sd_yun/downloading_en.py

##~   DOWNLOADING CODE | BY: ANXETY   ~##

from directory_setup import *
from models_data import model_list, vae_list, controlnet_list

import os
import re
import time
import json
import shutil
import zipfile
import requests
import subprocess
from datetime import timedelta
from subprocess import getoutput
from IPython.utils import capture
from IPython.display import clear_output
from urllib.parse import urlparse, parse_qs


# Setup Env
env = os.getenv('ENV_NAME')
root_path = os.getenv('ROOT_PATH')
webui_path = os.getenv('WEBUI_PATH')
free_plan = os.getenv('FREE_PLAN')

UI = os.getenv('SDW_UI')
OLD_UI = os.getenv('SDW_OLD_UI')

os.chdir(root_path)


# ============ loading settings V4 =============
def load_settings(path):
    if os.path.exists(path):
        with open(path, 'r') as file:
            return json.load(file)
    return {}

settings = load_settings(f'{root_path}/settings.json')

VARIABLES = [
    'model', 'model_num', 'inpainting_model',
    'vae', 'vae_num', 'latest_webui', 'latest_exstensions',
    'change_webui', 'detailed_download', 'controlnet',
    'controlnet_num', 'commit_hash', 'huggingface_token',
    'ngrok_token', 'zrok_token', 'commandline_arguments',
    'Model_url', 'Vae_url', 'LoRA_url', 'Embedding_url',
    'Extensions_url', 'custom_file_urls'
]

locals().update({key: settings.get(key) for key in VARIABLES})


# ================ LIBRARIES V2 ================
flag_file = f"{root_path}/libraries_installed.txt"

if not os.path.exists(flag_file):
    print("💿 Installing the libraries, it's going to take a while:\n")

    install_lib = {
        # "aria2": "apt -y install aria2",
        "aria2": "pip install aria2",
        "localtunnel": "npm install -g localtunnel",
    }
    if controlnet != 'none':
        install_lib["insightface"] = "pip install insightface"

    additional_libs = {
        "Google Colab": {
            "xformers": "pip install xformers==0.0.23.post1 --no-deps",
            "torch": "pip install torch==2.1.2+cu121 torchvision==0.16.2+cu121 torchaudio==2.1.2 --extra-index-url https://download.pytorch.org/whl/cu121",
        },
        "Kaggle": {
            "xformers": "pip install xformers==0.0.26.post1",
            # "torch": "pip install torch==2.1.2+cu121 torchvision==0.16.2+cu121 torchaudio==2.1.2 --extra-index-url https://download.pytorch.org/whl/cu121",
            # "aiohttp": "pip install trash-cli && trash-put /opt/conda/lib/python3.10/site-packages/aiohttp*" # fix install req
        }
    }
    if env in additional_libs:
        install_lib.update(additional_libs[env])

    # Loop through libraries
    for index, (package, install_cmd) in enumerate(install_lib.items(), start=1):
        print(f"\r[{index}/{len(install_lib)}] \033[32m>>\033[0m Installing \033[33m{package}\033[0m..." + " "*35, end='')
        subprocess.run(install_cmd, shell=True, capture_output=True)

    # Additional specific packages
    with capture.capture_output():
        get_ipython().system('curl -s -OL https://github.com/DEX-1101/sd-webui-notebook/raw/main/res/new_tunnel --output-dir {root_path}')
        get_ipython().system('curl -s -Lo /usr/bin/cl https://github.com/cloudflare/cloudflared/releases/latest/download/cloudflared-linux-amd64 && chmod +x /usr/bin/cl')
        get_ipython().system('curl -sLO https://github.com/openziti/zrok/releases/download/v0.4.32/zrok_0.4.32_linux_amd64.tar.gz && tar -xzf zrok_0.4.32_linux_amd64.tar.gz -C /usr/bin && rm -f zrok_0.4.32_linux_amd64.tar.gz')

    clear_output()

