Upload 6 files

Integrated "Gelbooru Image Fetcher" tab which is a big change, it will help you to find characters and other stuff pretty easy and faster, still working on it, so maybe in the future there will be even more features. Also fixed some other minor stuff and changed the description accordingly. Have fun!

app.py

@@ -1,593 +1,641 @@
-import os
-import io,copy,requests,spaces,gradio as gr,numpy as np
-from transformers import AutoProcessor,AutoModelForCausalLM
-from PIL import Image,ImageDraw,ImageFont
-from unittest.mock import patch
-import argparse,huggingface_hub,onnxruntime as rt,pandas as pd,traceback,tempfile,zipfile,re,ast,time
-from datetime import datetime,timezone
-from collections import defaultdict
-from apscheduler.schedulers.background import BackgroundScheduler
-import json
-from modules.classifyTags import classify_tags,process_tags
-from modules.florence2 import process_image,single_task_list,update_task_dropdown
-from modules.reorganizer_model import reorganizer_list,reorganizer_class
-from modules.tag_enhancer import prompt_enhancer
-os.environ['PYTORCH_ENABLE_MPS_FALLBACK']='1'
-
-TITLE = "Multi-Tagger"
-DESCRIPTION = """
-Multi-Tagger is a versatile application that combines the Waifu Diffusion and Florence 2 models for advanced image analysis and captioning. Perfect for AI artists and enthusiasts, it offers a range of features:
-
-- Batch processing for multiple images
-- Multi-category tagging with structured tag display.
-- CUDA or CPU support.
-- Image tagging, various captioning tasks which includes: Caption, Detailed Caption, Object Detection with visual outputs and much more.
-
-Example image by [me.](https://huggingface.co/Werli)
-"""
-
-# Dataset v3 series of models:
-SWINV2_MODEL_DSV3_REPO = "SmilingWolf/wd-swinv2-tagger-v3"
-CONV_MODEL_DSV3_REPO = "SmilingWolf/wd-convnext-tagger-v3"
-VIT_MODEL_DSV3_REPO = "SmilingWolf/wd-vit-tagger-v3"
-VIT_LARGE_MODEL_DSV3_REPO = "SmilingWolf/wd-vit-large-tagger-v3"
-EVA02_LARGE_MODEL_DSV3_REPO = "SmilingWolf/wd-eva02-large-tagger-v3"
-# Dataset v2 series of models:
-MOAT_MODEL_DSV2_REPO = "SmilingWolf/wd-v1-4-moat-tagger-v2"
-SWIN_MODEL_DSV2_REPO = "SmilingWolf/wd-v1-4-swinv2-tagger-v2"
-CONV_MODEL_DSV2_REPO = "SmilingWolf/wd-v1-4-convnext-tagger-v2"
-CONV2_MODEL_DSV2_REPO = "SmilingWolf/wd-v1-4-convnextv2-tagger-v2"
-VIT_MODEL_DSV2_REPO = "SmilingWolf/wd-v1-4-vit-tagger-v2"
-# IdolSankaku series of models:
-EVA02_LARGE_MODEL_IS_DSV1_REPO = "deepghs/idolsankaku-eva02-large-tagger-v1"
-SWINV2_MODEL_IS_DSV1_REPO = "deepghs/idolsankaku-swinv2-tagger-v1"
-# Files to download from the repos
-MODEL_FILENAME = "model.onnx"
-LABEL_FILENAME = "selected_tags.csv"
-
-kaomojis=['0_0','(o)_(o)','+_+','+_-','._.','<o>_<o>','<|>_<|>','=_=','>_<','3_3','6_9','>_o','@_@','^_^','o_o','u_u','x_x','|_|','||_||']
-def parse_args()->argparse.Namespace:parser=argparse.ArgumentParser();parser.add_argument('--score-slider-step',type=float,default=.05);parser.add_argument('--score-general-threshold',type=float,default=.35);parser.add_argument('--score-character-threshold',type=float,default=.85);parser.add_argument('--share',action='store_true');return parser.parse_args()
-def load_labels(dataframe)->list[str]:name_series=dataframe['name'];name_series=name_series.map(lambda x:x.replace('_',' ')if x not in kaomojis else x);tag_names=name_series.tolist();rating_indexes=list(np.where(dataframe['category']==9)[0]);general_indexes=list(np.where(dataframe['category']==0)[0]);character_indexes=list(np.where(dataframe['category']==4)[0]);return tag_names,rating_indexes,general_indexes,character_indexes
-def mcut_threshold(probs):sorted_probs=probs[probs.argsort()[::-1]];difs=sorted_probs[:-1]-sorted_probs[1:];t=difs.argmax();thresh=(sorted_probs[t]+sorted_probs[t+1])/2;return thresh
-
-class Timer:
-	def __init__(self):self.start_time=time.perf_counter();self.checkpoints=[('Start',self.start_time)]
-	def checkpoint(self,label='Checkpoint'):now=time.perf_counter();self.checkpoints.append((label,now))
-	def report(self,is_clear_checkpoints=True):
-		max_label_length=max(len(label)for(label,_)in self.checkpoints);prev_time=self.checkpoints[0][1]
-		for(label,curr_time)in self.checkpoints[1:]:elapsed=curr_time-prev_time;print(f"{label.ljust(max_label_length)}: {elapsed:.3f} seconds");prev_time=curr_time
-		if is_clear_checkpoints:self.checkpoints.clear();self.checkpoint()
-	def report_all(self):
-		print('\n> Execution Time Report:');max_label_length=max(len(label)for(label,_)in self.checkpoints)if len(self.checkpoints)>0 else 0;prev_time=self.start_time
-		for(label,curr_time)in self.checkpoints[1:]:elapsed=curr_time-prev_time;print(f"{label.ljust(max_label_length)}: {elapsed:.3f} seconds");prev_time=curr_time
-		total_time=self.checkpoints[-1][1]-self.start_time;print(f"{'Total Execution Time'.ljust(max_label_length)}: {total_time:.3f} seconds\n");self.checkpoints.clear()
-	def restart(self):self.start_time=time.perf_counter();self.checkpoints=[('Start',self.start_time)]
-class Predictor:
-    def __init__(self):
-        self.model_target_size = None
-        self.last_loaded_repo = None
-    def download_model(self, model_repo):
-        csv_path = huggingface_hub.hf_hub_download(
-            model_repo,
-            LABEL_FILENAME,
-        )
-        model_path = huggingface_hub.hf_hub_download(
-            model_repo,
-            MODEL_FILENAME,
-        )
-        return csv_path, model_path
-    def load_model(self, model_repo):
-        if model_repo == self.last_loaded_repo:
-            return
-
-        csv_path, model_path = self.download_model(model_repo)
-
-        tags_df = pd.read_csv(csv_path)
-        sep_tags = load_labels(tags_df)
-
-        self.tag_names = sep_tags[0]
-        self.rating_indexes = sep_tags[1]
-        self.general_indexes = sep_tags[2]
-        self.character_indexes = sep_tags[3]
-
-        model = rt.InferenceSession(model_path)
-        _, height, width, _ = model.get_inputs()[0].shape
-        self.model_target_size = height
-
-        self.last_loaded_repo = model_repo
-        self.model = model
-    def prepare_image(self, path):
-        image = Image.open(path)
-        image = image.convert("RGBA")
-        target_size = self.model_target_size
-
-        canvas = Image.new("RGBA", image.size, (255, 255, 255))
-        canvas.alpha_composite(image)
-        image = canvas.convert("RGB")
-
-        # Pad image to square
-        image_shape = image.size
-        max_dim = max(image_shape)
-        pad_left = (max_dim - image_shape[0]) // 2
-        pad_top = (max_dim - image_shape[1]) // 2
-
-        padded_image = Image.new("RGB", (max_dim, max_dim), (255, 255, 255))
-        padded_image.paste(image, (pad_left, pad_top))
-
-        # Resize
-        if max_dim != target_size:
-            padded_image = padded_image.resize(
-                (target_size, target_size),
-                Image.BICUBIC,
-            )
-        # Convert to numpy array
-        image_array = np.asarray(padded_image, dtype=np.float32)
-        # Convert PIL-native RGB to BGR
-        image_array = image_array[:, :, ::-1]
-        return np.expand_dims(image_array, axis=0)
-
-    def create_file(self, content: str, directory: str, fileName: str) -> str:
-        # Write the content to a file
-        file_path = os.path.join(directory, fileName)
-        if fileName.endswith('.json'):
-            with open(file_path, 'w', encoding="utf-8") as file:
-                file.write(content)
-        else:
-            with open(file_path, 'w+', encoding="utf-8") as file:
-                file.write(content)
-
-        return file_path
-
-    def predict(
-        self,
-        gallery,
-        model_repo,
-        general_thresh,
-        general_mcut_enabled,
-        character_thresh,
-        character_mcut_enabled,
-        characters_merge_enabled,
-        reorganizer_model_repo,
-        additional_tags_prepend,
-        additional_tags_append,
-        tag_results,
-        progress=gr.Progress()
-        ):
-            # Clear tag_results before starting a new prediction
-            tag_results.clear()
-
-            gallery_len = len(gallery)
-            print(f"Predict load model: {model_repo}, gallery length: {gallery_len}")
-
-            timer = Timer()  # Create a timer
-            progressRatio = 0.5 if reorganizer_model_repo else 1
-            progressTotal = gallery_len + 1
-            current_progress = 0
-
-            self.load_model(model_repo)
-            current_progress += progressRatio/progressTotal;
-            progress(current_progress, desc="Initialize wd model finished")
-            timer.checkpoint(f"Initialize wd model")
-
-            txt_infos = []
-            output_dir = tempfile.mkdtemp()
-            if not os.path.exists(output_dir):
-                os.makedirs(output_dir)
-
-            sorted_general_strings = ""
-            # Create categorized output string
-            categorized_output_strings = []
-            rating = None
-            character_res = None
-            general_res = None
-
-            if reorganizer_model_repo:
-                print(f"Reorganizer load model {reorganizer_model_repo}")
-                reorganizer = reorganizer_class(reorganizer_model_repo, loadModel=True)
-                current_progress += progressRatio/progressTotal;
-                progress(current_progress, desc="Initialize reoganizer model finished")
-                timer.checkpoint(f"Initialize reoganizer model")
-                
-            timer.report()
-
-            prepend_list = [tag.strip() for tag in additional_tags_prepend.split(",") if tag.strip()]
-            append_list = [tag.strip() for tag in additional_tags_append.split(",") if tag.strip()]
-            if prepend_list and append_list:
-                append_list = [item for item in append_list if item not in prepend_list]
-                
-            # Dictionary to track counters for each filename
-            name_counters = defaultdict(int)
-            
-            for idx, value in enumerate(gallery):
-                try:
-                    image_path = value[0]
-                    image_name = os.path.splitext(os.path.basename(image_path))[0]
-
-                    # Increment the counter for the current name
-                    name_counters[image_name] += 1
-                    
-                    if name_counters[image_name] > 1:
-                        image_name = f"{image_name}_{name_counters[image_name]:02d}"
-
-                    image = self.prepare_image(image_path)
-
-                    input_name = self.model.get_inputs()[0].name
-                    label_name = self.model.get_outputs()[0].name
-                    print(f"Gallery {idx:02d}: Starting run wd model...")
-                    preds = self.model.run([label_name], {input_name: image})[0]
-
-                    labels = list(zip(self.tag_names, preds[0].astype(float)))
-
-                    # First 4 labels are actually ratings: pick one with argmax
-                    ratings_names = [labels[i] for i in self.rating_indexes]
-                    rating = dict(ratings_names)
-
-                    # Then we have general tags: pick any where prediction confidence > threshold
-                    general_names = [labels[i] for i in self.general_indexes]
-
-                    if general_mcut_enabled:
-                        general_probs = np.array([x[1] for x in general_names])
-                        general_thresh = mcut_threshold(general_probs)
-
-                    general_res = [x for x in general_names if x[1] > general_thresh]
-                    general_res = dict(general_res)
-
-                    # Everything else is characters: pick any where prediction confidence > threshold
-                    character_names = [labels[i] for i in self.character_indexes]
-
-                    if character_mcut_enabled:
-                        character_probs = np.array([x[1] for x in character_names])
-                        character_thresh = mcut_threshold(character_probs)
-                        character_thresh = max(0.15, character_thresh)
-
-                    character_res = [x for x in character_names if x[1] > character_thresh]
-                    character_res = dict(character_res)
-                    character_list = list(character_res.keys())
-
-                    sorted_general_list = sorted(
-                        general_res.items(),
-                        key=lambda x: x[1],
-                        reverse=True,
-                    )
-                    sorted_general_list = [x[0] for x in sorted_general_list]
-                    # Remove values from character_list that already exist in sorted_general_list
-                    character_list = [item for item in character_list if item not in sorted_general_list]
-                    # Remove values from sorted_general_list that already exist in prepend_list or append_list
-                    if prepend_list:
-                        sorted_general_list = [item for item in sorted_general_list if item not in prepend_list]
-                    if append_list:
-                        sorted_general_list = [item for item in sorted_general_list if item not in append_list]
-
-                    sorted_general_list = prepend_list + sorted_general_list + append_list
-
-                    sorted_general_strings = ", ".join((character_list if characters_merge_enabled else []) + sorted_general_list).replace("(", "\(").replace(")", "\)")
-
-                    classified_tags, unclassified_tags = classify_tags(sorted_general_list)
-                    
-                    # Create a single string of ALL categorized tags for the current image
-                    categorized_output_string = ', '.join([', '.join(tags) for tags in classified_tags.values()])
-                    categorized_output_strings.append(categorized_output_string)
-                    # Collect all categorized output strings into a single string
-                    final_categorized_output = ', '.join(categorized_output_strings)
-                    
-                    # Create a .txt file for "Output (string)" and "Categorized Output (string)"
-                    txt_content = f"Output (string): {sorted_general_strings}\nCategorized Output (string): {final_categorized_output}"
-                    txt_file = self.create_file(txt_content, output_dir, f"{image_name}_output.txt")
-                    txt_infos.append({"path": txt_file, "name": f"{image_name}_output.txt"})
-
-                    # Create a .json file for "Categorized (tags)"
-                    json_content = json.dumps(classified_tags, indent=4)
-                    json_file = self.create_file(json_content, output_dir, f"{image_name}_categorized_tags.json")
-                    txt_infos.append({"path": json_file, "name": f"{image_name}_categorized_tags.json"})
-
-                    # Save a copy of the uploaded image in PNG format
-                    image_path = value[0]
-                    image = Image.open(image_path)
-                    image.save(os.path.join(output_dir, f"{image_name}.png"), format="PNG")
-                    txt_infos.append({"path": os.path.join(output_dir, f"{image_name}.png"), "name": f"{image_name}.png"})
-                    
-                    current_progress += progressRatio/progressTotal;
-                    progress(current_progress, desc=f"image{idx:02d}, predict finished")
-                    timer.checkpoint(f"image{idx:02d}, predict finished")
-                    
-                    if reorganizer_model_repo:
-                        print(f"Starting reorganizer...")
-                        reorganize_strings = reorganizer.reorganize(sorted_general_strings)
-                        reorganize_strings = re.sub(r" *Title: *", "", reorganize_strings)
-                        reorganize_strings = re.sub(r"\n+", ",", reorganize_strings)
-                        reorganize_strings = re.sub(r",,+", ",", reorganize_strings)
-                        sorted_general_strings += ",\n\n" + reorganize_strings
-
-                        current_progress += progressRatio/progressTotal;
-                        progress(current_progress, desc=f"image{idx:02d}, reorganizer finished")
-                        timer.checkpoint(f"image{idx:02d}, reorganizer finished")
-
-                    txt_file = self.create_file(sorted_general_strings, output_dir, image_name + ".txt")
-                    txt_infos.append({"path":txt_file, "name": image_name + ".txt"})
-                    
-                    # Store the result in tag_results using image_path as the key
-                    tag_results[image_path] = { 
-                        "strings": sorted_general_strings, 
-                        "strings2": categorized_output_string,  # Store the categorized output string here
-                        "classified_tags": classified_tags, 
-                        "rating": rating, 
-                        "character_res": character_res, 
-                        "general_res": general_res, 
-                        "unclassified_tags": unclassified_tags,
-                        "enhanced_tags": ""  # Initialize as empty string
-                    }
-
-                    timer.report()
-                except Exception as e:
-                    print(traceback.format_exc())
-                    print("Error predict: " + str(e))
-            # Zip creation logic:
-            download = []
-            if txt_infos is not None and len(txt_infos) > 0:
-                downloadZipPath = os.path.join(output_dir, "Multi-tagger-" + datetime.now().strftime("%Y%m%d-%H%M%S") + ".zip")
-                with zipfile.ZipFile(downloadZipPath, 'w', zipfile.ZIP_DEFLATED) as taggers_zip:
-                    for info in txt_infos:
-                        # Get file name from lookup
-                        taggers_zip.write(info["path"], arcname=info["name"])
-                download.append(downloadZipPath)
-            # End zip creation logic
-            if reorganizer_model_repo:
-                reorganizer.release_vram()
-                del reorganizer
-                 
-            progress(1, desc=f"Predict completed")
-            timer.report_all()  # Print all recorded times
-            print("Predict is complete.")
