civitai_api.py

import gradio as gr
from huggingface_hub import HfApi, hf_hub_url
import os
from pathlib import Path
import gc
import requests
from requests.adapters import HTTPAdapter
from urllib3.util import Retry
from civitai_constants import PERIOD, SORT
from utils import (get_token, set_token, is_repo_exists, get_user_agent, get_download_file,
                   list_uniq, list_sub, duplicate_hf_repo, HF_SUBFOLDER_NAME, get_state, set_state)
import re
from PIL import Image
import json
import pandas as pd
import tempfile
import hashlib

# Huge shoutout to @John6666, saved me many hours.

TEMP_DIR = tempfile.mkdtemp()


def parse_urls(s):
    url_pattern = "https?://[\\w/:%#\\$&\\?\\(\\)~\\.=\\+\\-]+"
    try:
        urls = re.findall(url_pattern, s)
        return list(urls)
    except Exception:
        return []


def parse_repos(s):
    repo_pattern = r'[^\w_\-\.]?([\w_\-\.]+/[\w_\-\.]+)[^\w_\-\.]?'
    try:
        s = re.sub("https?://[\\w/:%#\\$&\\?\\(\\)~\\.=\\+\\-]+", "", s)
        repos = re.findall(repo_pattern, s)
        return list(repos)
    except Exception:
        return []


def to_urls(l: list[str]):
    return "\n".join(l)


def uniq_urls(s):
    return to_urls(list_uniq(parse_urls(s) + parse_repos(s)))


def upload_safetensors_to_repo(filename, repo_id, repo_type, is_private, progress=gr.Progress(track_tqdm=True)):
    output_filename = Path(filename).name
    hf_token = get_token()
    api = HfApi(token=hf_token)
    try:
        if not is_repo_exists(repo_id, repo_type): api.create_repo(repo_id=repo_id, repo_type=repo_type, token=hf_token, private=is_private)
        progress(0, desc=f"Start uploading... {filename} to {repo_id}")
        api.upload_file(path_or_fileobj=filename, path_in_repo=output_filename, repo_type=repo_type, revision="main", token=hf_token, repo_id=repo_id)
        progress(1, desc="Uploaded.")
        url = hf_hub_url(repo_id=repo_id, repo_type=repo_type, filename=output_filename)
    except Exception as e:
        print(f"Error: Failed to upload to {repo_id}. {e}")
        gr.Warning(f"Error: Failed to upload to {repo_id}. {e}")
        return None
    finally:
        if Path(filename).exists(): Path(filename).unlink()
    return url


def is_same_file(filename: str, cmp_sha256: str, cmp_size: int):
    if cmp_sha256:
        sha256_hash = hashlib.sha256()
        with open(filename, "rb") as f:
            for byte_block in iter(lambda: f.read(4096), b""):
                sha256_hash.update(byte_block)
        sha256 = sha256_hash.hexdigest()
    else: sha256 = ""
    size = os.path.getsize(filename)
    if size == cmp_size and sha256 == cmp_sha256: return True
    else: return False


def get_safe_filename(filename, repo_id, repo_type):
    hf_token = get_token()
    api = HfApi(token=hf_token)
    new_filename = filename
    try:
        i = 1
        while api.file_exists(repo_id=repo_id, filename=Path(new_filename).name, repo_type=repo_type, token=hf_token):
            infos = api.get_paths_info(repo_id=repo_id, paths=[Path(new_filename).name], repo_type=repo_type, token=hf_token)
            if infos and len(infos) == 1:
                repo_fs = infos[0].size
                repo_sha256 = infos[0].lfs.sha256 if infos[0].lfs is not None else ""
                if is_same_file(filename, repo_sha256, repo_fs): break
            new_filename = str(Path(Path(filename).parent, f"{Path(filename).stem}_{i}{Path(filename).suffix}"))
            i += 1
        if filename != new_filename:
            print(f"{Path(filename).name} is already exists but file content is different. renaming to {Path(new_filename).name}.")
            Path(filename).rename(new_filename)
    except Exception as e:
        print(f"Error occured when renaming {filename}. {e}")
    finally:
        return new_filename