    # Save file install lib
    with open(flag_file, "w") as f:
        f.write(">W<'")

    print("🍪 Libraries are installed!" + " "*35)
    time.sleep(2)
    clear_output()


# =================== OTHER ====================
# Setup Timer
start_colab = int(os.environ.get("START_COLAB", time.time() - 5))
os.environ["START_COLAB"] = str(start_colab)

def download_cfg_files(file_paths, destination_path):
    base_url = "https://huggingface.co/NagisaNao/SD-CONFIGS/resolve/main"
    for filename in file_paths:
        file_name = filename.split('/')[-1]
        get_ipython().system('wget -O {destination_path}/{file_name} {base_url}/{filename}')

def cfg_download():
    common_files = ["styles.csv"]
    a1111_files = ["A1111/config.json", "A1111/ui-config.json"]
    forge_files = ["reForge/config.json", "reForge/ui-config.json"]

    with capture.capture_output():
        download_cfg_files(common_files, webui_path)
        ui_files = a1111_files if UI == 'A1111' else forge_files
        download_cfg_files(ui_files, webui_path)

def remove_dir(directory_path):
    if directory_path and os.path.exists(directory_path):
        try:
            shutil.rmtree(directory_path)
        except Exception:
            get_ipython().system('rm -rf {directory_path}')

TEMPORARY_DIR = f'{root_path}/temp_dir'
def copy_items_with_replace(src_base, dst_base):
    items_to_copy = [
        'embeddings',
        'models/Stable-diffusion',
        'models/VAE',
        'models/Lora',
        'models/ControlNet'
    ]

    print("⌚ Moving files...", end='')
    for item in items_to_copy:
        src = os.path.join(src_base, item)
        dst = os.path.join(dst_base, item)

        if os.path.exists(src):
            if os.path.exists(dst):
                remove_dir(dst)
            os.makedirs(os.path.dirname(dst), exist_ok=True)
            shutil.move(src, dst)
    print("\r🔥 Files moved!" + " "*15)

def handle_colab_timer(webui_path, timer_colab):
    timer_file_path = os.path.join(webui_path, 'static', 'colabTimer.txt')
    if not os.path.exists(timer_file_path):
        with open(timer_file_path, 'w') as timer_file:
            timer_file.write(str(timer_colab))
    else:
        with open(timer_file_path, 'r') as timer_file:
            timer_colab = float(timer_file.read())
    return timer_colab

def unpack_webui():
    start_install = time.time()
    print(f"⌚ Unpacking Stable Diffusion{' (Forge)' if UI == 'Forge' else ''}...", end='')

    with capture.capture_output():
        download_url = "https://huggingface.co/lang03383/hisd/resolve/main/iiiiiss_v3.zip"
        if UI == 'Forge':
            download_url = "https://huggingface.co/NagisaNao/fast_repo/resolve/main/FULL_REPO.zip"

        zip_path = f"{root_path}/repo.zip"
        get_ipython().system('aria2c --console-log-level=error -c -x 16 -s 16 -k 1M {download_url} -d {root_path} -o repo.zip')
        get_ipython().system('unzip -q -o {zip_path} -d {webui_path}')
        get_ipython().system('rm -rf {zip_path}')

        handle_colab_timer(webui_path, start_colab)

    install_time = time.time() - start_install
    minutes, seconds = divmod(int(install_time), 60)
    print(f"\r🚀 Unpacking complete! For {minutes:02}:{seconds:02} ⚡" + " "*15)

    if os.path.exists(TEMPORARY_DIR):
        copy_items_with_replace(TEMPORARY_DIR, webui_path)
        remove_dir(TEMPORARY_DIR)

# ================= MAIN CODE ==================
if os.path.exists(webui_path):
    if UI != OLD_UI:
        print(f'Switching the WebUI from \033[33m{OLD_UI}\033[0m to \033[33m{UI}\033[0m:')
        copy_items_with_replace(webui_path, TEMPORARY_DIR)
        remove_dir(webui_path)
        os.environ['SDW_OLD_UI'] = UI
        time.sleep(2)
        clear_output()

if not os.path.exists(webui_path):
    unpack_webui()
    # cfg_download()
else:
    print("🚀 All unpacked... Skip. ⚡")
    timer_colab = handle_colab_timer(webui_path, start_colab)
    elapsed_time = str(timedelta(seconds=time.time() - timer_colab)).split('.')[0]
    print(f"⌚️ You have been conducting this session for - \033[33m{elapsed_time}\033[0m")


