Non official fusion brain API client.

---

For actual official documentation please refer to https://fusionbrain.ai/docs/en/doc/api-dokumentaciya/

---

Features

1. Getting available models - GetModelsAsync.
2. Generation image
   1. Auto polling result status - GenerateImageAsync
   2. Manual pooling result status - StartGenerateAsync + GetGenerationStatusAsync

Steps for start

1. Configure ImageGeneratorSettings.

   1. BaseUrl and Authentication are mandatory.
   2. CheckGeneration are not mandatory with default values (Attempts = 10 + Delay = 10 seconds)

2. Create instance of IImageGeneratorClient.

   1. Use can use DI way or manually configuring, meanwhile you can use ConfigureImageGeneratorHttpClient HttpClient extension for simplify authentication configuration.

3. Get existing models by GetModelsAsync method

4. Use one of model for creating GeneratingImage

   1. Example:

      var generatingImage = new GeneratingImage(
          query: "<YOUR REQUEST>",
          excludingResultQuery: "<YOUR NEGATIVE PROMPT>",
          size: Size.Size1024X1024,
          numberOfImages: 1, // At now supporting only 1 image per request.
          mode: GenerationMode.Generate, // At now supporting only 1 mode - "GENERATE".
          model: model, // Model from step 4
          style: Style.Anime // One of style, available styles by API - https://cdn.fusionbrain.ai/static/styles/api)

5. Make request with auto or manual pooling result.

   1. Auto.
      1. Use IImageGeneratorClient.GenerateAsync(), result will be collection of images with processingId and content byte array.
   2. Manual
      1. Use IImageGeneratorClient.StartGenerateAsync(), result will be status with processing id and array of by arrays of case Done status (one of which is image byte array content).
      2. Depending on your application's polling configuration use IImageGeneratorClient.GetGenerationStatusAsync() to get actual status generation.
         1. Additionally u can get GeneratedImage by ToGeneratedImages collection of byte array extension.
   3. At each request you must have processing Model and sometimes processingId.