def download_file(dl_url, civitai_key, progress=gr.Progress(track_tqdm=True)):
    download_dir = TEMP_DIR
    progress(0, desc=f"Start downloading... {dl_url}")
    output_filename = get_download_file(download_dir, dl_url, civitai_key)
    return output_filename


def save_civitai_info(dl_url, filename, civitai_key="", progress=gr.Progress(track_tqdm=True)):
    json_str, html_str, image_path = get_civitai_json(dl_url, True, filename, civitai_key)
    if not json_str: return  "", "", ""
    json_path = str(Path(TEMP_DIR, Path(filename).stem + ".json"))
    html_path = str(Path(TEMP_DIR, Path(filename).stem + ".html"))
    try:
        with open(json_path, 'w') as f:
            json.dump(json_str, f, indent=2)
        with open(html_path, mode='w', encoding="utf-8") as f:
            f.write(html_str)
        return json_path, html_path, image_path
    except Exception as e:
        print(f"Error: Failed to save info file {json_path}, {html_path} {e}")
        return "", "", ""


def upload_info_to_repo(dl_url, filename, repo_id, repo_type, is_private, civitai_key="", progress=gr.Progress(track_tqdm=True)):
    def upload_file(api, filename, repo_id, repo_type, hf_token):
        if not Path(filename).exists(): return
        api.upload_file(path_or_fileobj=filename, path_in_repo=Path(filename).name, repo_type=repo_type, revision="main", token=hf_token, repo_id=repo_id)
        Path(filename).unlink()

    hf_token = get_token()
    api = HfApi(token=hf_token)
    try:
        if not is_repo_exists(repo_id, repo_type): api.create_repo(repo_id=repo_id, repo_type=repo_type, token=hf_token, private=is_private)
        progress(0, desc=f"Downloading info... {filename}")
        json_path, html_path, image_path = save_civitai_info(dl_url, filename, civitai_key)
        progress(0, desc=f"Start uploading info... {filename} to {repo_id}")
        if not json_path: return
        else: upload_file(api, json_path, repo_id, repo_type, hf_token)
        if html_path: upload_file(api, html_path, repo_id, repo_type, hf_token)
        if image_path: upload_file(api, image_path, repo_id, repo_type, hf_token)
        progress(1, desc="Info uploaded.")
        return
    except Exception as e:
        print(f"Error: Failed to upload info to {repo_id}. {e}")
        gr.Warning(f"Error: Failed to upload info to {repo_id}. {e}")
        return


def download_civitai(dl_url, civitai_key, hf_token, urls, 
                     newrepo_id, repo_type="model", is_private=True, is_info=False, is_rename=True, progress=gr.Progress(track_tqdm=True)):
    if hf_token: set_token(hf_token)
    else: set_token(os.environ.get("HF_TOKEN")) # default huggingface write token
    if not civitai_key: civitai_key = os.environ.get("CIVITAI_API_KEY") # default Civitai API key
    if not newrepo_id: newrepo_id = os.environ.get("HF_REPO") # default repo to upload
    if not get_token() or not civitai_key: raise gr.Error("HF write token and Civitai API key is required.")
    if not urls: urls = []
    dl_urls = parse_urls(dl_url)
    remain_urls = dl_urls.copy()
    try:
        md = f'### Your repo: [{newrepo_id}]({"https://huggingface.co/datasets/" if repo_type == "dataset" else "https://huggingface.co/"}{newrepo_id})\n'
        for u in dl_urls:
            file = download_file(u, civitai_key)
            if not Path(file).exists() or not Path(file).is_file(): continue
            if is_rename: file = get_safe_filename(file, newrepo_id, repo_type)
            url = upload_safetensors_to_repo(file, newrepo_id, repo_type, is_private)
            if url:
                if is_info: upload_info_to_repo(u, file, newrepo_id, repo_type, is_private, civitai_key)
                urls.append(url)
                remain_urls.remove(u)
                md += f"- Uploaded [{str(u)}]({str(u)})\n"
        dp_repos = parse_repos(dl_url)
        for r in dp_repos:
            url = duplicate_hf_repo(r, newrepo_id, "model", repo_type, is_private, HF_SUBFOLDER_NAME[1])
            if url: urls.append(url)
        return gr.update(value=urls, choices=urls), gr.update(value=md), gr.update(value="\n".join(remain_urls), visible=False)
    except Exception as e:
        gr.Info(f"Error occured: {e}")
        return gr.update(value=urls, choices=urls), gr.update(value=md), gr.update(value="\n".join(remain_urls), visible=True)
    finally:
        gc.collect()