## Changes extensions and WebUi
if latest_webui or latest_exstensions:
    action = "WebUI and Extensions" if latest_webui and latest_exstensions else ("WebUI" if latest_webui else "Extensions")
    print(f"⌚️ Updating {action}...", end='')
    with capture.capture_output():
        get_ipython().system('git config --global user.email "you@example.com"')
        get_ipython().system('git config --global user.name "Your Name"')

        ## Update Webui
        if latest_webui:
            get_ipython().run_line_magic('cd', '{webui_path}')
            get_ipython().system('git restore .')
            get_ipython().system('git pull -X theirs --rebase --autostash')

        ## Update extensions
        if latest_exstensions:
            get_ipython().system('{\'for dir in \' + webui_path + \'/extensions/*/; do cd \\"$dir\\" && git reset --hard && git pull; done\'}')
    print(f"\r✨ Updating {action} Completed!")


# === FIXING EXTENSIONS ===
anxety_repos = "https://huggingface.co/NagisaNao/fast_repo/resolve/main"
with capture.capture_output():
    # --- Umi-Wildcard ---
    get_ipython().system("sed -i '521s/open=\\(False\\|True\\)/open=False/' {webui_path}/extensions/Umi-AI-Wildcards/scripts/wildcard_recursive.py  # Closed accordion by default")
    # --- Encrypt-Image ---
    get_ipython().system("sed -i '9,37d' {webui_path}/extensions/Encrypt-Image/javascript/encrypt_images_info.js # Removes the weird text in webui")


## Version switching
if commit_hash:
    print('⏳ Time machine activation...', end="")
    with capture.capture_output():
        get_ipython().run_line_magic('cd', '{webui_path}')
        get_ipython().system('git config --global user.email "you@example.com"')
        get_ipython().system('git config --global user.name "Your Name"')
        get_ipython().system('git reset --hard {commit_hash}')
    print(f"\r⌛️ The time machine has been activated! Current commit: \033[34m{commit_hash}\033[0m")


## Downloading model and stuff | oh~ Hey! If you're freaked out by that code too, don't worry, me too!
print("📦 Downloading models and stuff...", end='')

extension_repo = []
PREFIXES = {
    "model": models_dir,
    "vae": vaes_dir,
    "lora": loras_dir,
    "embed": embeddings_dir,
    "extension": extensions_dir,
    "control": control_dir,
    "adetailer": adetailer_dir,
    "config": webui_path
}
get_ipython().system('mkdir -p {" ".join(PREFIXES.values())}')

''' Formatted Info Output '''

def center_text(text, terminal_width=45):
    padding = (terminal_width - len(text)) // 2
    return f"{' ' * padding}{text}{' ' * padding}"

def format_output(url, dst_dir, file_name, image_name=None, image_url=None):
    info = center_text(f"[{file_name.split('.')[0]}]")
    sep_line = '---' * 20

    print(f"\n\033[32m{sep_line}\033[36;1m{info}\033[32m{sep_line}\033[0m")
    print(f"\033[33mURL: {url}")
    print(f"\033[33mSAVE DIR: \033[34m{dst_dir}")
    print(f"\033[33mFILE NAME: \033[34m{file_name}\033[0m")
    if 'civitai' in url and image_url:
        print(f"\033[32m[Preview DL]:\033[0m {image_name} - {image_url}\n")