-            
-            return download, sorted_general_strings, final_categorized_output, classified_tags, rating, character_res, general_res, unclassified_tags, tag_results
-def get_selection_from_gallery(gallery: list, tag_results: dict, selected_state: gr.SelectData):
-    if not selected_state:
-        return selected_state
-    tag_result = { 
-        "strings": "", 
-        "strings2": "", 
-        "classified_tags": "{}", 
-        "rating": "", 
-        "character_res": "", 
-        "general_res": "", 
-        "unclassified_tags": "{}", 
-        "enhanced_tags": "" 
-    }
-    if selected_state.value["image"]["path"] in tag_results:
-        tag_result = tag_results[selected_state.value["image"]["path"]]
-    return (selected_state.value["image"]["path"], selected_state.value["caption"]), tag_result["strings"], tag_result["strings2"], tag_result["classified_tags"], tag_result["rating"], tag_result["character_res"], tag_result["general_res"], tag_result["unclassified_tags"], tag_result["enhanced_tags"]
-def append_gallery(gallery:list,image:str):
-	if gallery is None:gallery=[]
-	if not image:return gallery,None
-	gallery.append(image);return gallery,None
-def extend_gallery(gallery:list,images):
-	if gallery is None:gallery=[]
-	if not images:return gallery
-	gallery.extend(images);return gallery
-def remove_image_from_gallery(gallery:list,selected_image:str):
-	if not gallery or not selected_image:return gallery
-	selected_image=ast.literal_eval(selected_image)
-	if selected_image in gallery:gallery.remove(selected_image)
-	return gallery
-args = parse_args()
-predictor = Predictor()
-dropdown_list = [
-    EVA02_LARGE_MODEL_DSV3_REPO,
-    SWINV2_MODEL_DSV3_REPO,
-    CONV_MODEL_DSV3_REPO,
-    VIT_MODEL_DSV3_REPO,
-    VIT_LARGE_MODEL_DSV3_REPO,
-    # ---
-    MOAT_MODEL_DSV2_REPO,
-    SWIN_MODEL_DSV2_REPO,
-    CONV_MODEL_DSV2_REPO,
-    CONV2_MODEL_DSV2_REPO,
-    VIT_MODEL_DSV2_REPO,
-    # ---
-    SWINV2_MODEL_IS_DSV1_REPO,
-    EVA02_LARGE_MODEL_IS_DSV1_REPO,
-]
-
-def _restart_space():
-	HF_TOKEN=os.getenv('HF_TOKEN')
-	if not HF_TOKEN:raise ValueError('HF_TOKEN environment variable is not set.')
-	huggingface_hub.HfApi().restart_space(repo_id='Werli/Multi-Tagger',token=HF_TOKEN,factory_reboot=False)
-scheduler=BackgroundScheduler()
-# Add a job to restart the space every 2 days (172800 seconds)
-restart_space_job = scheduler.add_job(_restart_space, "interval", seconds=172800)
-scheduler.start()
-next_run_time_utc=restart_space_job.next_run_time.astimezone(timezone.utc)
-NEXT_RESTART=f"Next Restart: {next_run_time_utc.strftime('%Y-%m-%d %H:%M:%S')} (UTC) - The space will restart every 2 days to ensure stability and performance. It uses a background scheduler to handle the restart process."
-
-css = """
-#output {height: 500px; overflow: auto; border: 1px solid #ccc;}
-label.float.svelte-i3tvor {position: relative !important;}
-.reduced-height.svelte-11chud3 {height: calc(80% - var(--size-10));}
-"""
-
-with gr.Blocks(title=TITLE, css=css, theme="Werli/Multi-Tagger", fill_width=True) as demo:
-    gr.Markdown(value=f"<h1 style='text-align: center; margin-bottom: 1rem'>{TITLE}</h1>")
-    gr.Markdown(value=DESCRIPTION)
-    gr.Markdown(NEXT_RESTART)
-    with gr.Tab(label="Waifu Diffusion"):
-        with gr.Row():
-            with gr.Column():
-                submit = gr.Button(value="Submit", variant="primary", size="lg")
-                with gr.Column(variant="panel"):
-                    # Create an Image component for uploading images
-                    image_input = gr.Image(label="Upload an Image or clicking paste from clipboard button", type="filepath", sources=["upload", "clipboard"], height=150)
-                    with gr.Row():
-                        upload_button = gr.UploadButton("Upload multiple images", file_types=["image"], file_count="multiple", size="sm")
-                        remove_button = gr.Button("Remove Selected Image", size="sm")
-                    gallery = gr.Gallery(columns=5, rows=5, show_share_button=False, interactive=True, height="500px", label="Grid of images")
-                model_repo = gr.Dropdown(
-                    dropdown_list,
-                    value=EVA02_LARGE_MODEL_DSV3_REPO,
-                    label="Model",
-                )
-                with gr.Row():
-                    general_thresh = gr.Slider(
-                        0,
-                        1,
-                        step=args.score_slider_step,
-                        value=args.score_general_threshold,
-                        label="General Tags Threshold",
-                        scale=3,
-                    )
-                    general_mcut_enabled = gr.Checkbox(
-                        value=False,
-                        label="Use MCut threshold",
-                        scale=1,
-                    )
-                with gr.Row():
-                    character_thresh = gr.Slider(
-                        0,
-                        1,
-                        step=args.score_slider_step,
-                        value=args.score_character_threshold,
-                        label="Character Tags Threshold",
-                        scale=3,
-                    )
-                    character_mcut_enabled = gr.Checkbox(
-                        value=False,
-                        label="Use MCut threshold",
-                        scale=1,
-                    )
-                with gr.Row():
-                    characters_merge_enabled = gr.Checkbox(
-                        value=True,
-                        label="Merge characters into the string output",
-                        scale=1,
-                    )
-                with gr.Row():
-                    reorganizer_model_repo = gr.Dropdown(
-                        [None] + reorganizer_list,
-                        value=None,
-                        label="Reorganizer Model",
-                        info="Use a model to create a description for you",
-                    )
-                with gr.Row():
-                    additional_tags_prepend = gr.Text(label="Prepend Additional tags (comma split)")
-                    additional_tags_append  = gr.Text(label="Append Additional tags (comma split)")
-                with gr.Row():
-                    clear = gr.ClearButton(
-                        components=[
-                            gallery,
-                            model_repo,
-                            general_thresh,
-                            general_mcut_enabled,
-                            character_thresh,
-                            character_mcut_enabled,
-                            characters_merge_enabled,
-                            reorganizer_model_repo,
-                            additional_tags_prepend,
-                            additional_tags_append,
-                        ],
-                        variant="secondary",
-                        size="lg",
-                    )
-            with gr.Column(variant="panel"):
-                download_file = gr.File(label="Download includes: All outputs* and image(s)")  # 0
-                character_res = gr.Label(label="Output (characters)")  # 1
-                sorted_general_strings = gr.Textbox(label="Output (string)*", show_label=True, show_copy_button=True)  # 2
-                final_categorized_output = gr.Textbox(label="Categorized (string)* - If it's too long, select an image to display tags correctly.", show_label=True, show_copy_button=True)  # 3
-                pe_generate_btn = gr.Button(value="ENHANCE TAGS", size="lg", variant="primary") # 4
-                enhanced_tags = gr.Textbox(label="Enhanced Tags", show_label=True, show_copy_button=True)  # 5
-                prompt_enhancer_model = gr.Radio(["Medium", "Long", "Flux"], label="Model Choice", value="Medium", info="Enhance your prompts with Medium or Long answers") # 6
-                categorized = gr.JSON(label="Categorized (tags)* - JSON")  # 7
-                rating = gr.Label(label="Rating")  # 8
-                general_res = gr.Label(label="Output (tags)")  # 9
-                unclassified = gr.JSON(label="Unclassified (tags)")  # 10
-                clear.add(
-                    [
-                        download_file,
-                        sorted_general_strings,
-                        final_categorized_output,
-                        categorized,
-                        rating,
-                        character_res,
-                        general_res,
-                        unclassified,
-                        prompt_enhancer_model,
-                        enhanced_tags,
-                    ]
-                )      
-            tag_results = gr.State({})
-            # Define the event listener to add the uploaded image to the gallery
-            image_input.change(append_gallery, inputs=[gallery, image_input], outputs=[gallery, image_input])
-            # When the upload button is clicked, add the new images to the gallery
-            upload_button.upload(extend_gallery, inputs=[gallery, upload_button], outputs=gallery)
-            # Event to update the selected image when an image is clicked in the gallery
-            selected_image = gr.Textbox(label="Selected Image", visible=False)
-            gallery.select(get_selection_from_gallery,inputs=[gallery, tag_results],outputs=[selected_image, sorted_general_strings, final_categorized_output, categorized, rating, character_res, general_res, unclassified, enhanced_tags])
-            # Event to remove a selected image from the gallery
-            remove_button.click(remove_image_from_gallery, inputs=[gallery, selected_image], outputs=gallery)
-            # Event to for the Prompt Enhancer Button
-            pe_generate_btn.click(lambda tags,model:prompt_enhancer('','',tags,model)[0],inputs=[final_categorized_output,prompt_enhancer_model],outputs=[enhanced_tags])
-        submit.click(
-            predictor.predict,
-            inputs=[
-                gallery,
-                model_repo,
-                general_thresh,
-                general_mcut_enabled,
-                character_thresh,
-                character_mcut_enabled,
-                characters_merge_enabled,
-                reorganizer_model_repo,
-                additional_tags_prepend,
-                additional_tags_append,
-                tag_results,
-            ],
-            outputs=[download_file, sorted_general_strings, final_categorized_output, categorized, rating, character_res, general_res, unclassified, tag_results,],
-        )
-        gr.Examples(
-            [["images/1girl.png", VIT_LARGE_MODEL_DSV3_REPO, 0.35, False, 0.85, False]], 
-            inputs=[
-                image_input,
-                model_repo,
-                general_thresh,
-                general_mcut_enabled,
-                character_thresh,
-                character_mcut_enabled,
-            ],
-        )
-    with gr.Tab(label="Tag Categorizer + Enhancer"): 
-       with gr.Row():
-            with gr.Column(variant="panel"):
-                input_tags = gr.Textbox(label="Input Tags (Danbooru comma-separated) - (Experimental)", placeholder="1girl, cat, horns, blue hair, ...")
-                submit_button = gr.Button(value="Submit", variant="primary", size="lg")
-            with gr.Column(variant="panel"):
-                categorized_string = gr.Textbox(label="Categorized (string)", show_label=True, show_copy_button=True, lines=8)
-                categorized_json = gr.JSON(label="Categorized (tags) - JSON")
-                submit_button.click(process_tags, inputs=[input_tags], outputs=[categorized_string, categorized_json])
-            with gr.Column(variant="panel"):
-                pe_generate_btn = gr.Button(value="ENHANCE TAGS", size="lg", variant="primary")
-                enhanced_tags = gr.Textbox(label="Enhanced Tags", show_label=True, show_copy_button=True)
-                prompt_enhancer_model = gr.Radio(["Medium", "Long", "Flux"], label="Model Choice", value="Medium", info="Enhance your prompts with Medium or Long answers")
-                pe_generate_btn.click(lambda tags,model:prompt_enhancer('','',tags,model)[0],inputs=[categorized_string,prompt_enhancer_model],outputs=[enhanced_tags])
-    with gr.Tab(label="Florence 2 Image Captioning"):
-        with gr.Row():
-            with gr.Column(variant="panel"):
-                input_img = gr.Image(label="Input Picture")
-                task_type = gr.Radio(choices=['Single task', 'Cascaded task'], label='Task type selector', value='Single task')
-                task_prompt = gr.Dropdown(choices=single_task_list, label="Task Prompt", value="Caption")
-                task_type.change(fn=update_task_dropdown, inputs=task_type, outputs=task_prompt)
-                text_input = gr.Textbox(label="Text Input (optional)")
-                submit_btn = gr.Button(value="Submit")
-            with gr.Column(variant="panel"):
-                output_text = gr.Textbox(label="Output Text", show_label=True, show_copy_button=True, lines=8)
-                output_img = gr.Image(label="Output Image")
-        gr.Examples(
-            examples=[
-                ["images/image1.png", 'Object Detection'],
-                ["images/image2.png", 'OCR with Region']
-            ],
-            inputs=[input_img, task_prompt],
-            outputs=[output_text, output_img],
-            fn=process_image,
-            cache_examples=False,
-            label='Try examples'
-        )
-        submit_btn.click(process_image, [input_img, task_prompt, text_input], [output_text, output_img])
+import os
+import io,copy,requests,spaces,gradio as gr,numpy as np
+from PIL import Image, ImageOps
+import argparse,huggingface_hub,onnxruntime as rt,pandas as pd,traceback,tempfile,zipfile,re,ast,time
+from datetime import datetime,timezone
+from collections import defaultdict
+from apscheduler.schedulers.background import BackgroundScheduler
+import json
+from modules.classifyTags import classify_tags,process_tags
+from modules.florence2 import process_image,single_task_list,update_task_dropdown
+from modules.reorganizer_model import reorganizer_list,reorganizer_class
+from modules.tag_enhancer import prompt_enhancer
+from modules.booru import gelbooru_gradio,fetch_gelbooru_images,on_select
+
+os.environ['PYTORCH_ENABLE_MPS_FALLBACK']='1'
+
+TITLE = "Multi-Tagger v1.2"
+DESCRIPTION = """
+Multi-Tagger is a versatile application for advanced image analysis and captioning. Perfect for AI artists or enthusiasts, it offers a range of features:
+
+- **Automatic Tag Categorization**: Tags are grouped into categories.
+- **Tag Enhancement**: Boost your prompts with enhanced descriptions using a built-in prompt enhancer.
+- **Reorganizer**: Use a reorganizer model to format tags into a natural-language description.
+- **Batch Support**: Upload and process multiple images simultaneously.
+- **Downloadable Output**: Get almost all results as downloadable `.txt`, `.json`, and `.png` files in a `.zip` archive.
+- **Image Fetcher**: Search for images from **Gelbooru** using flexible tag filters.
+- CUDA or CPU support.
+
+Example image by [me](https://huggingface.co/Werli).