def search_on_civitai(query: str, types: list[str], allow_model: list[str] = [], limit: int = 100,
                      sort: str = "Highest Rated", period: str = "AllTime", tag: str = "", user: str = "", page: int = 1,
                      filetype: list[str] = [], api_key: str = "", progress=gr.Progress(track_tqdm=True)):
    user_agent = get_user_agent()
    headers = {'User-Agent': user_agent, 'content-type': 'application/json'}
    if api_key: headers['Authorization'] = f'Bearer {{{api_key}}}'
    base_url = 'https://civitai.com/api/v1/models'
    params = {'sort': sort, 'period': period, 'limit': int(limit), 'nsfw': 'true'}
    if len(types) != 0: params["types"] = types
    if query: params["query"] = query
    if tag: params["tag"] = tag
    if user: params["username"] = user
    if page != 0: params["page"] = int(page)
    session = requests.Session()
    retries = Retry(total=5, backoff_factor=1, status_forcelist=[500, 502, 503, 504])
    session.mount("https://", HTTPAdapter(max_retries=retries))
    rs = []
    try:
        if page == 0:
            progress(0, desc="Searching page 1...")
            print("Searching page 1...")
            r = session.get(base_url, params=params | {'page': 1}, headers=headers, stream=True, timeout=(7.0, 30))
            rs.append(r)
            if r.ok:
                json = r.json()
                next_url = json['metadata']['nextPage'] if 'metadata' in json and 'nextPage' in json['metadata'] else None
                i = 2
                while(next_url is not None):
                    progress(0, desc=f"Searching page {i}...")
                    print(f"Searching page {i}...")
                    r = session.get(next_url, headers=headers, stream=True, timeout=(7.0, 30))
                    rs.append(r)
                    if r.ok:
                        json = r.json()
                        next_url = json['metadata']['nextPage'] if 'metadata' in json and 'nextPage' in json['metadata'] else None
                    else: next_url = None
                    i += 1
        else:
            progress(0, desc="Searching page 1...")
            print("Searching page 1...")
            r = session.get(base_url, params=params, headers=headers, stream=True, timeout=(7.0, 30))
            rs.append(r)
    except requests.exceptions.ConnectTimeout:
        print("Request timed out.")
    except Exception as e:
        print(e)
    items = []
    for r in rs:
        if not r.ok: continue
        json = r.json()
        if 'items' not in json: continue
        for j in json['items']:
            for model in j['modelVersions']:
                item = {}
                if len(allow_model) != 0 and model['baseModel'] not in set(allow_model): continue
                item['name'] = j['name']
                item['creator'] = j['creator']['username'] if 'creator' in j.keys() and 'username' in j['creator'].keys() else ""
                item['tags'] = j['tags'] if 'tags' in j.keys() else []
                item['model_name'] = model['name'] if 'name' in model.keys() else ""
                item['base_model'] = model['baseModel'] if 'baseModel' in model.keys() else ""
                item['description'] = model['description'] if 'description' in model.keys() else ""
                item['md'] = ""
                