''' GET CivitAi API - DATA '''

def CivitAi_API(url, file_name=None):
    SUPPORT_TYPES = ('Checkpoint', 'TextualInversion', 'LORA')
    CIVITAI_TOKEN = "62c0c5956b2f9defbd844d754000180b"

    url = url.split('?token=')[0] if '?token=' in url else url
    url = url.replace('?type=', f'?token={CIVITAI_TOKEN}&type=') if '?type=' in url else f"{url}?token={CIVITAI_TOKEN}"

    def get_model_data(url):
        base_url = "https://civitai.com/api/v1"
        try:
            if "civitai.com/models/" in url:
                if '?modelVersionId=' in url:
                    version_id = url.split('?modelVersionId=')[1]
                else:
                    model_id = url.split('/models/')[1].split('/')[0]
                    model_data = requests.get(f"{base_url}/models/{model_id}").json()
                    version_id = model_data['modelVersions'][0].get('id')
            else:
                version_id = url.split('/models/')[1].split('/')[0]

            return requests.get(f"{base_url}/model-versions/{version_id}").json()
        except (KeyError, IndexError, requests.RequestException) as e:
            return None

    data = get_model_data(url)

    if not data:
        print("\033[31m[Data Info]:\033[0m Failed to retrieve data from the API.\n")
        return 'None', None, None, None, None, None, None

    def get_model_info(url, data):
        model_type = data['model']['type']
        model_name = data['files'][0]['name']

        if 'type=' in url:
            url_model_type = parse_qs(urlparse(url).query).get('type', [''])[0].lower()
            if 'vae' in url_model_type:
                model_type = data['files'][1]['type']
                model_name = data['files'][1]['name']

        if file_name and '.' not in file_name:
            file_extension = model_name.split('.')[-1]
            model_name = f"{file_name}.{file_extension}"
        elif file_name:
            model_name = file_name

        return model_type, model_name

    def get_download_url(data, model_type):
        if any(t.lower() in model_type.lower() for t in SUPPORT_TYPES):
            return data['files'][0]['downloadUrl']

        return data['files'][1]['downloadUrl'] if 'type' in url else data['files'][0]['downloadUrl']

    def get_image_info(data, model_type, model_name):
        if not any(t in model_type for t in SUPPORT_TYPES):
            return None, None

        for image in data.get('images', []):
            if image['nsfwLevel'] >= 4 and env == 'Kaggle':  # Filter NSFW images for Kaggle
                continue
            image_url = image['url']
            image_extension = image_url.split('.')[-1]
            image_name = f"{model_name.split('.')[0]}.preview.{image_extension}" if image_url else None
            return image_url, image_name
        return None, None

    model_type, model_name = get_model_info(url, data)
    download_url = get_download_url(data, model_type)
    image_url, image_name = get_image_info(data, model_type, model_name)

    return f"{download_url}{'&' if '?' in download_url else '?'}token={CIVITAI_TOKEN}", download_url, model_type, model_name, image_url, image_name, data

''' Main Download Code '''

def strip_(url):
    if 'github.com' in url:
        return url.replace('/blob/', '/raw/')
    elif "huggingface.co" in url:
        url = url.replace('/blob/', '/resolve/')
        return url.split('?')[0] if '?' in url else url
    return url

def download(url):
    links_and_paths = [link_or_path.strip() for link_or_path in url.split(',') if link_or_path.strip()]

    for link_or_path in links_and_paths:
        if any(link_or_path.lower().startswith(prefix) for prefix in PREFIXES):
            handle_manual(link_or_path)
        else:
            url, dst_dir, file_name = link_or_path.split()
            manual_download(url, dst_dir, file_name)

