Flux inference fails: AttributeError: 'dict' object has no attribute 'to_request_dict'

[rensortino]

Problem Description

In the documentation, the input request parameters are all reported as dicts that will be parsed by the IImageInference class.

A typical input is as follows, as reported by the documentation:

{
  "taskType": "imageInference",
  "taskUUID": "6ba7b832-9dad-11d1-80b4-00c04fd430c8",
  "model": "bfl:7@1",
  "positivePrompt": "Create a realistic, high-quality image of a person using all provided reference images to preserve the same character identity."
  "inputs": {
    "referenceImages": [
      "c64351d5-4c59-42f7-95e1-eace013eddab",
      "d7e8f9a0-2b5c-4e7f-a1d3-9c8b7a6e5d4f",
      "e8f9a0b1-3c6d-4e8f-b2e4-0d9e8f7c6b5a",
      "f9a0b1c2-4d7e-5f9g-c3f5-1e0f9g8d7c6b",
      "a1b2c3d4-5e6f-7g8h-d4f6-2f1g0h9i8h7g"
    ]
  },
  "width": 1280,
  "height": 720
}

When passing the inputs in this format, the server raises an error and produces the following stack trace:

Traceback (most recent call last):
  File "/Users/rsortino/projects/glare-removal/sdk-python/repro_error.py", line 68, in <module>
    asyncio.run(main())
  File "/Users/rsortino/.local/share/uv/python/cpython-3.11.12-macos-aarch64-none/lib/python3.11/asyncio/runners.py", line 190, in run
    return runner.run(main)
           ^^^^^^^^^^^^^^^^
  File "/Users/rsortino/.local/share/uv/python/cpython-3.11.12-macos-aarch64-none/lib/python3.11/asyncio/runners.py", line 118, in run
    return self._loop.run_until_complete(task)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/rsortino/.local/share/uv/python/cpython-3.11.12-macos-aarch64-none/lib/python3.11/asyncio/base_events.py", line 654, in run_until_complete
    return future.result()
           ^^^^^^^^^^^^^^^
  File "/Users/rsortino/projects/glare-removal/sdk-python/repro_error.py", line 63, in main
    images = await runware.imageInference(request)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/rsortino/projects/glare-removal/sdk-python/runware/base.py", line 445, in imageInference
    return await self._retry_with_reconnect(self._imageInference, requestImage)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/rsortino/projects/glare-removal/sdk-python/runware/base.py", line 146, in _retry_with_reconnect
    result = await func(*args, **kwargs)
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/rsortino/projects/glare-removal/sdk-python/runware/base.py", line 567, in _imageInference
    raise e
  File "/Users/rsortino/projects/glare-removal/sdk-python/runware/base.py", line 538, in _imageInference
    request_object = self._buildImageRequest(requestImage, prompt, control_net_data_dicts, instant_id_data, ip_adapters_data, ace_plus_plus_data, pulid_data)
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/Users/rsortino/projects/glare-removal/sdk-python/runware/base.py", line 2207, in _buildImageRequest
    self._addImageProviderSettings(request_object, requestImage)
  File "/Users/rsortino/projects/glare-removal/sdk-python/runware/base.py", line 2322, in _addImageProviderSettings
    provider_dict = requestImage.providerSettings.to_request_dict()

Code to reproduce

import asyncio
from runware import Runware, IImageInference
import os
import base64
from pathlib import Path

# Get API key from environment variables
API_KEY = os.getenv("RUNWARE_API_KEY")

PROMPT = "Replace the dog with a cat. Keep the background the same."

def encode_image(image_path: str) -> str:
    with open(image_path, "rb") as f:
        image_data = base64.b64encode(f.read()).decode("utf-8")
    ext = Path(image_path).suffix.lower()
    mime = "image/jpeg" if ext in {".jpg", ".jpeg"} else f"image/{ext.lstrip('.')}"
    return f"data:{mime};base64,{image_data}"


async def main():
    runware = Runware(api_key=API_KEY)
    await runware.connect()

    image_data_uri = encode_image("examples/dalmatian.jpg")

    request_params = {
        "positivePrompt": PROMPT,
        "model": "bfl:7@1",
        "width": 1280,
        "height": 720,
        "includeCost": True,
        "inputs": {
            "referenceImages": [image_data_uri],
        },
    }
   
    request = IImageInference(**request_params)

    images = await runware.imageInference(request)
    print(f"Generated image: {images[0].imageURL}")


if __name__ == "__main__":
    asyncio.run(main())

A solution would be to assign the "inputs" dict to a IInputs object and then use this in the request, but IMHO it would be cleaner to handle this in the library rather than in the client code.

[Sirsho1997]

Added post_init() for other taskTypes as well
#246