constants.py

import re
from huggingface_hub import HfApi
from pprint import pprint
from stablepy import BUILTIN_UPSCALERS

# Initial UPSCALER_DICT_GUI
UPSCALER_DICT_GUI = {
    **{bu.replace(" ", "_"): bu for bu in BUILTIN_UPSCALERS[2:]},
    "RealESRNet_x4plus": "https://github.com/xinntao/Real-ESRGAN/releases/download/v0.1.1/RealESRNet_x4plus.pth",
    "4x_foolhardy_Remacri": "https://huggingface.co/FacehugmanIII/4x_foolhardy_Remacri/resolve/main/4x_foolhardy_Remacri.pth",
    "Remacri4xExtraSmoother": "https://huggingface.co/hollowstrawberry/upscalers-backup/resolve/main/ESRGAN/Remacri%204x%20ExtraSmoother.pth",
    "Lollypop": "https://huggingface.co/hollowstrawberry/upscalers-backup/resolve/main/ESRGAN/lollypop.pth",
    "NickelbackFS4x": "https://huggingface.co/hollowstrawberry/upscalers-backup/resolve/main/ESRGAN/NickelbackFS%204x.pth",
    "Real_HAT_GAN_SRx4": "https://huggingface.co/halffried/gyre_upscalers/resolve/main/hat_ganx4/Real_HAT_GAN_SRx4.safetensors",
    "HAT-L_SRx4_ImageNet-pretrain": "https://huggingface.co/halffried/gyre_upscalers/resolve/main/hat_lx4/HAT-L_SRx4_ImageNet-pretrain.safetensors",
    "Real-ESRGAN-Anime-finetuning": "https://huggingface.co/danhtran2mind/Real-ESRGAN-Anime-finetuning/resolve/main/Real-ESRGAN-Anime-finetuning.pth",
    "8x_NMKD-Superscale_150000_G": "https://huggingface.co/lantianhang/8x_NMKD-Superscale_150000_G/resolve/main/8x_NMKD-Superscale_150000_G.pth",
    "4x_Valar_v1": "https://huggingface.co/halffried/gyre_upscalers/resolve/main/esrgan_valar_x4/4x_Valar_v1.pth",
    "Ghibli_Grain": "https://huggingface.co/anonderpling/upscalers/resolve/main/ESRGAN/ghibli_grain.pth",
    "Detoon4x": "https://huggingface.co/anonderpling/upscalers/resolve/main/ESRGAN/4x_detoon_225k.pth",
    "2x_Text2HD_v1-RealPLKSR": "https://github.com/starinspace/StarinspaceUpscale/releases/download/Models/2x_Text2HD_v.1-RealPLKSR.pth",
    "4xNomosWebPhoto_esrgan": "https://github.com/Phhofm/models/releases/download/4xNomosWebPhoto_esrgan/4xNomosWebPhoto_esrgan.pth",
    "4xTextureDAT2_otf": "https://github.com/Phhofm/models/releases/download/4xTextureDAT2_otf/4xTextureDAT2_otf.pth",
    "2xVHS2HD-RealPLKSR": "https://github.com/starinspace/StarinspaceUpscale/releases/download/Models/2xVHS2HD-RealPLKSR.pth",
}

def clean_filename_for_key(filename):
    """Removes the extension and replaces special characters with underscores."""
    name_without_extension = filename.rsplit('.', 1)[0]
    # Replace common separators and any non-alphanumeric characters (except underscore) with an underscore
    cleaned_name = re.sub(r'[^a-zA-Z0-9_]+', '_', name_without_extension)
    return cleaned_name

def add_upscalers_from_author(author_name, upscaler_dict, repos_to_avoid=None):
    """
    Scans an author's Hugging Face repos and adds model files to the upscaler dictionary.

    Args:
        author_name (str): The Hugging Face username of the author.
        upscaler_dict (dict): The dictionary of upscalers to update.
        repos_to_avoid (list, optional): A list of repository names to skip. Defaults to None.
    """
    if repos_to_avoid is None:
        repos_to_avoid = []

    print(f"--- Processing author: {author_name} ---")
    api = HfApi()
    try:
        models = api.list_models(author=author_name)
        all_repo_ids = [model.modelId for model in models]

        # Filter out repositories that should be avoided
        repo_ids = [
            repo for repo in all_repo_ids
            if not any(avoid_name in repo for avoid_name in repos_to_avoid)
        ]
        
        filtered_count = len(all_repo_ids) - len(repo_ids)
        print(f"Found {len(all_repo_ids)} repositories. Skipping {filtered_count}, processing {len(repo_ids)}.")

    except Exception as e:
        print(f"Could not fetch repositories for {author_name}: {e}")
        return

    for repo_id in repo_ids:
        try:
            files = api.list_repo_files(repo_id)
            # Prioritize .pth files
            pth_files = [f for f in files if f.endswith(".pth")]
            model_files_to_add = []

            if pth_files:
                model_files_to_add = pth_files
            else:
                # If no .pth files, look for .safetensors
                safetensors_files = [f for f in files if f.endswith(".safetensors")]
                if safetensors_files:
                    model_files_to_add = safetensors_files

            if not model_files_to_add:
                continue

            print(f"Found {len(model_files_to_add)} model(s) in {repo_id}")
            for file_path in model_files_to_add:
                # Get just the filename from the full path
                filename = file_path.split('/')[-1]
                key = clean_filename_for_key(filename)

                if key in upscaler_dict:
                    print(f"  - Skipping duplicate key: {key}")
                    continue

                url = f"https://huggingface.co/{repo_id}/resolve/main/{file_path}"
                upscaler_dict[key] = url
                print(f"  + Added: {key}")

        except Exception as e:
            print(f"Could not process repo {repo_id}: {e}")

# --- Main Execution ---
# Create a copy to avoid modifying the original dict while iterating if needed later
updated_upscaler_dict = UPSCALER_DICT_GUI.copy()

# Define the list of repositories to skip
REPOS_TO_AVOID = ["BHI_Filtering_Post", "ppo-LunarLander-v2", "dqn-SpaceInvadersNoFrameskip-v4", "dqn-BeamRiderNoFrameskip-v4"]

# Add models from Kim2091 (no repos to avoid in this case)
add_upscalers_from_author("Kim2091", updated_upscaler_dict)

# Add models from Phips, passing the list of repos to avoid
add_upscalers_from_author("Phips", updated_upscaler_dict, repos_to_avoid=REPOS_TO_AVOID)