // now supported
});

Testing

• TypeScript compiles without errors
• LSP diagnostics clean