+"""
+
+# Dataset v3 series of models:
+SWINV2_MODEL_DSV3_REPO = "SmilingWolf/wd-swinv2-tagger-v3"
+CONV_MODEL_DSV3_REPO = "SmilingWolf/wd-convnext-tagger-v3"
+VIT_MODEL_DSV3_REPO = "SmilingWolf/wd-vit-tagger-v3"
+VIT_LARGE_MODEL_DSV3_REPO = "SmilingWolf/wd-vit-large-tagger-v3"
+EVA02_LARGE_MODEL_DSV3_REPO = "SmilingWolf/wd-eva02-large-tagger-v3"
+# Dataset v2 series of models:
+MOAT_MODEL_DSV2_REPO = "SmilingWolf/wd-v1-4-moat-tagger-v2"
+SWIN_MODEL_DSV2_REPO = "SmilingWolf/wd-v1-4-swinv2-tagger-v2"
+CONV_MODEL_DSV2_REPO = "SmilingWolf/wd-v1-4-convnext-tagger-v2"
+CONV2_MODEL_DSV2_REPO = "SmilingWolf/wd-v1-4-convnextv2-tagger-v2"
+VIT_MODEL_DSV2_REPO = "SmilingWolf/wd-v1-4-vit-tagger-v2"
+# IdolSankaku series of models:
+EVA02_LARGE_MODEL_IS_DSV1_REPO = "deepghs/idolsankaku-eva02-large-tagger-v1"
+SWINV2_MODEL_IS_DSV1_REPO = "deepghs/idolsankaku-swinv2-tagger-v1"
+# Files to download from the repos
+MODEL_FILENAME = "model.onnx"
+LABEL_FILENAME = "selected_tags.csv"
+
+kaomojis=['0_0','(o)_(o)','+_+','+_-','._.','<o>_<o>','<|>_<|>','=_=','>_<','3_3','6_9','>_o','@_@','^_^','o_o','u_u','x_x','|_|','||_||']
+def parse_args()->argparse.Namespace:parser=argparse.ArgumentParser();parser.add_argument('--score-slider-step',type=float,default=.05);parser.add_argument('--score-general-threshold',type=float,default=.35);parser.add_argument('--score-character-threshold',type=float,default=.85);parser.add_argument('--share',action='store_true');return parser.parse_args()
+def load_labels(dataframe)->list[str]:name_series=dataframe['name'];name_series=name_series.map(lambda x:x.replace('_',' ')if x not in kaomojis else x);tag_names=name_series.tolist();rating_indexes=list(np.where(dataframe['category']==9)[0]);general_indexes=list(np.where(dataframe['category']==0)[0]);character_indexes=list(np.where(dataframe['category']==4)[0]);return tag_names,rating_indexes,general_indexes,character_indexes
+def mcut_threshold(probs):sorted_probs=probs[probs.argsort()[::-1]];difs=sorted_probs[:-1]-sorted_probs[1:];t=difs.argmax();thresh=(sorted_probs[t]+sorted_probs[t+1])/2;return thresh
+
+class Timer:
+	def __init__(self):self.start_time=time.perf_counter();self.checkpoints=[('Start',self.start_time)]
+	def checkpoint(self,label='Checkpoint'):now=time.perf_counter();self.checkpoints.append((label,now))
+	def report(self,is_clear_checkpoints=True):
+		max_label_length=max(len(label)for(label,_)in self.checkpoints);prev_time=self.checkpoints[0][1]
+		for(label,curr_time)in self.checkpoints[1:]:elapsed=curr_time-prev_time;print(f"{label.ljust(max_label_length)}: {elapsed:.3f} seconds");prev_time=curr_time
+		if is_clear_checkpoints:self.checkpoints.clear();self.checkpoint()
+	def report_all(self):
+		print('\n> Execution Time Report:');max_label_length=max(len(label)for(label,_)in self.checkpoints)if len(self.checkpoints)>0 else 0;prev_time=self.start_time
+		for(label,curr_time)in self.checkpoints[1:]:elapsed=curr_time-prev_time;print(f"{label.ljust(max_label_length)}: {elapsed:.3f} seconds");prev_time=curr_time
+		total_time=self.checkpoints[-1][1]-self.start_time;print(f"{'Total Execution Time'.ljust(max_label_length)}: {total_time:.3f} seconds\n");self.checkpoints.clear()
+	def restart(self):self.start_time=time.perf_counter();self.checkpoints=[('Start',self.start_time)]
+class Predictor:
+    def __init__(self):
+        self.model_target_size = None
+        self.last_loaded_repo = None
+    def download_model(self, model_repo):
+        csv_path = huggingface_hub.hf_hub_download(
+            model_repo,
+            LABEL_FILENAME,
+        )
+        model_path = huggingface_hub.hf_hub_download(
+            model_repo,
+            MODEL_FILENAME,
+        )
+        return csv_path, model_path
+    def load_model(self, model_repo):
+        if model_repo == self.last_loaded_repo:
+            return
+
+        csv_path, model_path = self.download_model(model_repo)
+
+        tags_df = pd.read_csv(csv_path)
+        sep_tags = load_labels(tags_df)
+
+        self.tag_names = sep_tags[0]
+        self.rating_indexes = sep_tags[1]
+        self.general_indexes = sep_tags[2]
+        self.character_indexes = sep_tags[3]
+
+        model = rt.InferenceSession(model_path)
+        _, height, width, _ = model.get_inputs()[0].shape
+        self.model_target_size = height
+
+        self.last_loaded_repo = model_repo
+        self.model = model
+    def prepare_image(self, path):
+        image = Image.open(path)
+        image = image.convert("RGBA")
+        target_size = self.model_target_size
+
+        canvas = Image.new("RGBA", image.size, (255, 255, 255))
+        canvas.alpha_composite(image)
+        image = canvas.convert("RGB")
+
+        # Pad image to square
+        image_shape = image.size
+        max_dim = max(image_shape)
+        pad_left = (max_dim - image_shape[0]) // 2
+        pad_top = (max_dim - image_shape[1]) // 2
+
+        padded_image = Image.new("RGB", (max_dim, max_dim), (255, 255, 255))
+        padded_image.paste(image, (pad_left, pad_top))
+
+        # Resize
+        if max_dim != target_size:
+            padded_image = padded_image.resize(
+                (target_size, target_size),
+                Image.BICUBIC,
+            )
+        # Convert to numpy array
+        image_array = np.asarray(padded_image, dtype=np.float32)
+        # Convert PIL-native RGB to BGR
+        image_array = image_array[:, :, ::-1]
+        return np.expand_dims(image_array, axis=0)
+
+    def create_file(self, content: str, directory: str, fileName: str) -> str:
+        # Write the content to a file
+        file_path = os.path.join(directory, fileName)
+        if fileName.endswith('.json'):
+            with open(file_path, 'w', encoding="utf-8") as file:
+                file.write(content)
+        else:
+            with open(file_path, 'w+', encoding="utf-8") as file:
+                file.write(content)
+
+        return file_path
+
+    def predict(
+        self,
+        gallery,
+        model_repo,
+        general_thresh,
+        general_mcut_enabled,
+        character_thresh,
+        character_mcut_enabled,
+        characters_merge_enabled,
+        reorganizer_model_repo,
+        additional_tags_prepend,
+        additional_tags_append,
+        tag_results,
+        progress=gr.Progress()
+        ):
+            # Clear tag_results before starting a new prediction
+            tag_results.clear()
+
+            gallery_len = len(gallery)
+            print(f"Predict load model: {model_repo}, gallery length: {gallery_len}")
+
+            timer = Timer()  # Create a timer
+            progressRatio = 0.5 if reorganizer_model_repo else 1
+            progressTotal = gallery_len + 1
+            current_progress = 0
+
+            self.load_model(model_repo)
+            current_progress += progressRatio/progressTotal;
+            progress(current_progress, desc="Initialize wd model finished")
+            timer.checkpoint(f"Initialize wd model")
+
+            txt_infos = []
+            output_dir = tempfile.mkdtemp()
+            if not os.path.exists(output_dir):
+                os.makedirs(output_dir)
+
+            sorted_general_strings = ""
+            # Create categorized output string
+            categorized_output_strings = []
+            rating = None
+            character_res = None
+            general_res = None
+
+            if reorganizer_model_repo:
+                print(f"Reorganizer load model {reorganizer_model_repo}")
+                reorganizer = reorganizer_class(reorganizer_model_repo, loadModel=True)
+                current_progress += progressRatio/progressTotal;
+                progress(current_progress, desc="Initialize reoganizer model finished")
+                timer.checkpoint(f"Initialize reoganizer model")
+                
+            timer.report()
+
+            prepend_list = [tag.strip() for tag in additional_tags_prepend.split(",") if tag.strip()]
+            append_list = [tag.strip() for tag in additional_tags_append.split(",") if tag.strip()]
+            if prepend_list and append_list:
+                append_list = [item for item in append_list if item not in prepend_list]
+                
+            # Dictionary to track counters for each filename
+            name_counters = defaultdict(int)
+            
+            for idx, value in enumerate(gallery):
+                try:
+                    image_path = value[0]
+                    image_name = os.path.splitext(os.path.basename(image_path))[0]
+
+                    # Increment the counter for the current name
+                    name_counters[image_name] += 1
+                    
+                    if name_counters[image_name] > 1:
+                        image_name = f"{image_name}_{name_counters[image_name]:02d}"
+
+                    image = self.prepare_image(image_path)
+
+                    input_name = self.model.get_inputs()[0].name
+                    label_name = self.model.get_outputs()[0].name
+                    print(f"Gallery {idx:02d}: Starting run wd model...")
+                    preds = self.model.run([label_name], {input_name: image})[0]
+
+                    labels = list(zip(self.tag_names, preds[0].astype(float)))
+
+                    # First 4 labels are actually ratings: pick one with argmax
+                    ratings_names = [labels[i] for i in self.rating_indexes]
+                    rating = dict(ratings_names)
+
+                    # Then we have general tags: pick any where prediction confidence > threshold
+                    general_names = [labels[i] for i in self.general_indexes]
+
+                    if general_mcut_enabled:
+                        general_probs = np.array([x[1] for x in general_names])
+                        general_thresh = mcut_threshold(general_probs)
+
+                    general_res = [x for x in general_names if x[1] > general_thresh]
+                    general_res = dict(general_res)
+
+                    # Everything else is characters: pick any where prediction confidence > threshold
+                    character_names = [labels[i] for i in self.character_indexes]
+
+                    if character_mcut_enabled:
+                        character_probs = np.array([x[1] for x in character_names])
+                        character_thresh = mcut_threshold(character_probs)
+                        character_thresh = max(0.15, character_thresh)
+
+                    character_res = [x for x in character_names if x[1] > character_thresh]
+                    character_res = dict(character_res)
+                    character_list = list(character_res.keys())
+
+                    sorted_general_list = sorted(
+                        general_res.items(),
+                        key=lambda x: x[1],
+                        reverse=True,
+                    )
+                    sorted_general_list = [x[0] for x in sorted_general_list]
+                    # Remove values from character_list that already exist in sorted_general_list
+                    character_list = [item for item in character_list if item not in sorted_general_list]
+                    # Remove values from sorted_general_list that already exist in prepend_list or append_list
+                    if prepend_list:
+                        sorted_general_list = [item for item in sorted_general_list if item not in prepend_list]
+                    if append_list:
+                        sorted_general_list = [item for item in sorted_general_list if item not in append_list]
+
+                    sorted_general_list = prepend_list + sorted_general_list + append_list
+
+                    sorted_general_strings = ", ".join((character_list if characters_merge_enabled else []) + sorted_general_list).replace("(", "\(").replace(")", "\)")
+
+                    classified_tags, unclassified_tags = classify_tags(sorted_general_list)
+                    
+                    # Create a single string of ALL categorized tags for the current image
+                    categorized_output_string = ', '.join([', '.join(tags) for tags in classified_tags.values()])
+                    categorized_output_strings.append(categorized_output_string)
+                    # Collect all categorized output strings into a single string
+                    final_categorized_output = ', '.join(categorized_output_strings)
+                    
+                    # Create a .txt file for "Output (string)" and "Categorized Output (string)"
+                    txt_content = f"Output (string): {sorted_general_strings}\nCategorized Output (string): {final_categorized_output}"
+                    txt_file = self.create_file(txt_content, output_dir, f"{image_name}_output.txt")
+                    txt_infos.append({"path": txt_file, "name": f"{image_name}_output.txt"})
+
+                    # Create a .json file for "Categorized (tags)"
+                    json_content = json.dumps(classified_tags, indent=4)
+                    json_file = self.create_file(json_content, output_dir, f"{image_name}_categorized_tags.json")
+                    txt_infos.append({"path": json_file, "name": f"{image_name}_categorized_tags.json"})
+
+                    # Save a copy of the uploaded image in PNG format
+                    image_path = value[0]
+                    image = Image.open(image_path)
+                    image.save(os.path.join(output_dir, f"{image_name}.png"), format="PNG")
+                    txt_infos.append({"path": os.path.join(output_dir, f"{image_name}.png"), "name": f"{image_name}.png"})
+                    
+                    current_progress += progressRatio/progressTotal;
+                    progress(current_progress, desc=f"image{idx:02d}, predict finished")
+                    timer.checkpoint(f"image{idx:02d}, predict finished")
+                    
+                    if reorganizer_model_repo:
+                        print(f"Starting reorganizer...")
+                        reorganize_strings = reorganizer.reorganize(sorted_general_strings)
+                        reorganize_strings = re.sub(r" *Title: *", "", reorganize_strings)
+                        reorganize_strings = re.sub(r"\n+", ",", reorganize_strings)
+                        reorganize_strings = re.sub(r",,+", ",", reorganize_strings)
+                        sorted_general_strings += ",\n\n" + reorganize_strings
+
+                        current_progress += progressRatio/progressTotal;
+                        progress(current_progress, desc=f"image{idx:02d}, reorganizer finished")
+                        timer.checkpoint(f"image{idx:02d}, reorganizer finished")
+
+                    txt_file = self.create_file(sorted_general_strings, output_dir, image_name + ".txt")
+                    txt_infos.append({"path":txt_file, "name": image_name + ".txt"})
+                    
+                    # Store the result in tag_results using image_path as the key
+                    tag_results[image_path] = { 
+                        "strings": sorted_general_strings, 
+                        "strings2": categorized_output_string,  # Store the categorized output string here
+                        "classified_tags": classified_tags, 
+                        "rating": rating, 
+                        "character_res": character_res, 
+                        "general_res": general_res, 
+                        "unclassified_tags": unclassified_tags,
+                        "enhanced_tags": ""  # Initialize as empty string
+                    }
+
+                    timer.report()
+                except Exception as e:
+                    print(traceback.format_exc())
+                    print("Error predict: " + str(e))
+            # Zip creation logic:
+            download = []
+            if txt_infos is not None and len(txt_infos) > 0:
+                downloadZipPath = os.path.join(output_dir, "Multi-tagger-" + datetime.now().strftime("%Y%m%d-%H%M%S") + ".zip")
+                with zipfile.ZipFile(downloadZipPath, 'w', zipfile.ZIP_DEFLATED) as taggers_zip:
+                    for info in txt_infos:
+                        # Get file name from lookup
+                        taggers_zip.write(info["path"], arcname=info["name"])
+                download.append(downloadZipPath)
+            # End zip creation logic
+            if reorganizer_model_repo:
+                reorganizer.release_vram()
+                del reorganizer
+                 
+            progress(1, desc=f"Predict completed")
+            timer.report_all()  # Print all recorded times
+            print("Predict is complete.")
+            
+            return download, sorted_general_strings, final_categorized_output, classified_tags, rating, character_res, general_res, unclassified_tags, tag_results
+def get_selection_from_gallery(gallery: list, tag_results: dict, selected_state: gr.SelectData):
+    if not selected_state:
+        return selected_state
+    tag_result = { 
+        "strings": "", 
+        "strings2": "", 
+        "classified_tags": "{}", 
+        "rating": "", 
+        "character_res": "", 
+        "general_res": "", 
+        "unclassified_tags": "{}", 
+        "enhanced_tags": "" 
+    }
+    if selected_state.value["image"]["path"] in tag_results:
+        tag_result = tag_results[selected_state.value["image"]["path"]]
+    return (selected_state.value["image"]["path"], selected_state.value["caption"]), tag_result["strings"], tag_result["strings2"], tag_result["classified_tags"], tag_result["rating"], tag_result["character_res"], tag_result["general_res"], tag_result["unclassified_tags"], tag_result["enhanced_tags"]
+def append_gallery(gallery:list,image:str):
+	if gallery is None:gallery=[]
+	if not image:return gallery,None
+	gallery.append(image);return gallery,None
+def extend_gallery(gallery:list,images):
+	if gallery is None:gallery=[]
+	if not images:return gallery
+	gallery.extend(images);return gallery
+def remove_image_from_gallery(gallery:list,selected_image:str):
+	if not gallery or not selected_image:return gallery
+	selected_image=ast.literal_eval(selected_image)
+	if selected_image in gallery:gallery.remove(selected_image)
+	return gallery
+args = parse_args()
+predictor = Predictor()
+dropdown_list = [
+    EVA02_LARGE_MODEL_DSV3_REPO,
+    SWINV2_MODEL_DSV3_REPO,
+    CONV_MODEL_DSV3_REPO,
+    VIT_MODEL_DSV3_REPO,
+    VIT_LARGE_MODEL_DSV3_REPO,
+    # ---
+    MOAT_MODEL_DSV2_REPO,
+    SWIN_MODEL_DSV2_REPO,
+    CONV_MODEL_DSV2_REPO,
+    CONV2_MODEL_DSV2_REPO,
+    VIT_MODEL_DSV2_REPO,
+    # ---
+    SWINV2_MODEL_IS_DSV1_REPO,
+    EVA02_LARGE_MODEL_IS_DSV1_REPO,
+]
+
+def _restart_space():
+	HF_TOKEN=os.getenv('HF_TOKEN')
+	if not HF_TOKEN:raise ValueError('HF_TOKEN environment variable is not set.')