    # Unpuck ZIPs Files
    for directory in PREFIXES.values():
        for root, _, files in os.walk(directory):
            for file in files:
                if file.endswith(".zip"):
                    zip_path = os.path.join(root, file)
                    extract_path = os.path.splitext(zip_path)[0]
                    with zipfile.ZipFile(zip_path, 'r') as zip_ref:
                        zip_ref.extractall(extract_path)
                    os.remove(zip_path)

def handle_manual(url):
    url_parts = url.split(':', 1)
    prefix, path = url_parts[0], url_parts[1]

    file_name_match = re.search(r'\[(.*?)\]', path)
    file_name = file_name_match.group(1) if file_name_match else None
    if file_name:
        path = re.sub(r'\[.*?\]', '', path)

    if prefix in PREFIXES:
        dir = PREFIXES[prefix]
        if prefix != "extension":
            try:
                manual_download(path, dir, file_name=file_name, prefix=prefix)
            except Exception as e:
                print(f"Error downloading file: {e}")
        else:
            extension_repo.append((path, file_name))

def manual_download(url, dst_dir, file_name, prefix=None):
    hf_header = f"--header='Authorization: Bearer {huggingface_token}'" if huggingface_token else ""
    aria2c_header = "--header='User-Agent: Mozilla/5.0' --allow-overwrite=true"
    aria2_args = "--optimize-concurrent-downloads --console-log-level=error --summary-interval=10 --stderr=true -c -x16 -s16 -k1M -j5"

    clean_url = strip_(url)

    if 'civitai' in url:
        url, clean_url, model_type, file_name, image_url, image_name, data = CivitAi_API(url, file_name)
        if image_url and image_name:
            command = ["aria2c"] + aria2_args.split() + ["-d", dst_dir, "-o", image_name, image_url]
            subprocess.run(command, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)

    elif 'github' in url or 'huggingface.co' in url:
        if file_name and '.' not in file_name:
            file_extension = f"{clean_url.split('/')[-1].split('.', 1)[1]}"
            file_name = f"{file_name}.{file_extension}"
        if not file_name:
            file_name = clean_url.split("/")[-1]

    """ Formatted info output """
    try:
        format_output(clean_url, dst_dir, file_name, image_name, image_url)
    except UnboundLocalError:
        format_output(clean_url, dst_dir, file_name, None, None)

    # =====================
    def run_aria2c(url, dst_dir, file_name=None, args="", header=""):
        file_path = os.path.join(dst_dir, file_name) # replaces config files
        if os.path.exists(file_path) and prefix == 'config':
            os.remove(file_path)

        out = f"-o '{file_name}'" if file_name else ""
        get_ipython().system("aria2c {header} {args} -d {dst_dir} {out} '{url}'")

    # -- Google Drive --
    if 'drive.google' in url:
        if not globals().get('have_drive_link', False):
            os.system("pip install -U gdown > /dev/null")
            globals()['have_drive_link'] = True

        if 'folders' in url:
            os.system(f"gdown --folder \"{url}\" -O {dst_dir} --fuzzy -c")
        else:
            out_path = f"{dst_dir}/{file_name}" if file_name else dst_dir
            os.system(f"gdown \"{url}\" -O {out_path} --fuzzy -c")

    # -- GitHub or Hugging Face --
    elif 'github' in url or 'huggingface' in url:
        run_aria2c(clean_url, dst_dir, file_name, aria2_args, hf_header if 'huggingface' in url else '')

    # -- Other HTTP/Sources --
    elif 'http' in url:
        run_aria2c(url, dst_dir, file_name, aria2_args, aria2c_header)

''' SubModels - Added URLs '''

# Separation of merged numbers
def split_numbers(num_str, max_num):
    result = []
    i = 0
    while i < len(num_str):
        found = False
        for length in range(2, 0, -1):
            if i + length <= len(num_str):
                part = int(num_str[i:i + length])
                if part <= max_num:
                    result.append(part)
                    i += length
                    found = True
                    break
        if not found:
            break
    return result

def add_submodels(selection, num_selection, model_dict, dst_dir):
    if selection == "none":
        return []
    selected_models = []

    if selection == "ALL":
        selected_models = sum(model_dict.values(), [])
    else:
        if selection in model_dict:
            selected_models.extend(model_dict[selection])

        nums = num_selection.replace(',', ' ').split()
        max_num = len(model_dict)
        unique_nums = set()

        for num_part in nums:
            split_nums = split_numbers(num_part, max_num)
            unique_nums.update(split_nums)