                # Handle both images and videos
                if 'images' in model.keys() and len(model["images"]) != 0:
                    first_media = model["images"][0]
                    item['img_url'] = first_media["url"]
                    item['is_video'] = first_media.get("type", "image") == "video"
                    item['video_url'] = first_media.get("meta", {}).get("video", "") if item['is_video'] else ""
                    
                    if item['is_video']:
                        item['md'] += f'<video src="{item["img_url"]}" poster="{item["img_url"]}" muted loop autoplay width="300" height="480" style="float:right;padding:16px;"></video><br>'
                    else:
                        item['md'] += f'<img src="{item["img_url"]}#float" alt="thumbnail" width="150" height="240"><br>'
                else:
                    item['img_url'] = "/home/user/app/null.png"
                    item['is_video'] = False
                    item['video_url'] = ""
                
                item['md'] += f'''Model URL: [https://civitai.com/models/{j["id"]}](https://civitai.com/models/{j["id"]})<br>Model Name: {item["name"]}<br>
                    Creator: {item["creator"]}<br>Tags: {", ".join(item["tags"])}<br>Base Model: {item["base_model"]}<br>Description: {item["description"]}'''
                if 'files' in model.keys():
                    for f in model['files']:
                        i = item.copy()
                        i['dl_url'] = f['downloadUrl']
                        if len(filetype) != 0 and f['type'] not in set(filetype): continue
                        items.append(i)
                else:
                    item['dl_url'] = model['downloadUrl']
                    items.append(item)
    return items if len(items) > 0 else None


def search_civitai(query, types, base_model=[], sort=SORT[0], period=PERIOD[0], tag="", user="", limit=100, page=1,
                   filetype=[], api_key="", gallery=[], state={}, progress=gr.Progress(track_tqdm=True)):
    civitai_last_results = {}
    set_state(state, "civitai_last_choices", [("", "")])
    set_state(state, "civitai_last_gallery", [])
    set_state(state, "civitai_last_results", civitai_last_results)
    results_info = "No item found."
    items = search_on_civitai(query, types, base_model, int(limit), sort, period, tag, user, int(page), filetype, api_key)
    if not items: return gr.update(choices=[("", "")], value=[], visible=True),\
          gr.update(value="", visible=False), gr.update(), gr.update(), gr.update(), gr.update(), results_info, state
    choices = []
    gallery = []
    for item in items:
        base_model_name = "Pony🐴" if item['base_model'] == "Pony" else item['base_model']
        name = f"{item['name']} (for {base_model_name} / By: {item['creator']})"
        value = item['dl_url']
        choices.append((name, value))
        
        # For gallery, use tuples with HTML that includes both image and video
        if item.get('is_video') and item.get('video_url'):
            # Create an HTML element that contains both image and video
            media_html = f"""
            <div class="media-container">
                <img src="{item['img_url']}" alt="{name}">
                <video src="{item['video_url']}" muted loop poster="{item['img_url']}"></video>
            </div>
            """
            gallery.append((item['img_url'], name))  # Keep using image as thumbnail
        else:
            gallery.append((item['img_url'], name))
            
        civitai_last_results[value] = item
    if len(choices) >= 1: 
        results_info = f"{int(len(choices))} items found."
    else: 
        choices = [("", "")]
        
    md = ""
    set_state(state, "civitai_last_choices", choices)
    set_state(state, "civitai_last_gallery", gallery)
    set_state(state, "civitai_last_results", civitai_last_results)
    
    return gr.update(choices=choices, value=[], visible=True),\
           gr.update(value=md, visible=True),\
           gr.update(),\
           gr.update(),\
           gr.update(value=gallery),\
           gr.update(choices=choices, value=[]),\
           results_info,\
           state