+	huggingface_hub.HfApi().restart_space(repo_id='Werli/Multi-Tagger',token=HF_TOKEN,factory_reboot=False)
+scheduler=BackgroundScheduler()
+# Add a job to restart the space every 2 days (172800 seconds)
+restart_space_job = scheduler.add_job(_restart_space, "interval", seconds=172800)
+scheduler.start()
+next_run_time_utc=restart_space_job.next_run_time.astimezone(timezone.utc)
+NEXT_RESTART=f"Next Restart: {next_run_time_utc.strftime('%Y-%m-%d %H:%M:%S')} (UTC) - The space will restart every 2 days to ensure stability and performance. It uses a background scheduler to handle the restart process."
+
+css = """
+#output {height: 500px; overflow: auto; border: 1px solid #ccc;}
+label.float.svelte-i3tvor {position: relative !important;}
+.reduced-height.svelte-11chud3 {height: calc(80% - var(--size-10));}
+"""
+
+with gr.Blocks(title=TITLE, css=css, theme="Werli/Multi-Tagger", fill_width=True) as demo:
+    gr.Markdown(value=f"<h1 style='text-align: center; margin-bottom: 1rem'>{TITLE}</h1>")
+    gr.Markdown(value=DESCRIPTION)
+    gr.Markdown(NEXT_RESTART)
+    with gr.Tab(label="Waifu Diffusion"):
+        with gr.Row():
+            with gr.Column():
+                submit = gr.Button(value="Submit", variant="primary", size="lg")
+                with gr.Column(variant="panel"):
+                    # Create an Image component for uploading images
+                    image_input = gr.Image(label="Upload an Image or clicking paste from clipboard button", type="filepath", sources=["upload", "clipboard"], height=150)
+                    with gr.Row():
+                        upload_button = gr.UploadButton("Upload multiple images", file_types=["image"], file_count="multiple", size="sm")
+                        remove_button = gr.Button("Remove Selected Image", size="sm")
+                    gallery = gr.Gallery(columns=5, rows=5, show_share_button=False, interactive=True, height="500px", label="Grid of images")
+                model_repo = gr.Dropdown(
+                    dropdown_list,
+                    value=EVA02_LARGE_MODEL_DSV3_REPO,
+                    label="Model",
+                )
+                with gr.Row():
+                    general_thresh = gr.Slider(
+                        0,
+                        1,
+                        step=args.score_slider_step,
+                        value=args.score_general_threshold,
+                        label="General Tags Threshold",
+                        scale=3,
+                    )
+                    general_mcut_enabled = gr.Checkbox(
+                        value=False,
+                        label="Use MCut threshold",
+                        scale=1,
+                    )
+                with gr.Row():
+                    character_thresh = gr.Slider(
+                        0,
+                        1,
+                        step=args.score_slider_step,
+                        value=args.score_character_threshold,
+                        label="Character Tags Threshold",
+                        scale=3,
+                    )
+                    character_mcut_enabled = gr.Checkbox(
+                        value=False,
+                        label="Use MCut threshold",
+                        scale=1,
+                    )
+                with gr.Row():
+                    characters_merge_enabled = gr.Checkbox(
+                        value=True,
+                        label="Merge characters into the string output",
+                        scale=1,
+                    )
+                with gr.Row():
+                    reorganizer_model_repo = gr.Dropdown(
+                        [None] + reorganizer_list,
+                        value=None,
+                        label="Reorganizer Model",
+                        info="Use a model to create a description for you",
+                    )
+                with gr.Row():
+                    additional_tags_prepend = gr.Text(label="Prepend Additional tags (comma split)")
+                    additional_tags_append  = gr.Text(label="Append Additional tags (comma split)")
+                with gr.Row():
+                    clear = gr.ClearButton(
+                        components=[
+                            gallery,
+                            model_repo,
+                            general_thresh,
+                            general_mcut_enabled,
+                            character_thresh,
+                            character_mcut_enabled,
+                            characters_merge_enabled,
+                            reorganizer_model_repo,
+                            additional_tags_prepend,
+                            additional_tags_append,
+                        ],
+                        variant="secondary",
+                        size="lg",
+                    )
+            with gr.Column(variant="panel"):
+                download_file = gr.File(label="Download includes: All outputs* and image(s)")  # 0
+                character_res = gr.Label(label="Output (characters)")  # 1
+                sorted_general_strings = gr.Textbox(label="Output (string)*", show_label=True, show_copy_button=True)  # 2
+                final_categorized_output = gr.Textbox(label="Categorized (string)* - If it's too long, select an image to display tags correctly.", show_label=True, show_copy_button=True)  # 3
+                pe_generate_btn = gr.Button(value="ENHANCE TAGS", size="lg", variant="primary") # 4
+                enhanced_tags = gr.Textbox(label="Enhanced Tags", show_label=True, show_copy_button=True)  # 5
+                prompt_enhancer_model = gr.Radio(["Medium", "Long", "Flux"], label="Model Choice", value="Medium", info="Enhance your prompts with Medium or Long answers") # 6
+                categorized = gr.JSON(label="Categorized (tags)* - JSON")  # 7
+                rating = gr.Label(label="Rating")  # 8
+                general_res = gr.Label(label="Output (tags)")  # 9
+                unclassified = gr.JSON(label="Unclassified (tags)")  # 10
+                clear.add(
+                    [
+                        download_file,
+                        sorted_general_strings,
+                        final_categorized_output,
+                        categorized,
+                        rating,
+                        character_res,
+                        general_res,
+                        unclassified,
+                        prompt_enhancer_model,
+                        enhanced_tags,
+                    ]
+                )      
+            tag_results = gr.State({})
+            # Define the event listener to add the uploaded image to the gallery
+            image_input.change(append_gallery, inputs=[gallery, image_input], outputs=[gallery, image_input])
+            # When the upload button is clicked, add the new images to the gallery
+            upload_button.upload(extend_gallery, inputs=[gallery, upload_button], outputs=gallery)
+            # Event to update the selected image when an image is clicked in the gallery
+            selected_image = gr.Textbox(label="Selected Image", visible=False)
+            gallery.select(get_selection_from_gallery,inputs=[gallery, tag_results],outputs=[selected_image, sorted_general_strings, final_categorized_output, categorized, rating, character_res, general_res, unclassified, enhanced_tags])
+            # Event to remove a selected image from the gallery
+            remove_button.click(remove_image_from_gallery, inputs=[gallery, selected_image], outputs=gallery)
+            # Event to for the Prompt Enhancer Button
+            pe_generate_btn.click(lambda tags,model:prompt_enhancer('','',tags,model)[0],inputs=[final_categorized_output,prompt_enhancer_model],outputs=[enhanced_tags])
+        submit.click(
+            predictor.predict,
+            inputs=[
+                gallery,
+                model_repo,
+                general_thresh,
+                general_mcut_enabled,
+                character_thresh,
+                character_mcut_enabled,
+                characters_merge_enabled,
+                reorganizer_model_repo,
+                additional_tags_prepend,
+                additional_tags_append,
+                tag_results,
+            ],
+            outputs=[download_file, sorted_general_strings, final_categorized_output, categorized, rating, character_res, general_res, unclassified, tag_results,],
+        )
+        gr.Examples(
+            [["images/1girl.png", VIT_LARGE_MODEL_DSV3_REPO, 0.35, False, 0.85, False]], 
+            inputs=[
+                image_input,
+                model_repo,
+                general_thresh,
+                general_mcut_enabled,
+                character_thresh,
+                character_mcut_enabled,
+            ],
+        )
+    with gr.Tab(label="Florence 2 Image Captioning"):
+        with gr.Row():
+            with gr.Column(variant="panel"):
+                input_img = gr.Image(label="Input Picture")
+                task_type = gr.Radio(choices=['Single task', 'Cascaded task'], label='Task type selector', value='Single task')
+                task_prompt = gr.Dropdown(choices=single_task_list, label="Task Prompt", value="Caption")
+                task_type.change(fn=update_task_dropdown, inputs=task_type, outputs=task_prompt)
+                text_input = gr.Textbox(label="Text Input (optional)")
+                submit_btn = gr.Button(value="Submit")
+            with gr.Column(variant="panel"):
+                output_text = gr.Textbox(label="Output Text", show_label=True, show_copy_button=True, lines=8)
+                output_img = gr.Image(label="Output Image")
+        gr.Examples(
+            examples=[
+                ["images/image1.png", 'Object Detection'],
+                ["images/image2.png", 'OCR with Region']
+            ],
+            inputs=[input_img, task_prompt],
+            outputs=[output_text, output_img],
+            fn=process_image,
+            cache_examples=False,
+            label='Try examples'
+        )
+        submit_btn.click(process_image, [input_img, task_prompt, text_input], [output_text, output_img])
+    with gr.Tab(label="Gelbooru Image Fetcher"):
+        with gr.Row():
+            with gr.Column():
+                gr.Markdown("### ⚙️ Search Parameters")
+                site = gr.Dropdown(label="Select Source", choices=["Gelbooru", "None (will not work)"], value="Gelbooru")
+                OR_tags = gr.Textbox(label="OR Tags (comma-separated)", placeholder="e.g. solo, 1girl, 1boy, artist, character, ...")
+                AND_tags = gr.Textbox(label="AND Tags (comma-separated)", placeholder="e.g. black hair, cat ears, black hair, granblue fantasy, ...")
+                exclude_tags = gr.Textbox(label="Exclude Tags (comma-separated)", placeholder="e.g. animated, watermark, username, ...")
+                score = gr.Number(label="Minimum Score", value=0)
+                count = gr.Slider(label="Number of Images", minimum=1, maximum=4, step=1, value=1) # Increase if necessary (not recommend)
+                Safe = gr.Checkbox(label="Include Safe", value=True)
+                Questionable = gr.Checkbox(label="Include Questionable", value=True)
+                Explicit = gr.Checkbox(label="Include Explicit", value=False)
+                #user_id = gr.Textbox(label="User ID (Optional)", value="")
+                #api_key = gr.Textbox(label="API Key (Optional)", value="", type="password")
+
+                submit_btn = gr.Button("Fetch Images", variant="primary")
+
+            with gr.Column():
+                gr.Markdown("### 📄 Results")
+                images_output = gr.Gallery(label="Images", columns=3, rows=2, object_fit="contain", height=500)
+                tags_output = gr.Textbox(label="Tags", placeholder="Select an image to show tags", lines=5, show_copy_button=True)
+                post_url_output = gr.Textbox(label="Post URL", lines=1, show_copy_button=True)
+                image_url_output = gr.Textbox(label="Image URL", lines=1, show_copy_button=True)
+
+        # State to store tags, URLs
+        tags_state = gr.State([])
+        post_url_state = gr.State([])
+        image_url_state = gr.State([])
+
+        submit_btn.click(
+            fn=gelbooru_gradio,
+            inputs=[OR_tags, AND_tags, exclude_tags, score, count, Safe, Questionable, Explicit, site], # add 'api_key' and 'user_id' if necessary
+            outputs=[images_output, tags_state, post_url_state, image_url_state],
+        )
+
+        images_output.select(
+            fn=on_select,
+            inputs=[tags_state, post_url_state, image_url_state],
+            outputs=[tags_output, post_url_output, image_url_output],
+        )
+        gr.Markdown("""
+        ---
+        ComfyUI version: [Comfyui-Gelbooru](https://github.com/1mckw/Comfyui-Gelbooru)
+        """)
+    with gr.Tab(label="Categorizer++"): 
+        with gr.Row():
+            with gr.Column(variant="panel"):
+                input_tags = gr.Textbox(label="Input Tags", placeholder="1girl, cat, horns, blue hair, ...\nor\n? 1girl 1234567? cat 1234567? horns 1234567? blue hair 1234567? ...", lines=4)
+                submit_button = gr.Button(value="Submit", variant="primary", size="lg")
+            with gr.Column(variant="panel"):
+                categorized_string = gr.Textbox(label="Categorized (string)", show_label=True, show_copy_button=True, lines=8)
+                categorized_json = gr.JSON(label="Categorized (tags) - JSON")
+                submit_button.click(process_tags, inputs=[input_tags], outputs=[categorized_string, categorized_json])
+            with gr.Column(variant="panel"):
+                pe_generate_btn = gr.Button(value="ENHANCE TAGS", size="lg", variant="primary")
+                enhanced_tags = gr.Textbox(label="Enhanced Tags", show_label=True, show_copy_button=True)
+                prompt_enhancer_model = gr.Radio(["Medium", "Long", "Flux"], label="Model Choice", value="Medium", info="Enhance your prompts with Medium or Long answers")
+                pe_generate_btn.click(lambda tags,model:prompt_enhancer('','',tags,model)[0],inputs=[categorized_string,prompt_enhancer_model],outputs=[enhanced_tags])
 demo.queue(max_size=2).launch()

modules/booru.py

@@ -0,0 +1,137 @@
+import requests
+import re
+import base64
+import io
+import numpy as np
+from PIL import Image, ImageOps
+import torch
+import gradio as gr
+
+# Helper to load image from URL
+def loadImageFromUrl(url):
+    if url.startswith("data:image/"):
+        i = Image.open(io.BytesIO(base64.b64decode(url.split(",")[1])))
+    elif url.startswith("s3://"):