        for num in unique_nums:
            if 1 <= num <= max_num:
                name = list(model_dict.keys())[num - 1]
                selected_models.extend(model_dict[name])

    unique_models = {model['name']: model for model in selected_models}.values()

    for model in unique_models:
        model['dst_dir'] = dst_dir

    return list(unique_models)

def handle_submodels(selection, num_selection, model_dict, dst_dir, url):
    submodels = add_submodels(selection, num_selection, model_dict, dst_dir)
    for submodel in submodels:
        if not inpainting_model and "inpainting" in submodel['name']:
            continue
        url += f"{submodel['url']} {submodel['dst_dir']} {submodel['name']}, "
    return url

url = ""
url = handle_submodels(model, model_num, model_list, models_dir, url)
url = handle_submodels(vae, vae_num, vae_list, vaes_dir, url)
url = handle_submodels(controlnet, controlnet_num, controlnet_list, control_dir, url)

''' file.txt - added urls '''

def process_file_download(file_url, PREFIXES, unique_urls):
    files_urls = ""

    if file_url.startswith("http"):
        if "blob" in file_url:
            file_url = file_url.replace("blob", "raw")
        response = requests.get(file_url)
        lines = response.text.split('\n')
    else:
        with open(file_url, 'r') as file:
            lines = file.readlines()

    current_tag = None
    for line in lines:
        line = line.strip()
        if any(f'# {tag}' in line.lower() for tag in PREFIXES):
            current_tag = next((tag for tag in PREFIXES if tag in line.lower()))

        urls = [url.split('#')[0].strip() for url in line.split(',')]  # filter urls
        for url in urls:
            filter_url = url.split('[')[0] # same url filter

            if url.startswith("http") and filter_url not in unique_urls:
                files_urls += f"{current_tag}:{url}, "
                unique_urls.add(filter_url)

    return files_urls

file_urls = ""
unique_urls = set()

if custom_file_urls:
    for custom_file_url in custom_file_urls.replace(',', '').split():
        if not custom_file_url.endswith('.txt'):
            custom_file_url += '.txt'
        if not custom_file_url.startswith('http'):
            if not custom_file_url.startswith(root_path):
                custom_file_url = f'{root_path}/{custom_file_url}'

        try:
            file_urls += process_file_download(custom_file_url, PREFIXES, unique_urls)
        except FileNotFoundError:
            pass

# url prefixing
urls = (Model_url, Vae_url, LoRA_url, Embedding_url, Extensions_url)
prefixed_urls = (f"{prefix}:{url}" for prefix, url in zip(PREFIXES.keys(), urls) if url for url in url.replace(',', '').split())
url += ", ".join(prefixed_urls) + ", " + file_urls

if detailed_download == "on":
    print("\n\n\033[33m# ====== Detailed Download ====== #\n\033[0m")
    download(url)
    print("\n\033[33m# =============================== #\n\033[0m")
else:
    with capture.capture_output():
        download(url)

print("\r🏁 Download Complete!" + " "*15)


# Cleaning shit after downloading...
get_ipython().system('find {webui_path} \\( -type d \\( -name ".ipynb_checkpoints" -o -name ".aria2" \\) -o -type f -name "*.aria2" \\) -exec rm -r {{}} \\; >/dev/null 2>&1')


## Install of Custom extensions
if len(extension_repo) > 0:
    print("✨ Installing custom extensions...", end='')
    with capture.capture_output():
        for repo, repo_name in extension_repo:
            if not repo_name:
                repo_name = repo.split('/')[-1]
            get_ipython().system('cd {extensions_dir}                  && git clone {repo} {repo_name}                  && cd {repo_name}                  && git fetch')
    print(f"\r📦 Installed '{len(extension_repo)}', Custom extensions!")


## List Models and stuff V2
if detailed_download == "off":
    print("\n\n\033[33mIf you don't see any downloaded files, enable the 'Detailed Downloads' feature in the widget.")

get_ipython().run_line_magic('run', '{root_path}/file_cell/special/dl_display_results.py # display widgets result')