def get_civitai_json(dl_url: str, is_html: bool=False, image_baseurl: str="", api_key=""):
    if not image_baseurl: image_baseurl = dl_url
    default = ("", "", "") if is_html else ""
    if "https://civitai.com/api/download/models/" not in dl_url: return default
    user_agent = get_user_agent()
    headers = {'User-Agent': user_agent, 'content-type': 'application/json'}
    if api_key: headers['Authorization'] = f'Bearer {{{api_key}}}'
    base_url = 'https://civitai.com/api/v1/model-versions/'
    params = {}
    session = requests.Session()
    retries = Retry(total=5, backoff_factor=1, status_forcelist=[500, 502, 503, 504])
    session.mount("https://", HTTPAdapter(max_retries=retries))
    model_id = re.sub('https://civitai.com/api/download/models/(\\d+)(?:.+)?', '\\1', dl_url)
    url = base_url + model_id
    #url = base_url + str(dl_url.split("/")[-1])
    try:
        r = session.get(url, params=params, headers=headers, stream=True, timeout=(5.0, 15))
        if not r.ok: return default
        json = dict(r.json()).copy()
        html = ""
        image = ""
        if "modelId" in json.keys():
            url = f"https://civitai.com/models/{json['modelId']}"
            r = session.get(url, params=params, headers=headers, stream=True, timeout=(5.0, 15))
            if not r.ok: return json, html, image
            html = r.text
        if 'images' in json.keys() and len(json["images"]) != 0:
            url = json["images"][0]["url"]
            r = session.get(url, params=params, headers=headers, stream=True, timeout=(5.0, 15))
            if not r.ok: return json, html, image
            image_temp = str(Path(TEMP_DIR, "image" + Path(url.split("/")[-1]).suffix))
            image = str(Path(TEMP_DIR, Path(image_baseurl.split("/")[-1]).stem + ".png"))
            with open(image_temp, 'wb') as f:
                f.write(r.content)
            Image.open(image_temp).convert('RGBA').save(image)
        return json, html, image
    except Exception as e:
        print(e)
        return default


def get_civitai_tag():
    default = [""]
    user_agent = get_user_agent()
    headers = {'User-Agent': user_agent, 'content-type': 'application/json'}
    base_url = 'https://civitai.com/api/v1/tags'
    params = {'limit': 200}
    session = requests.Session()
    retries = Retry(total=5, backoff_factor=1, status_forcelist=[500, 502, 503, 504])
    session.mount("https://", HTTPAdapter(max_retries=retries))
    url = base_url
    try:
        r = session.get(url, params=params, headers=headers, stream=True, timeout=(7.0, 15))
        if not r.ok: return default
        j = dict(r.json()).copy()
        if "items" not in j.keys(): return default
        items = []
        for item in j["items"]:
            items.append([str(item.get("name", "")), int(item.get("modelCount", 0))])
        df = pd.DataFrame(items)
        df.sort_values(1, ascending=False)
        tags = df.values.tolist()
        tags = [""] + [l[0] for l in tags]
        return tags
    except Exception as e:
        print(e)
        return default


def select_civitai_item(results: list[str], state: dict):
    json = {}
    if "http" not in "".join(results) or len(results) == 0: return gr.update(value="", visible=True), gr.update(value=json, visible=False), state
    result = get_state(state, "civitai_last_results")
    last_selects = get_state(state, "civitai_last_selects")
    selects = list_sub(results, last_selects if last_selects else [])
    md = result.get(selects[-1]).get('md', "") if result and isinstance(result, dict) and len(selects) > 0 else ""
    set_state(state, "civitai_last_selects", results)
    return gr.update(value=md, visible=True), gr.update(value=json, visible=False), state


def add_civitai_item(results: list[str], dl_url: str):
    if "http" not in "".join(results): return gr.update(value=dl_url)
    new_url = dl_url if dl_url else ""
    for result in results:
        if "http" not in result: continue
        new_url += f"\n{result}" if new_url else f"{result}"
    new_url = uniq_urls(new_url)
    return gr.update(value=new_url)


def select_civitai_all_item(button_name: str, state: dict):
    if button_name not in ["Select All", "Deselect All"]: return gr.update(value=button_name), gr.Update(visible=True)
    civitai_last_choices = get_state(state, "civitai_last_choices")
    selected = [t[1] for t in civitai_last_choices if t[1] != ""] if button_name == "Select All" else []
    new_button_name = "Select All" if button_name == "Deselect All" else "Deselect All"
    return gr.update(value=new_button_name), gr.update(value=selected, choices=civitai_last_choices)


def update_civitai_selection(evt: gr.SelectData, value: list[str], state: dict):
    try:
        civitai_last_choices = get_state(state, "civitai_last_choices")
        selected_index = evt.index
        selected = list_uniq([v for v in value if v != ""] + [civitai_last_choices[selected_index][1]])
        return gr.update(value=selected)
    except Exception:
        return gr.update()


def update_civitai_checkbox(selected: list[str]):
    return gr.update(value=selected)


def from_civitai_checkbox(selected: list[str]):
    return gr.update(value=selected)