+        raise Exception("S3 URLs not supported in this interface")
+    else:
+        response = requests.get(url, timeout=5)
+        if response.status_code != 200:
+            raise Exception(response.text)
+        i = Image.open(io.BytesIO(response.content))
+
+    i = ImageOps.exif_transpose(i)
+    if i.mode != "RGBA":
+        i = i.convert("RGBA")
+
+    alpha = i.split()[-1]
+    image = Image.new("RGB", i.size, (0, 0, 0))
+    image.paste(i, mask=alpha)
+
+    image = np.array(image).astype(np.float32) / 255.0
+    image = torch.from_numpy(image)[None,]
+    return image
+
+# Fetch data from Gelbooru or None
+def fetch_gelbooru_images(site, OR_tags, AND_tags, exclude_tag, score, count, Safe, Questionable, Explicit): # add 'api_key' and 'user_id' if necessary
+    # AND_tags
+    AND_tags = AND_tags.rstrip(',').rstrip(' ')
+    AND_tags = AND_tags.split(',')
+    AND_tags = [item.strip().replace(' ', '_').replace('\\', '') for item in AND_tags]
+    AND_tags = [item for item in AND_tags if item]
+    if len(AND_tags) > 1:
+        AND_tags = '+'.join(AND_tags)
+    else:
+        AND_tags = AND_tags[0] if AND_tags else ''
+
+    # OR_tags
+    OR_tags = OR_tags.rstrip(',').rstrip(' ')
+    OR_tags = OR_tags.split(',')
+    OR_tags = [item.strip().replace(' ', '_').replace('\\', '') for item in OR_tags]
+    OR_tags = [item for item in OR_tags if item]
+    if len(OR_tags) > 1:
+        OR_tags = '{' + ' ~ '.join(OR_tags) + '}'
+    else:
+        OR_tags = OR_tags[0] if OR_tags else ''
+
+    # Exclude tags
+    exclude_tag = '+'.join('-' + item.strip().replace(' ', '_') for item in exclude_tag.split(','))
+
+    rate_exclusion = ""
+    if not Safe:
+        if site == "None":
+            rate_exclusion += "+-rating%3asafe"
+        else:
+            rate_exclusion += "+-rating%3ageneral"
+    if not Questionable:
+        if site == "None":
+            rate_exclusion += "+-rating%3aquestionable"
+        else:
+            rate_exclusion += "+-rating%3aquestionable+-rating%3aSensitive"
+    if not Explicit:
+        if site == "None":
+            rate_exclusion += "+-rating%3aexplicit"
+        else:
+            rate_exclusion += "+-rating%3aexplicit"
+
+    if site == "None":
+        base_url = "https://api.example.com/index.php"
+    else:
+        base_url = "https://gelbooru.com/index.php"
+
+    query_params = (
+        f"page=dapi&s=post&q=index&tags=sort%3arandom+"
+        f"{exclude_tag}+{OR_tags}+{AND_tags}+{rate_exclusion}"
+        f"+score%3a>{score}&limit={count}&json=1"
+        #f"+score%3a>{score}&api_key={api_key}&user_id={user_id}&limit={count}&json=1"
+    )
+    url = f"{base_url}?{query_params}".replace("-+", "")
+    url = re.sub(r"\++", "+", url)
+
+    response = requests.get(url, verify=True)
+    if site == "None":
+        posts = response.json()
+    else:
+        posts = response.json().get('post', [])
+
+    image_urls = [post.get("file_url", "") for post in posts]
+    tags_list = [post.get("tags", "").replace(" ", ", ").replace("_", " ").replace("(", "\\(").replace(")", "\\)").strip() for post in posts]
+    #tags_list = [post.get("tags", "").replace("_", " ").replace(" ", ", ").strip() for post in posts]
+    ids_list = [str(post.get("id", "")) for post in posts]
+
+    if site == "Gelbooru":
+        post_urls = [f"https://gelbooru.com/index.php?page=post&s=view&id={id}" for id in ids_list]
+    #else:
+    #   post_urls = [f"https://api.none.com/index.php?page=post&s=view&id={id}" for id in ids_list]
+
+    return image_urls, tags_list, post_urls
+
+# Main function to fetch and return processed images
+def gelbooru_gradio(
+    OR_tags, AND_tags, exclude_tags, score, count, Safe, Questionable, Explicit, site  # add 'api_key' and 'user_id' if necessary
+):
+    image_urls, tags_list, post_urls = fetch_gelbooru_images(
+        site, OR_tags, AND_tags, exclude_tags, score, count, Safe, Questionable, Explicit # 'api_key' and 'user_id' if necessary
+    )
+
+    if not image_urls:
+        return [], [], [], []
+
+    image_data = []
+    for url in image_urls:
+        try:
+            image = loadImageFromUrl(url)
+            image = (image * 255).clamp(0, 255).cpu().numpy().astype(np.uint8)[0]
+            image = Image.fromarray(image)
+            image_data.append(image)
+        except Exception as e:
+            print(f"Error loading image from {url}: {e}")
+            continue
+
+    return image_data, tags_list, post_urls, image_urls
+
+# Update UI on image click
+def on_select(evt: gr.SelectData, tags_list, post_url_list, image_url_list):
+    idx = evt.index
+    if idx < len(tags_list):
+        return tags_list[idx], post_url_list[idx], image_url_list[idx]
+    return "No tags", "", ""

modules/classifyTags.py

@@ -1,174 +1,191 @@
-from collections import defaultdict
-
-# Define grouping rules (categories and keywords)
-# Provided categories and reversed_categories
-categories = {
-"Explicit"            : ["sex", "69", "paizuri", "cum", "precum", "areola_slip", "hetero", "erection", "oral", "fellatio", "yaoi", "ejaculation", "ejaculating", "masturbation", "handjob", "bulge", "rape", "_rape", "doggystyle", "threesome", "missionary", "object_insertion", "nipple", "nipples", "pussy", "anus", "penis", "groin", "testicles", "testicle", "anal", "cameltoe", "areolae", "dildo", "clitoris", "top-down_bottom-up", "gag", "groping", "gagged", "gangbang", "orgasm", "femdom", "incest", "bukkake", "breast_out", "vaginal", "vagina", "public_indecency", "breast_sucking", "folded", "cunnilingus", "_cunnilingus", "foreskin", "bestiality", "footjob", "uterus", "womb", "flaccid", "defloration", "butt_plug", "cowgirl_position", "reverse_cowgirl_position", "squatting_cowgirl_position", "reverse_upright_straddle", "irrumatio", "deepthroat", "pokephilia", "gaping", "orgy", "cleft_of_venus", "futanari", "futasub", "futa", "cumdrip", "fingering", "vibrator", "partially_visible_vulva", "penetration", "penetrated", "cumshot", "exhibitionism", "breast_milk", "grinding", "clitoral", "urethra", "phimosis", "cervix", "impregnation", "tribadism", "molestation", "pubic_hair", "clothed_female_nude_male", "clothed_male_nude_female", "clothed_female_nude_female", "clothed_male_nude_male", "sex_machine", "milking_machine", "ovum", "chikan", "pussy_juice_drip_through_clothes", "ejaculating_while_penetrated", "suspended_congress", "reverse_suspended_congress", "spread_pussy_under_clothes", "anilingus", "reach-around", "humping", "consensual_tentacles", "tentacle_pit", "cum_in_", ],
-#外観状態／外觀狀態
-"Appearance Status"   : ["backless", "bandaged_neck", "bleeding", "blood", "_blood", "blush", "body_writing", "bodypaint", "bottomless", "breath", "bruise", "butt_crack", "cold", "covered_mouth", "crack", "cross-section", "crotchless", "crying", "curvy", "cuts", "dirty", "dripping", "drunk", "from_mouth", "glowing", "hairy", "halterneck", "hot", "injury", "latex", "leather", "levitation", "lipstick_mark", "_markings", "makeup", "mole", "moles", "no_bra", "nosebleed", "nude", "outfit", "pantylines", "peeing", "piercing", "_piercing", "piercings", "pregnant", "public_nudity", "reverse", "_skin", "_submerged", "saliva", "scar", "scratches", "see-through", "shadow", "shibari", "sideless", "skindentation", "sleeping","tan", "soap_bubbles", "steam", "steaming_body", "stitches", "sweat", "sweatdrop", "sweaty", "tanlines", "tattoo", "tattoo", "tears", "topless", "transparent", "trefoil", "trembling", "veins", "visible_air", "wardrobe_malfunction", "wet", "x-ray", "unconscious", "handprint", ],
-#動作姿勢／動作姿勢
-"Action Pose"         : ["afloat", "afterimage", "against_fourth_wall", "against_wall", "aiming", "all_fours", "another's_mouth", "arm_", "arm_support", "arms_", "arms_behind_back", "asphyxiation", "attack", "back", "ballet", "bara", "bathing", "battle", "bdsm", "beckoning", "bent_over", "bite_mark", "biting", "bondage", "breast_suppress", "breathing", "burning", "bust_cup", "carry", "carrying", "caught", "chained", "cheek_squash", "chewing", "cigarette", "clapping", "closed_eye", "come_hither", "cooking", "covering", "cuddling", "dancing", "_docking", "destruction", "dorsiflexion", "dreaming", "dressing", "drinking", "driving", "dropping", "eating", "exercise", "expansion", "exposure", "facing", "failure", "fallen_down", "falling", "feeding", "fetal_position", "fighting", "finger_on_trigger", "finger_to_cheek", "finger_to_mouth", "firing", "fishing", "flashing", "fleeing", "flexible", "flexing", "floating", "flying", "fourth_wall", "freediving", "frogtie", "_grab", "girl_on_top", "giving", "grabbing", "grabbing_", "gymnastics", "_hold", "hadanugi_dousa", "hairdressing", "hand_", "hand_on", "hand_on_wall", "hands_", "headpat", "hiding", "holding", "hug", "hugging", "imagining", "in_container", "in_mouth", "in_palm", "jealous", "jumping", "kabedon", "kicking", "kiss", "kissing", "kneeling", "_lift", "lactation", "laundry", "licking", "lifted_by_self", "looking", "lowleg", "lying", "melting", "midair", "moaning", "_open", "on_back", "on_bed", "on_ground", "on_lap", "on_one_knee", "one_eye_closed", "open_", "over_mouth", "own_mouth", "_peek", "_pose", "_press", "_pull", "padding", "paint", "painting_(action)", "palms_together", "pee", "peeking", "pervert", "petting", "pigeon-toed", "piggyback", "pinching", "pinky_out", "pinned", "plantar_flexion", "planted", "playing", "pocky", "pointing", "poke", "poking", "pouring", "pov", "praying", "presenting", "profanity", "pulled_by_self", "pulling", "pump_action", "punching", "_rest", "raised", "reaching", "reading", "reclining", "reverse_grip", "riding", "running", "_slip", "salute", "screaming", "seiza", "selfie", "sewing", "shaking", "shoe_dangle", "shopping", "shouting", "showering", "shushing", "singing", "sitting", "slapping", "smell", "smelling", "smoking", "smother", "solo", "spanked", "spill", "spilling", "spinning", "splashing", "split", "squatting", "squeezed", "breasts_squeezed_together", "standing", "standing_on_", "staring", "straddling", "strangling", "stretching", "surfing", "suspension", "swimming", "talking", "teardrop", "tearing_clothes", "throwing", "tied_up", "tiptoes", "toe_scrunch", "toothbrush", "trigger_discipline", "tripping", "tsundere", "turning_head", "twitching", "two-handed", "tying", "_up",  "unbuttoned", "undressed", "undressing", "unsheathed", "unsheathing", "unzipped", "unzipping", "upright_straddle", "v", "V", "vore", "_wielding","wading", "walk-in", "walking", "wariza", "waving", "wedgie", "wrestling", "writing", "yawning", "yokozuwari", "_conscious", "massage", "struggling", "shrugging", "drugged", "tentacles_under_clothes", "restrained_by_tentacles", "tentacles_around_arms", "tentacles_around_legs", "restrained_legs", "restrained_tail", "restrained_arms", "tentacles_on_female", "archery", "cleaning", "tempura", "facepalm", "sadism", ],
-#頭部装飾／頭部服飾
-"Headwear"            : ["antennae", "antlers", "aura", "bandaged_head", "bandana", "bandeau", "beanie", "beanie", "beret", "bespectacled", "blindfold", "bonnet", "_cap", "circlet", "crown", "_drill", "_drills", "diadem", "_eyewear", "ear_covers", "ear_ornament", "ear_tag", "earbuds", "earclip", "earmuffs", "earphones", "earpiece", "earring", "earrings", "eyeliner", "eyepatch", "eyewear_on_head", "facial", "fedora", "glasses", "goggles", "_headwear", "hachimaki", "hair_bobbles", "hair_ornament", "hair_rings", "hair_tie", "hairband", "hairclip", "hairpin", "hairpods", "halo", "hat", "head-mounted_display", "head_wreath", "headband", "headdress", "headgear", "headphones", "headpiece", "headset", "helm", "helmet", "hood", "kabuto_(helmet)", "kanzashi", "_mask", "maid_headdress", "mask", "mask", "mechanical_ears", "mechanical_eye", "mechanical_horns", "mob_cap", "monocle", "neck_ruff", "nightcap", "on_head", "pince-nez", "qingdai_guanmao", "scarf_over_mouth", "scrunchie", "sunglasses", "tam_o'_shanter", "tate_eboshi", "tiara", "topknot", "turban", "veil", "visor", "wig", "mitre", "tricorne", "bicorne", ],
-#手部装飾／手部服飾
-"Handwear"            : ["arm_warmers", "armband", "armlet", "bandaged_arm", "bandaged_fingers", "bandaged_hand", "bandaged_wrist", "bangle", "bracelet", "bracelets", "bracer", "cuffs", "elbow_pads", "_gauntlets", "_glove", "_gloves", "gauntlets", "gloves", "kote", "kurokote", "mechanical_arm", "mechanical_arms", "mechanical_hands", "mittens", "mitts", "nail_polish", "prosthetic_arm", "wrist_cuffs", "wrist_guards", "wristband", "yugake", ],
-#ワンピース衣装／一件式服裝
-"One-Piece Outfit"    : ["bodystocking", "bodysuit", "dress", "furisode", "gown", "hanfu", "jumpsuit", "kimono", "leotard", "microdress", "one-piece", "overalls", "robe", "spacesuit", "sundress", "yukata", ],
-#上半身衣装／上半身服裝
-"Upper Body Clothing" : ["aiguillette", "apron", "_apron", "armor", "_armor", "ascot", "babydoll", "bikini", "_bikini", "blazer", "_blazer", "blouse", "_blouse", "bowtie", "_bowtie", "bra", "_bra", "breast_curtain", "breast_curtains", "breast_pocket", "breastplate", "bustier", "camisole", "cape", "capelet", "cardigan", "center_opening", "chemise", "chest_jewel", "choker", "cloak", "coat", "coattails", "collar", "_collar", "corset", "criss-cross_halter", "crop_top", "dougi", "feather_boa", "gakuran", "hagoromo", "hanten_(clothes)", "haori", "harem_pants", "harness", "hoodie", "jacket", "_jacket", "japanese_clothes", "kappougi", "kariginu", "lapels", "lingerie", "_lingerie", "maid", "mechanical_wings", "mizu_happi", "muneate", "neckerchief", "necktie", "negligee", "nightgown", "pajamas", "_pajamas", "pauldron", "pauldrons", "plunging_neckline", "raincoat", "rei_no_himo", "sailor_collar", "sarashi", "scarf", "serafuku", "shawl", "shirt", "shoulder_", "sleepwear", "sleeve", "sleeveless", "sleeves", "_sleeves", "sode", "spaghetti_strap", "sportswear", "strapless", "suit", "sundress", "suspenders", "sweater", "swimsuit", "_top", "_torso", "t-shirt", "tabard", "tailcoat", "tank_top", "tasuki", "tie_clip", "tunic", "turtleneck", "tuxedo", "_uniform", "undershirt", "uniform", "v-neck", "vambraces", "vest", "waistcoat", ],
-#下半身衣装／下半身服裝
-"Lower Body Clothing" : ["bare_hips", "bloomers", "briefs", "buruma", "crotch_seam", "cutoffs", "denim", "faulds", "fundoshi", "g-string", "garter_straps", "hakama", "hip_vent", "jeans", "knee_pads", "loincloth", "mechanical_tail", "microskirt", "miniskirt", "overskirt", "panties", "pants", "pantsu", "panty_straps", "pelvic_curtain", "petticoat", "sarong", "shorts", "side_slit", "skirt", "sweatpants", "swim_trunks", "thong", "underwear", "waist_cape", ],
-#足元・レッグウェア／腳與腿部服飾
-"Foot & Legwear"      : ["anklet", "bandaged_leg", "boot", "boots", "_footwear", "flats", "flip-flops", "geta", "greaves", "_heels", "kneehigh", "kneehighs", "_legwear", "leg_warmers", "leggings", "loafers", "mary_janes", "mechanical_legs", "okobo", "over-kneehighs", "pantyhose", "prosthetic_leg", "pumps", "_shoe", "_sock", "sandals", "shoes", "skates", "slippers", "sneakers", "socks", "spikes", "tabi", "tengu-geta", "thigh_strap", "thighhighs", "uwabaki", "zouri", "legband", "ankleband", ],
-#その他の装飾／其他服飾
-"Other Accessories"   : ["alternate_", "anklet", "badge", "beads", "belt", "belts", "bow", "brooch", "buckle", "button", "buttons", "_clothes", "_costume", "_cutout", "casual", "charm", "clothes_writing", "clothing_aside", "costume", "cow_print", "cross", "d-pad", "double-breasted", "drawstring", "epaulettes", "fabric", "fishnets", "floral_print", "formal", "frills", "_garter", "gem", "holster", "jewelry", "_knot", "lace", "lanyard", "leash", "magatama", "mechanical_parts", "medal", "medallion", "naked_bandage", "necklace", "_ornament", "(ornament)", "o-ring", "obi", "obiage", "obijime", "_pin", "_print", "padlock", "patterned_clothing", "pendant", "piercing", "plaid", "pocket", "polka_dot", "pom_pom_(clothes)", "pom_pom_(clothes)", "pouch", "ribbon", "_ribbon", "_stripe", "_stripes", "sash", "shackles", "shimenawa", "shrug_(clothing)", "skin_tight", "spandex", "strap", "sweatband", "_trim", "tassel", "zettai_ryouiki", "zipper", ],
-#表情／表情
-"Facial Expression"   : ["ahegao", "anger_vein", "angry", "annoyed", "confused", "drooling", "embarrassed", "expressionless", "eye_contact", "_face", "frown", "fucked_silly", "furrowed_brow", "glaring", "gloom_(expression)", "grimace", "grin", "happy", "jitome", "laughing", "_mouth", "nervous", "notice_lines", "o_o", "parted_lips", "pout", "puff_of_air", "restrained", "sad", "sanpaku", "scared", "scowl", "serious", "shaded_face", "shy", "sigh", "sleepy", "smile", "smirk", "smug", "snot", "spoken_ellipsis", "spoken_exclamation_mark", "spoken_interrobang", "spoken_question_mark", "squiggle", "surprised", "tareme", "tearing_up", "thinking", "tongue", "tongue_out", "torogao", "tsurime", "turn_pale", "wide-eyed", "wince", "worried", "heartbeat", ],
-#絵文字／表情符號
-"Facial Emoji"        : ["!!", "!", "!?", "+++", "+_+", "...", "...?", "._.", "03:00", "0_0", ":/", ":3", ":<", ":>", ":>=", ":d", ":i", ":o", ":p", ":q", ":t", ":x", ":|", ";(", ";)", ";3", ";d", ";o", ";p", ";q", "=_=", ">:(", ">:)", ">_<", ">_o", ">o<", "?", "??", "@_@", "\m/", "\n/", "\o/", "\||/", "^^^", "^_^", "c:", "d:", "o_o", "o3o", "u_u", "w", "x", "x_x", "xd", "zzz", "|_|", ],
-#頭部／頭部
-"Head"                : ["afro", "ahoge", "animal_ear_fluff", "_bangs", "_bun", "bald", "beard", "blunt_bangs", "blunt_ends", "bob_cut", "bowl_cut", "braid", "braids", "buzz_cut", "circle_cut", "colored_tips", "cowlick", "dot_nose", "dreadlocks", "_ear", "_ears", "_eye", "_eyes", "enpera", "eyeball", "eyebrow", "eyebrow_cut", "eyebrows", "eyelashes", "eyeshadow", "faceless", "facepaint", "facial_mark", "fang", "forehead", "freckles", "goatee", "_hair", "_horn", "_horns", "hair_", "hair_bun", "hair_flaps", "hair_intakes", "hair_tubes", "half_updo", "head_tilt", "heterochromia", "hime_cut", "hime_cut", "horns", "in_eye", "inverted_bob", "kemonomimi_mode", "lips", "mascara", "mohawk", "mouth_", "mustache", "nose", "one-eyed", "one_eye", "one_side_up", "_pupils", "parted_bangs", "pompadour", "ponytail", "ringlets", "_sclera", "sideburns", "sidecut", "sidelock", "sidelocks", "skull", "snout", "stubble", "swept_bangs", "tails", "teeth", "third_eye", "twintails", "two_side_up", "undercut", "updo", "v-shaped_eyebrows", "whiskers", "tentacle_hair", ],
-#手部／手部
-"Hands"               : ["_arm", "_arms", "claws", "_finger", "_fingers", "fingernails", "_hand", "_nail", "_nails", "palms", "rings", "thumbs_up", ],
-#上半身／上半身
-"Upper Body"          : ["abs", "armpit", "armpits", "backboob", "belly", "biceps", "breast_rest", "breasts", "button_gap", "cleavage", "collarbone", "dimples_of_venus", "downblouse", "flat_chest", "linea_alba", "median_furrow", "midriff", "nape", "navel", "pectorals", "ribs", "_shoulder", "_shoulders", "shoulder_blades", "sideboob", "sidetail", "spine", "stomach", "strap_gap", "toned", "underboob", "underbust", ],
-#下半身／下半身
-"Lower Body"          : ["ankles", "ass", "barefoot", "crotch", "feet", "highleg", "hip_bones", "hooves", "kneepits", "knees", "legs", "soles", "tail", "thigh_gap", "thighlet", "thighs", "toenail", "toenails", "toes", "wide_hips", ],
-#生物／生物
-"Creature"            : ["(animal)", "anglerfish", "animal", "bear", "bee", "bird", "bug", "butterfly", "cat", "chick", "chicken", "chinese_zodiac", "clownfish", "coral", "crab", "creature", "crow", "dog", "dove", "dragon", "duck", "eagle", "fish", "fish", "fox", "fox", "frog", "frog", "goldfish", "hamster", "horse", "jellyfish", "ladybug", "lion", "mouse", "octopus", "owl", "panda", "penguin", "pig", "pigeon", "rabbit", "rooster", "seagull", "shark", "sheep", "shrimp", "snail", "snake", "squid", "starfish", "tanuki", "tentacles", "goo_tentacles", "plant_tentacles", "crotch_tentacles", "mechanical_tentacles", "squidward_tentacles", "suction_tentacles", "penis_tentacles", "translucent_tentacles", "back_tentacles", "red_tentacles", "green_tentacles", "blue_tentacles", "black_tentacles", "pink_tentacles", "purple_tentacles", "face_tentacles", "tentacles_everywhere", "milking_tentacles", "tiger", "turtle", "weasel", "whale", "wolf", "parrot", "sparrow", "unicorn", ],
-#植物／植物
-"Plant"               : ["bamboo", "bouquet", "branch", "bush", "cherry_blossoms", "clover", "daisy", "(flower)", "flower", "flower", "gourd", "hibiscus", "holly", "hydrangea", "leaf", "lily_pad", "lotus", "moss", "palm_leaf", "palm_tree", "petals", "plant", "plum_blossoms", "rose", "spider_lily", "sunflower", "thorns", "tree", "tulip", "vines", "wisteria", "acorn", ],
-#食べ物／食物
-"Food"                : ["apple", "baguette", "banana", "baozi", "beans", "bento", "berry", "blueberry", "bread", "broccoli", "burger", "cabbage", "cake", "candy", "carrot", "cheese", "cherry", "chili_pepper", "chocolate", "coconut", "cookie", "corn", "cream", "crepe", "cucumber", "cucumber", "cupcake", "curry", "dango", "dessert", "doughnut", "egg", "eggplant", "_(food)", "_(fruit)", "food", "french_fries", "fruit", "grapes", "ice_cream", "icing", "lemon", "lettuce", "lollipop", "macaron", "mandarin_orange", "meat", "melon", "mochi", "mushroom", "noodles", "omelet", "omurice", "onigiri", "onion", "pancake", "parfait", "pasties", "pastry", "peach", "pineapple", "pizza", "popsicle", "potato", "pudding", "pumpkin", "radish", "ramen", "raspberry", "rice", "roasted_sweet_potato", "sandwich", "sausage", "seaweed", "skewer", "spitroast", "spring_onion", "strawberry", "sushi", "sweet_potato", "sweets", "taiyaki", "takoyaki", "tamagoyaki", "tempurakanbea", "toast", "tomato", "vegetable", "wagashi", "wagashi", "watermelon", "jam", "popcorn", ],
-#飲み物／飲品
-"Beverage"            : ["alcohol", "beer", "coffee", "cola", "drink", "juice", "juice_box", "milk", "sake", "soda", "tea", "_tea", "whiskey", "wine", "cocktail", ],
-#音楽／音樂
-"Music"               : ["band", "baton_(conducting)", "beamed", "cello", "concert", "drum", "drumsticks", "eighth_note", "flute", "guitar", "harp", "horn", "(instrument)", "idol", "instrument", "k-pop", "lyre", "(music)", "megaphone", "microphone", "music", "musical_note", "phonograph", "piano", "plectrum", "quarter_note", "recorder", "sixteenth_note", "sound_effects", "trumpet", "utaite", "violin", "whistle", ],
-#武器・装備／武器・裝備
-"Weapons & Equipment" : ["ammunition", "arrow_(projectile)", "axe", "bandolier", "baseball_bat", "beretta_92", "bolt_action", "bomb", "bullet", "bullpup", "cannon", "chainsaw", "crossbow", "dagger", "energy_sword", "explosive", "fighter_jet", "gohei", "grenade", "gun", "hammer", "handgun", "holstered", "jet", "katana", "knife", "kunai", "lance", "mallet", "nata_(tool)", "polearm", "quiver", "rapier", "revolver", "rifle", "rocket_launcher", "scabbard", "scope", "scythe", "sheath", "sheathed", "shield", "shotgun", "shuriken", "spear", "staff", "suppressor", "sword", "tank", "tantou", "torpedo", "trident", "(weapon)", "wand", "weapon", "whip", "yumi_(bow)", "h&k_hk416", "rocket_launcher", "heckler_&_koch", "_weapon", ],
-#乗り物／交通器具
-"Vehicles"            : ["aircraft", "airplane", "bicycle", "boat", "car", "caterpillar_tracks", "flight_deck", "helicopter", "motor_vehicle", "motorcycle", "ship", "spacecraft", "spoiler_(automobile)", "train", "truck", "watercraft", "wheel", "wheelbarrow", "wheelchair", "inflatable_raft", ],
-#建物／建物
-"Buildings"           : ["apartment", "aquarium", "architecture", "balcony", "building", "cafe", "castle", "church", "gym", "hallway", "hospital", "house", "library", "(place)", "porch", "restaurant", "restroom", "rooftop", "shop", "skyscraper", "stadium", "stage", "temple", "toilet", "tower", "train_station", "veranda", ],
-#室内／室內
-"Indoor"              : ["bath", "bathroom", "bathtub", "bed", "bed_sheet", "bedroom", "blanket", "bookshelf", "carpet", "ceiling", "chair", "chalkboard", "classroom", "counter", "cupboard", "curtains", "cushion", "dakimakura", "desk", "door", "doorway", "drawer", "_floor", "floor", "futon", "indoors", "interior", "kitchen", "kotatsu", "locker", "mirror", "pillow", "room", "rug", "school_desk", "shelf", "shouji", "sink", "sliding_doors", "stairs", "stool", "storeroom", "table", "tatami", "throne", "window", "windowsill", "bathhouse", "chest_of_drawers", ],
-#屋外／室外
-"Outdoor"             : ["alley", "arch", "beach", "bridge", "bus_stop", "bush", "cave", "(city)", "city", "cliff", "crescent", "crosswalk", "day", "desert", "fence", "ferris_wheel", "field", "forest", "grass", "graveyard", "hill", "lake", "lamppost", "moon", "mountain", "night", "ocean", "onsen", "outdoors", "path", "pool", "poolside", "railing", "railroad", "river", "road", "rock", "sand", "shore", "sky", "smokestack", "snow", "snowball", "snowman", "street", "sun", "sunlight", "sunset", "tent", "torii", "town", "tree", "turret", "utility_pole", "valley", "village", "waterfall", ],
-#物品／物品
-"Objects"             : ["anchor", "android", "armchair", "(bottle)", "backpack", "bag", "ball", "balloon", "bandages", "bandaid", "bandaids", "banknote", "banner", "barcode", "barrel", "baseball", "basket", "basketball", "beachball", "bell", "bench", "binoculars", "board_game", "bone", "book", "bottle", "bowl", "box", "box_art", "briefcase", "broom", "bucket", "(chess)", "(computer)", "(computing)", "(container)", "cage", "calligraphy_brush", "camera", "can", "candle", "candlestand", "cane", "card", "cartridge", "cellphone", "chain", "chandelier", "chess", "chess_piece", "choko_(cup)", "chopsticks", "cigar", "clipboard", "clock", "clothesline", "coin", "comb", "computer", "condom", "controller", "cosmetics", "couch", "cowbell", "crazy_straw", "cup", "cutting_board", "dice", "digital_media_player", "doll", "drawing_tablet", "drinking_straw", "easel", "electric_fan", "emblem", "envelope", "eraser", "feathers", "figure", "fire", "fishing_rod", "flag", "flask", "folding_fan", "fork", "frying_pan", "(gemstone)", "game_console", "gears", "gemstone", "gift", "glass", "glowstick", "gold", "handbag", "handcuffs", "handheld_game_console", "hose", "id_card", "innertube", "iphone", "jack-o'-lantern", "jar", "joystick", "key", "keychain", "kiseru", "ladder", "ladle", "lamp", "lantern", "laptop", "letter", "letterboxed", "lifebuoy", "lipstick", "liquid", "lock", "lotion", "_machine", "map", "marker", "model_kit", "money", "monitor", "mop", "mug", "needle", "newspaper", "nintendo", "nintendo_switch", "notebook", "(object)", "ofuda", "orb", "origami", "(playing_card)", "pack", "paddle", "paintbrush", "pan", "paper", "parasol", "patch", "pc", "pen", "pencil", "pencil", "pendant_watch", "phone", "pill", "pinwheel", "plate", "playstation", "pocket_watch", "pointer", "poke_ball", "pole", "quill", "racket", "randoseru", "remote_control", "ring", "rope", "sack", "saddle", "sakazuki", "satchel", "saucer", "scissors", "scroll", "seashell", "seatbelt", "shell", "shide", "shopping_cart", "shovel", "shower_head", "silk", "sketchbook", "smartphone", "soap", "sparkler", "spatula", "speaker", "spoon", "statue", "stethoscope", "stick", "sticker", "stopwatch", "string", "stuffed_", "stylus", "suction_cups", "suitcase", "surfboard", "syringe", "talisman", "tanzaku", "tape", "teacup", "teapot", "teddy_bear", "television", "test_tube", "tiles", "tokkuri", "tombstone", "torch", "towel", "toy", "traffic_cone", "tray", "treasure_chest", "uchiwa", "umbrella", "vase", "vial", "video_game", "viewfinder", "volleyball", "wallet", "watch", "watch", "whisk", "whiteboard", "wreath", "wrench", "wristwatch", "yunomi", "ace_of_hearts", "inkwell", "compass", "ipod", "sunscreen", "rocket", "cobblestone", ],
-#キャラクター設定／角色設定
-"Character Design"    : ["+boys", "+girls", "1other", "39", "_boys", "_challenge", "_connection", "_female", "_fur", "_girls", "_interface", "_male", "_man", "_person", "abyssal_ship", "age_difference", "aged_down", "aged_up", "albino", "alien", "alternate_muscle_size", "ambiguous_gender", "amputee", "androgynous", "angel", "animalization", "ass-to-ass", "assault_visor", "au_ra", "baby", "bartender", "beak", "bishounen", "borrowed_character", "boxers", "boy", "breast_envy", "breathing_fire", "bride", "broken", "brother_and_sister", "brothers", "camouflage", "cheating_(relationship)", "cheerleader", "chibi", "child", "clone", "command_spell", "comparison", "contemporary", "corpse", "corruption", "cosplay", "couple", "creature_and_personification", "crossdressing", "crossover", "cyberpunk", "cyborg", "cyclops", "damaged", "dancer", "danmaku", "darkness", "death", "defeat", "demon", "disembodied_", "draph", "drone", "duel", "dwarf", "egyptian", "electricity", "elezen", "elf", "enmaided", "erune", "everyone", "evolutionary_line", "expressions", "fairy", "family", "fangs", "fantasy", "fashion", "fat", "father_and_daughter", "father_and_son", "fewer_digits", "fins", "flashback", "fluffy", "fumo_(doll)", "furry", "fusion", "fuuin_no_tsue", "gameplay_mechanics", "genderswap", "ghost", "giant", "giantess", "gibson_les_paul", "girl", "goblin", "groom", "guro", "gyaru", "habit", "harem", "harpy", "harvin", "heads_together", "health_bar", "height_difference", "hitodama", "horror_(theme)", "humanization", "husband_and_wife", "hydrokinesis", "hypnosis", "hyur", "idol", "insignia", "instant_loss", "interracial", "interspecies", "japari_bun", "jeweled_branch_of_hourai", "jiangshi", "jirai_kei", "joints", "karakasa_obake", "keyhole", "kitsune", "knight", "kodona", "kogal", "kyuubi", "lamia", "left-handed", "loli", "lolita", "look-alike", "machinery", "magic", "male_focus", "manly", "matching_outfits", "mature_female", "mecha", "mermaid", "meta", "miko", "milestone_celebration", "military", "mind_control", "miniboy", "minigirl", "miqo'te", "monster", "monsterification", "mother_and_daughter", "mother_and_son", "multiple_others", "muscular", "nanodesu_(phrase)", "narrow_waist", "nekomata", "netorare", "ninja", "no_humans", "nontraditional", "nun", "nurse", "object_namesake", "obliques", "office_lady", "old", "on_body", "onee-shota", "oni", "orc", "others", "otoko_no_ko", "oversized_object", "paint_splatter", "pantyshot", "pawpads", "persona", "personality", "personification", "pet_play", "petite", "pirate", "playboy_bunny", "player_2", "plugsuit", "plump", "poi", "pokemon", "police", "policewoman", "pom_pom_(cheerleading)", "princess", "prosthesis", "pun", "puppet", "race_queen", "radio_antenna", "real_life_insert", "redesign", "reverse_trap", "rigging", "robot", "rod_of_remorse", "sailor", "salaryman", "samurai", "sangvis_ferri", "scales", "scene_reference", "school", "sheikah", "shota", "shrine", "siblings", "side-by-side", "sidesaddle", "sisters", "size_difference", "skeleton", "skinny", "slave", "slime_(substance)", "soldier", "spiked_shell", "spokencharacter", "steampunk", "streetwear", "striker_unit", "strongman", "submerged", "suggestive", "super_saiyan", "superhero", "surreal", "take_your_pick", "tall", "talons", "taur", "teacher", "team_rocket", "three-dimensional_maneuver_gear", "time_paradox", "tomboy", "traditional_youkai", "transformation", "trick_or_treat", "tusks", "twins", "ufo", "under_covers", "v-fin", "v-fin", "vampire", "virtual_youtuber", "waitress", "watching_television", "wedding", "what", "when_you_see_it", "wife_and_wife", "wing", "wings", "witch", "world_war_ii", "yandere", "year_of", "yes", "yin_yang", "yordle", "you're_doing_it_wrong", "you_gonna_get_raped", "yukkuri_shiteitte_ne", "yuri", "zombie", "(alice_in_wonderland)", "(arknights)", "(blue_archive)", "(cosplay)", "(creature)", "(emblem)", "(evangelion)", "(fate)", "(fate/stay_night)", "(ff11)", "(fire_emblem)", "(genshin_impact)", "(grimm)", "(houseki_no_kuni)", "(hyouka)", "(idolmaster)", "(jojo)", "(kancolle)", "(kantai_collection)", "(kill_la_kill)", "(league_of_legends)", "(legends)", "(lyomsnpmp)", "(machimazo)", "(madoka_magica)", "(mecha)", "(meme)", "(nier:automata)", "(organ)", "(overwatch)", "(pokemon)", "(project_moon)", "(project_sekai)", "(sao)", "(senran_kagura)", "(splatoon)", "(touhou)", "(tsukumo_sana)", "(youkai_watch)", "(yu-gi-oh!_gx)", "(zelda)", "sextuplets", "imperial_japanese_army", "extra_faces", "_miku", ], 
-#構図／構圖
-"Composition"         : ["abstract", "anime_coloring", "animification", "back-to-back", "bad_anatomy", "blurry", "border", "bound", "cameo", "cheek-to-cheek", "chromatic_aberration", "close-up", "collage", "color_guide", "colorful", "comic", "contrapposto", "cover", "cowboy_shot", "crosshatching", "depth_of_field", "dominatrix", "dutch_angle", "_focus", "face-to-face", "fake_screenshot", "film_grain", "fisheye", "flat_color", "foreshortening", "from_above", "from_behind", "from_below", "from_side", "full_body", "glitch", "greyscale", "halftone", "head_only", "heads-up_display", "high_contrast", "horizon", "_inset", "inset", "jaggy_lines", "1koma", "2koma", "3koma", "4koma", "5koma", "leaning", "leaning_forward", "leaning_to_the_side", "left-to-right_manga", "lens_flare", "limited_palette", "lineart", "lineup", "lower_body", "(medium)", "marker_(medium)", "meme", "mixed_media", "monochrome", "multiple_views", "muted_color", "oekaki", "on_side", "out_of_frame", "outline", "painting", "parody", "partially_colored", "partially_underwater_shot", "perspective", "photorealistic", "picture_frame", "pillarboxed", "portrait", "poster_(object)", "product_placement", "profile", "realistic", "recording", "retro_artstyle", "(style)", "_style", "sandwiched", "science_fiction", "sepia", "shikishi", "side-by-side", "sideways", "sideways_glance", "silhouette", "sketch", "spot_color", "still_life", "straight-on", "symmetry", "(texture)", "tachi-e", "taking_picture", "tegaki", "too_many", "traditional_media", "turnaround", "underwater", "upper_body", "upside-down", "upskirt", "variations", "wide_shot", "_design", "symbolism", "rounded_corners", "surrounded", ],
-#季節／季節
-"Season"              : ["akeome", "anniversary", "autumn", "birthday", "christmas", "_day", "festival", "halloween", "kotoyoro", "nengajou", "new_year", "spring_(season)", "summer", "tanabata", "valentine", "winter", ],
-#背景／背景
-"Background"          : ["_background", "backlighting", "bloom", "bokeh", "brick_wall", "bubble", "cable", "caustics", "cityscape", "cloud", "confetti", "constellation", "contrail", "crowd", "crystal", "dark", "debris", "dusk", "dust", "egasumi", "embers", "emphasis_lines", "energy", "evening", "explosion", "fireworks", "fog", "footprints", "glint", "graffiti", "ice", "industrial_pipe", "landscape", "light", "light_particles", "light_rays", "lightning", "lights", "moonlight", "motion_blur", "motion_lines", "mountainous_horizon", "nature", "(planet)", "pagoda", "people", "pillar", "planet", "power_lines", "puddle", "rain", "rainbow", "reflection", "ripples", "rubble", "ruins", "scenery", "shade", "shooting_star", "sidelighting", "smoke", "snowflakes", "snowing", "space", "sparkle", "sparks", "speed_lines", "spider_web", "spotlight", "star_(sky)", "stone_wall", "sunbeam", "sunburst", "sunrise", "_theme", "tile_wall", "twilight", "wall_clock", "wall_of_text", "water", "waves", "wind", "wire", "wooden_wall", "lighthouse", ],
-# パターン／圖案
-"Patterns"            : ["arrow", "bass_clef", "blank_censor", "circle", "cube", "heart", "hexagon", "hexagram", "light_censor", "(pattern)", "pattern", "pentagram", "roman_numeral", "(shape)", "(symbol)", "shape", "sign", "symbol", "tally", "treble_clef", "triangle", "tube", "yagasuri", ],
-#検閲／審查
-"Censorship"          : ["blur_censor", "_censor", "_censoring", "censored", "character_censor", "convenient", "hair_censor", "heart_censor", "identity_censor", "maebari", "novelty_censor", "soap_censor", "steam_censor", "tail_censor", "uncensored", ],
-#その他／其他
-"Others"              : ["2007", "2008", "2009", "2010", "2011", "2012", "2013", "2014", "2015", "2016", "2017", "2018", "2019", "2020", "2021", "2022", "2023", "2024", "artist", "artist_name", "artistic_error", "asian", "(company)", "character_name", "content_rating", "copyright", "cover_page", "dated", "english_text", "japan", "layer", "logo", "name", "numbered", "page_number", "pixiv_id", "ranguage", "reference_sheet", "signature", "speech_bubble", "subtitled", "text", "thank_you", "typo", "username", "wallpaper", "watermark", "web_address", "screwdriver", "translated", ],
-"Quality Tags"              : ["masterpiece", "_quality", "highres", "absurdres", "ultra-detailed", "lowres", ],
-}
-
-reversed_categories = {value: key for key, values in categories.items() for value in values}
-
-# Precompute keyword lengths
-keyword_lengths = {keyword: len(keyword) for keyword in reversed_categories}
-
-# Trie for efficient keyword matching
-class TrieNode:
-    def __init__(self):
-        self.children = {}
-        self.category = None
-
-def build_trie(keywords):
-    root = TrieNode()
-    for keyword, category in reversed_categories.items():
-        node = root
-        for char in keyword:
-            if char not in node.children:
-                node.children[char] = TrieNode()
-            node = node.children[char]
-        node.category = category
-    return root
-
-trie_root = build_trie(reversed_categories)
-
-def find_category(trie_root, tag):
-    node = trie_root
-    for char in tag:
-        if char in node.children:
-            node = node.children[char]
-            if node.category:
-                return node.category
-        else:
-            break
-    return None
-
-def classify_tags(tags: list[str], local_test: bool = False):
-    # Dictionary for automatic classification
-    classified_tags: defaultdict[str, list] = defaultdict(list)
-    fuzzy_match_tags: defaultdict[str, list] = defaultdict(list)
-    unclassified_tags: list[str] = []
-
-    # Logic for automatic grouping
-    for tag in tags:
-        classified = False
-        tag_new = tag.replace(" ", "_").replace("-", "_").replace("\\(", "(").replace("\\)", ")")  # Replace spaces in source tags with underscores
-
-        # Exact match using the trie
-        category = find_category(trie_root, tag_new)
-        if category:
-            classified = True
-        else:
-            # Fuzzy match
-            tag_parts = tag_new.split("_")
-            for keyword, keyword_length in keyword_lengths.items():
-                if keyword in tag_new and keyword_length > 3:  # Adjust the threshold if needed
-                    classified = True
-                    category = reversed_categories[keyword]
-                    break
-
-        if classified and tag not in classified_tags[category]:  # Avoid duplicates
-            classified_tags[category].append(tag)
-        elif not classified and tag not in unclassified_tags:
-            unclassified_tags.append(tag)  # Unclassified tags
-
-    if local_test:
-        # Output the grouping result
-        for category, tags in classified_tags.items():
-            print(f"{category}:")
-            print(", ".join(tags))
-            print()
-
-        print()
-        print("Fuzzy match:")
-        for category, tags in fuzzy_match_tags.items():
-            print(f"{category}:")
-            print(", ".join(tags))
-            print()
-        print()
-
-        if len(unclassified_tags) > 0:
-            print(f"\nUnclassified tags: {len(unclassified_tags)}")
-            print(f"{unclassified_tags[:200]}")  # Display some unclassified tags
-
-    return classified_tags, unclassified_tags
-
-# Code for "Tag Categorizer" tab
-def process_tags(input_tags: str):
-    # Clean and split the input tags <- Fix later
-    # tags = [tag.strip().split()[0] for tag in input_tags.split('?') if tag.strip()]
-    # tags = [tag.replace('_', ' ') for tag in tags]
-    tags = [tag.strip() for tag in input_tags.split(',') if tag.strip()]
-    classified_tags, unclassified_tags = classify_tags(tags)
-
-    categorized_string = ', '.join([tag for category in classified_tags.values() for tag in category])
-    categorized_json = {category: tags for category, tags in classified_tags.items()}
-
-    return categorized_string, categorized_json, ""  # Initialize enhanced_prompt as empty
-
-tags = []
+from collections import defaultdict
+import re
+# Define grouping rules (categories and keywords)
+# Provided categories and reversed_categories
+categories = {
+"Explicit"            : ["sex", "69", "paizuri", "cum", "precum", "areola_slip", "hetero", "erection", "oral", "fellatio", "yaoi", "ejaculation", "ejaculating", "masturbation", "handjob", "bulge", "rape", "_rape", "doggystyle", "threesome", "missionary", "object_insertion", "nipple", "nipples", "pussy", "anus", "penis", "groin", "testicles", "testicle", "anal", "cameltoe", "areolae", "dildo", "clitoris", "top-down_bottom-up", "gag", "groping", "gagged", "gangbang", "orgasm", "femdom", "incest", "bukkake", "breast_out", "vaginal", "vagina", "public_indecency", "breast_sucking", "folded", "cunnilingus", "_cunnilingus", "foreskin", "bestiality", "footjob", "uterus", "womb", "flaccid", "defloration", "butt_plug", "cowgirl_position", "reverse_cowgirl_position", "squatting_cowgirl_position", "reverse_upright_straddle", "irrumatio", "deepthroat", "pokephilia", "gaping", "orgy", "cleft_of_venus", "futanari", "futasub", "futa", "cumdrip", "fingering", "vibrator", "partially_visible_vulva", "penetration", "penetrated", "cumshot", "exhibitionism", "breast_milk", "grinding", "clitoral", "urethra", "phimosis", "cervix", "impregnation", "tribadism", "molestation", "pubic_hair", "clothed_female_nude_male", "clothed_male_nude_female", "clothed_female_nude_female", "clothed_male_nude_male", "sex_machine", "milking_machine", "ovum", "chikan", "pussy_juice_drip_through_clothes", "ejaculating_while_penetrated", "suspended_congress", "reverse_suspended_congress", "spread_pussy_under_clothes", "anilingus", "reach-around", "humping", "consensual_tentacles", "tentacle_pit", "cum_in_", ],
+#外観状態／外觀狀態
+"Appearance Status"   : ["backless", "bandaged_neck", "bleeding", "blood", "_blood", "blush", "body_writing", "bodypaint", "bottomless", "breath", "bruise", "butt_crack", "cold", "covered_mouth", "crack", "cross-section", "crotchless", "crying", "curvy", "cuts", "dirty", "dripping", "drunk", "from_mouth", "glowing", "hairy", "halterneck", "hot", "injury", "latex", "leather", "levitation", "lipstick_mark", "_markings", "makeup", "mole", "moles", "no_bra", "nosebleed", "nude", "outfit", "pantylines", "peeing", "piercing", "_piercing", "piercings", "pregnant", "public_nudity", "reverse", "_skin", "_submerged", "saliva", "scar", "scratches", "see-through", "shadow", "shibari", "sideless", "skindentation", "sleeping","tan", "soap_bubbles", "steam", "steaming_body", "stitches", "sweat", "sweatdrop", "sweaty", "tanlines", "tattoo", "tattoo", "tears", "topless", "transparent", "trefoil", "trembling", "veins", "visible_air", "wardrobe_malfunction", "wet", "x-ray", "unconscious", "handprint", ],
+#動作姿勢／動作姿勢
+"Action Pose"         : ["afloat", "afterimage", "against_fourth_wall", "against_wall", "aiming", "all_fours", "another's_mouth", "arm_", "arm_support", "arms_", "arms_behind_back", "asphyxiation", "attack", "back", "ballet", "bara", "bathing", "battle", "bdsm", "beckoning", "bent_over", "bite_mark", "biting", "bondage", "breast_suppress", "breathing", "burning", "bust_cup", "carry", "carrying", "caught", "chained", "cheek_squash", "chewing", "cigarette", "clapping", "closed_eye", "come_hither", "cooking", "covering", "cuddling", "dancing", "_docking", "destruction", "dorsiflexion", "dreaming", "dressing", "drinking", "driving", "dropping", "eating", "exercise", "expansion", "exposure", "facing", "failure", "fallen_down", "falling", "feeding", "fetal_position", "fighting", "finger_on_trigger", "finger_to_cheek", "finger_to_mouth", "firing", "fishing", "flashing", "fleeing", "flexible", "flexing", "floating", "flying", "fourth_wall", "freediving", "frogtie", "_grab", "girl_on_top", "giving", "grabbing", "grabbing_", "gymnastics", "_hold", "hadanugi_dousa", "hairdressing", "hand_", "hand_on", "hand_on_wall", "hands_", "headpat", "hiding", "holding", "hug", "hugging", "imagining", "in_container", "in_mouth", "in_palm", "jealous", "jumping", "kabedon", "kicking", "kiss", "kissing", "kneeling", "_lift", "lactation", "laundry", "licking", "lifted_by_self", "looking", "lowleg", "lying", "melting", "midair", "moaning", "_open", "on_back", "on_bed", "on_ground", "on_lap", "on_one_knee", "one_eye_closed", "open_", "over_mouth", "own_mouth", "_peek", "_pose", "_press", "_pull", "padding", "paint", "painting_(action)", "palms_together", "pee", "peeking", "pervert", "petting", "pigeon-toed", "piggyback", "pinching", "pinky_out", "pinned", "plantar_flexion", "planted", "playing", "pocky", "pointing", "poke", "poking", "pouring", "pov", "praying", "presenting", "profanity", "pulled_by_self", "pulling", "pump_action", "punching", "_rest", "raised", "reaching", "reading", "reclining", "reverse_grip", "riding", "running", "_slip", "salute", "screaming", "seiza", "selfie", "sewing", "shaking", "shoe_dangle", "shopping", "shouting", "showering", "shushing", "singing", "sitting", "slapping", "smell", "smelling", "smoking", "smother", "solo", "spanked", "spill", "spilling", "spinning", "splashing", "split", "squatting", "squeezed", "breasts_squeezed_together", "standing", "standing_on_", "staring", "straddling", "strangling", "stretching", "surfing", "suspension", "swimming", "talking", "teardrop", "tearing_clothes", "throwing", "tied_up", "tiptoes", "toe_scrunch", "toothbrush", "trigger_discipline", "tripping", "tsundere", "turning_head", "twitching", "two-handed", "tying", "_up",  "unbuttoned", "undressed", "undressing", "unsheathed", "unsheathing", "unzipped", "unzipping", "upright_straddle", "v", "V", "vore", "_wielding","wading", "walk-in", "walking", "wariza", "waving", "wedgie", "wrestling", "writing", "yawning", "yokozuwari", "_conscious", "massage", "struggling", "shrugging", "drugged", "tentacles_under_clothes", "restrained_by_tentacles", "tentacles_around_arms", "tentacles_around_legs", "restrained_legs", "restrained_tail", "restrained_arms", "tentacles_on_female", "archery", "cleaning", "tempura", "facepalm", "sadism", ],
+#頭部装飾／頭部服飾
+"Headwear"            : ["antennae", "antlers", "aura", "bandaged_head", "bandana", "bandeau", "beanie", "beanie", "beret", "bespectacled", "blindfold", "bonnet", "_cap", "circlet", "crown", "_drill", "_drills", "diadem", "_eyewear", "ear_covers", "ear_ornament", "ear_tag", "earbuds", "earclip", "earmuffs", "earphones", "earpiece", "earring", "earrings", "eyeliner", "eyepatch", "eyewear_on_head", "facial", "fedora", "glasses", "goggles", "_headwear", "hachimaki", "hair_bobbles", "hair_ornament", "hair_rings", "hair_tie", "hairband", "hairclip", "hairpin", "hairpods", "halo", "hat", "head-mounted_display", "head_wreath", "headband", "headdress", "headgear", "headphones", "headpiece", "headset", "helm", "helmet", "hood", "kabuto_(helmet)", "kanzashi", "_mask", "maid_headdress", "mask", "mask", "mechanical_ears", "mechanical_eye", "mechanical_horns", "mob_cap", "monocle", "neck_ruff", "nightcap", "on_head", "pince-nez", "qingdai_guanmao", "scarf_over_mouth", "scrunchie", "sunglasses", "tam_o'_shanter", "tate_eboshi", "tiara", "topknot", "turban", "veil", "visor", "wig", "mitre", "tricorne", "bicorne", ],
+#手部装飾／手部服飾
+"Handwear"            : ["arm_warmers", "armband", "armlet", "bandaged_arm", "bandaged_fingers", "bandaged_hand", "bandaged_wrist", "bangle", "bracelet", "bracelets", "bracer", "cuffs", "elbow_pads", "_gauntlets", "_glove", "_gloves", "gauntlets", "gloves", "kote", "kurokote", "mechanical_arm", "mechanical_arms", "mechanical_hands", "mittens", "mitts", "nail_polish", "prosthetic_arm", "wrist_cuffs", "wrist_guards", "wristband", "yugake", ],
+#ワンピース衣装／一件式服裝
+"One-Piece Outfit"    : ["bodystocking", "bodysuit", "dress", "furisode", "gown", "hanfu", "jumpsuit", "kimono", "leotard", "microdress", "one-piece", "overalls", "robe", "spacesuit", "sundress", "yukata", ],
+#上半身衣装／上半身服裝
+"Upper Body Clothing" : ["aiguillette", "apron", "_apron", "armor", "_armor", "ascot", "babydoll", "bikini", "_bikini", "blazer", "_blazer", "blouse", "_blouse", "bowtie", "_bowtie", "bra", "_bra", "breast_curtain", "breast_curtains", "breast_pocket", "breastplate", "bustier", "camisole", "cape", "capelet", "cardigan", "center_opening", "chemise", "chest_jewel", "choker", "cloak", "coat", "coattails", "collar", "_collar", "corset", "criss-cross_halter", "crop_top", "dougi", "feather_boa", "gakuran", "hagoromo", "hanten_(clothes)", "haori", "harem_pants", "harness", "hoodie", "jacket", "_jacket", "japanese_clothes", "kappougi", "kariginu", "lapels", "lingerie", "_lingerie", "maid", "mechanical_wings", "mizu_happi", "muneate", "neckerchief", "necktie", "negligee", "nightgown", "pajamas", "_pajamas", "pauldron", "pauldrons", "plunging_neckline", "raincoat", "rei_no_himo", "sailor_collar", "sarashi", "scarf", "serafuku", "shawl", "shirt", "shoulder_", "sleepwear", "sleeve", "sleeveless", "sleeves", "_sleeves", "sode", "spaghetti_strap", "sportswear", "strapless", "suit", "sundress", "suspenders", "sweater", "swimsuit", "_top", "_torso", "t-shirt", "tabard", "tailcoat", "tank_top", "tasuki", "tie_clip", "tunic", "turtleneck", "tuxedo", "_uniform", "undershirt", "uniform", "v-neck", "vambraces", "vest", "waistcoat", ],
+#下半身衣装／下半身服裝
+"Lower Body Clothing" : ["bare_hips", "bloomers", "briefs", "buruma", "crotch_seam", "cutoffs", "denim", "faulds", "fundoshi", "g-string", "garter_straps", "hakama", "hip_vent", "jeans", "knee_pads", "loincloth", "mechanical_tail", "microskirt", "miniskirt", "overskirt", "panties", "pants", "pantsu", "panty_straps", "pelvic_curtain", "petticoat", "sarong", "shorts", "side_slit", "skirt", "sweatpants", "swim_trunks", "thong", "underwear", "waist_cape", ],
+#足元・レッグウェア／腳與腿部服飾
+"Foot & Legwear"      : ["anklet", "bandaged_leg", "boot", "boots", "_footwear", "flats", "flip-flops", "geta", "greaves", "_heels", "kneehigh", "kneehighs", "_legwear", "leg_warmers", "leggings", "loafers", "mary_janes", "mechanical_legs", "okobo", "over-kneehighs", "pantyhose", "prosthetic_leg", "pumps", "_shoe", "_sock", "sandals", "shoes", "skates", "slippers", "sneakers", "socks", "spikes", "tabi", "tengu-geta", "thigh_strap", "thighhighs", "uwabaki", "zouri", "legband", "ankleband", ],
+#その他の装飾／其他服飾
+"Other Accessories"   : ["alternate_", "anklet", "badge", "beads", "belt", "belts", "bow", "brooch", "buckle", "button", "buttons", "_clothes", "_costume", "_cutout", "casual", "charm", "clothes_writing", "clothing_aside", "costume", "cow_print", "cross", "d-pad", "double-breasted", "drawstring", "epaulettes", "fabric", "fishnets", "floral_print", "formal", "frills", "_garter", "gem", "holster", "jewelry", "_knot", "lace", "lanyard", "leash", "magatama", "mechanical_parts", "medal", "medallion", "naked_bandage", "necklace", "_ornament", "(ornament)", "o-ring", "obi", "obiage", "obijime", "_pin", "_print", "padlock", "patterned_clothing", "pendant", "piercing", "plaid", "pocket", "polka_dot", "pom_pom_(clothes)", "pom_pom_(clothes)", "pouch", "ribbon", "_ribbon", "_stripe", "_stripes", "sash", "shackles", "shimenawa", "shrug_(clothing)", "skin_tight", "spandex", "strap", "sweatband", "_trim", "tassel", "zettai_ryouiki", "zipper", ],
+#表情／表情
+"Facial Expression"   : ["ahegao", "anger_vein", "angry", "annoyed", "confused", "drooling", "embarrassed", "expressionless", "eye_contact", "_face", "frown", "fucked_silly", "furrowed_brow", "glaring", "gloom_(expression)", "grimace", "grin", "happy", "jitome", "laughing", "_mouth", "nervous", "notice_lines", "o_o", "parted_lips", "pout", "puff_of_air", "restrained", "sad", "sanpaku", "scared", "scowl", "serious", "shaded_face", "shy", "sigh", "sleepy", "smile", "smirk", "smug", "snot", "spoken_ellipsis", "spoken_exclamation_mark", "spoken_interrobang", "spoken_question_mark", "squiggle", "surprised", "tareme", "tearing_up", "thinking", "tongue", "tongue_out", "torogao", "tsurime", "turn_pale", "wide-eyed", "wince", "worried", "heartbeat", ],
+#絵文字／表情符號
+"Facial Emoji"        : ["!!", "!", "!?", "+++", "+_+", "...", "...?", "._.", "03:00", "0_0", ":/", ":3", ":<", ":>", ":>=", ":d", ":i", ":o", ":p", ":q", ":t", ":x", ":|", ";(", ";)", ";3", ";d", ";o", ";p", ";q", "=_=", ">:(", ">:)", ">_<", ">_o", ">o<", "?", "??", "@_@", "\m/", "\n/", "\o/", "\||/", "^^^", "^_^", "c:", "d:", "o_o", "o3o", "u_u", "w", "x", "x_x", "xd", "zzz", "|_|", ],
+#頭部／頭部
+"Head"                : ["afro", "ahoge", "animal_ear_fluff", "_bangs", "_bun", "bald", "beard", "blunt_bangs", "blunt_ends", "bob_cut", "bowl_cut", "braid", "braids", "buzz_cut", "circle_cut", "colored_tips", "cowlick", "dot_nose", "dreadlocks", "_ear", "_ears", "_eye", "_eyes", "enpera", "eyeball", "eyebrow", "eyebrow_cut", "eyebrows", "eyelashes", "eyeshadow", "faceless", "facepaint", "facial_mark", "fang", "forehead", "freckles", "goatee", "_hair", "_horn", "_horns", "hair_", "hair_bun", "hair_flaps", "hair_intakes", "hair_tubes", "half_updo", "head_tilt", "heterochromia", "hime_cut", "hime_cut", "horns", "in_eye", "inverted_bob", "kemonomimi_mode", "lips", "mascara", "mohawk", "mouth_", "mustache", "nose", "one-eyed", "one_eye", "one_side_up", "_pupils", "parted_bangs", "pompadour", "ponytail", "ringlets", "_sclera", "sideburns", "sidecut", "sidelock", "sidelocks", "skull", "snout", "stubble", "swept_bangs", "tails", "teeth", "third_eye", "twintails", "two_side_up", "undercut", "updo", "v-shaped_eyebrows", "whiskers", "tentacle_hair", ],
+#手部／手部
+"Hands"               : ["_arm", "_arms", "claws", "_finger", "_fingers", "fingernails", "_hand", "_nail", "_nails", "palms", "rings", "thumbs_up", ],
+#上半身／上半身
+"Upper Body"          : ["abs", "armpit", "armpits", "backboob", "belly", "biceps", "breast_rest", "breasts", "button_gap", "cleavage", "collarbone", "dimples_of_venus", "downblouse", "flat_chest", "linea_alba", "median_furrow", "midriff", "nape", "navel", "pectorals", "ribs", "_shoulder", "_shoulders", "shoulder_blades", "sideboob", "sidetail", "spine", "stomach", "strap_gap", "toned", "underboob", "underbust", ],
+#下半身／下半身
+"Lower Body"          : ["ankles", "ass", "barefoot", "crotch", "feet", "highleg", "hip_bones", "hooves", "kneepits", "knees", "legs", "soles", "tail", "thigh_gap", "thighlet", "thighs", "toenail", "toenails", "toes", "wide_hips", ],
+#生物／生物
+"Creature"            : ["(animal)", "anglerfish", "animal", "bear", "bee", "bird", "bug", "butterfly", "cat", "chick", "chicken", "chinese_zodiac", "clownfish", "coral", "crab", "creature", "crow", "dog", "dove", "dragon", "duck", "eagle", "fish", "fish", "fox", "fox", "frog", "frog", "goldfish", "hamster", "horse", "jellyfish", "ladybug", "lion", "mouse", "octopus", "owl", "panda", "penguin", "pig", "pigeon", "rabbit", "rooster", "seagull", "shark", "sheep", "shrimp", "snail", "snake", "squid", "starfish", "tanuki", "tentacles", "goo_tentacles", "plant_tentacles", "crotch_tentacles", "mechanical_tentacles", "squidward_tentacles", "suction_tentacles", "penis_tentacles", "translucent_tentacles", "back_tentacles", "red_tentacles", "green_tentacles", "blue_tentacles", "black_tentacles", "pink_tentacles", "purple_tentacles", "face_tentacles", "tentacles_everywhere", "milking_tentacles", "tiger", "turtle", "weasel", "whale", "wolf", "parrot", "sparrow", "unicorn", ],
+#植物／植物
+"Plant"               : ["bamboo", "bouquet", "branch", "bush", "cherry_blossoms", "clover", "daisy", "(flower)", "flower", "flower", "gourd", "hibiscus", "holly", "hydrangea", "leaf", "lily_pad", "lotus", "moss", "palm_leaf", "palm_tree", "petals", "plant", "plum_blossoms", "rose", "spider_lily", "sunflower", "thorns", "tree", "tulip", "vines", "wisteria", "acorn", ],
+#食べ物／食物
+"Food"                : ["apple", "baguette", "banana", "baozi", "beans", "bento", "berry", "blueberry", "bread", "broccoli", "burger", "cabbage", "cake", "candy", "carrot", "cheese", "cherry", "chili_pepper", "chocolate", "coconut", "cookie", "corn", "cream", "crepe", "cucumber", "cucumber", "cupcake", "curry", "dango", "dessert", "doughnut", "egg", "eggplant", "_(food)", "_(fruit)", "food", "french_fries", "fruit", "grapes", "ice_cream", "icing", "lemon", "lettuce", "lollipop", "macaron", "mandarin_orange", "meat", "melon", "mochi", "mushroom", "noodles", "omelet", "omurice", "onigiri", "onion", "pancake", "parfait", "pasties", "pastry", "peach", "pineapple", "pizza", "popsicle", "potato", "pudding", "pumpkin", "radish", "ramen", "raspberry", "rice", "roasted_sweet_potato", "sandwich", "sausage", "seaweed", "skewer", "spitroast", "spring_onion", "strawberry", "sushi", "sweet_potato", "sweets", "taiyaki", "takoyaki", "tamagoyaki", "tempurakanbea", "toast", "tomato", "vegetable", "wagashi", "wagashi", "watermelon", "jam", "popcorn", ],
+#飲み物／飲品
+"Beverage"            : ["alcohol", "beer", "coffee", "cola", "drink", "juice", "juice_box", "milk", "sake", "soda", "tea", "_tea", "whiskey", "wine", "cocktail", ],
+#音楽／音樂
+"Music"               : ["band", "baton_(conducting)", "beamed", "cello", "concert", "drum", "drumsticks", "eighth_note", "flute", "guitar", "harp", "horn", "(instrument)", "idol", "instrument", "k-pop", "lyre", "(music)", "megaphone", "microphone", "music", "musical_note", "phonograph", "piano", "plectrum", "quarter_note", "recorder", "sixteenth_note", "sound_effects", "trumpet", "utaite", "violin", "whistle", ],
+#武器・装備／武器・裝備
+"Weapons & Equipment" : ["ammunition", "arrow_(projectile)", "axe", "bandolier", "baseball_bat", "beretta_92", "bolt_action", "bomb", "bullet", "bullpup", "cannon", "chainsaw", "crossbow", "dagger", "energy_sword", "explosive", "fighter_jet", "gohei", "grenade", "gun", "hammer", "handgun", "holstered", "jet", "katana", "knife", "kunai", "lance", "mallet", "nata_(tool)", "polearm", "quiver", "rapier", "revolver", "rifle", "rocket_launcher", "scabbard", "scope", "scythe", "sheath", "sheathed", "shield", "shotgun", "shuriken", "spear", "staff", "suppressor", "sword", "tank", "tantou", "torpedo", "trident", "(weapon)", "wand", "weapon", "whip", "yumi_(bow)", "h&k_hk416", "rocket_launcher", "heckler_&_koch", "_weapon", ],
+#乗り物／交通器具
+"Vehicles"            : ["aircraft", "airplane", "bicycle", "boat", "car", "caterpillar_tracks", "flight_deck", "helicopter", "motor_vehicle", "motorcycle", "ship", "spacecraft", "spoiler_(automobile)", "train", "truck", "watercraft", "wheel", "wheelbarrow", "wheelchair", "inflatable_raft", ],
+#建物／建物
+"Buildings"           : ["apartment", "aquarium", "architecture", "balcony", "building", "cafe", "castle", "church", "gym", "hallway", "hospital", "house", "library", "(place)", "porch", "restaurant", "restroom", "rooftop", "shop", "skyscraper", "stadium", "stage", "temple", "toilet", "tower", "train_station", "veranda", ],
+#室内／室內
+"Indoor"              : ["bath", "bathroom", "bathtub", "bed", "bed_sheet", "bedroom", "blanket", "bookshelf", "carpet", "ceiling", "chair", "chalkboard", "classroom", "counter", "cupboard", "curtains", "cushion", "dakimakura", "desk", "door", "doorway", "drawer", "_floor", "floor", "futon", "indoors", "interior", "kitchen", "kotatsu", "locker", "mirror", "pillow", "room", "rug", "school_desk", "shelf", "shouji", "sink", "sliding_doors", "stairs", "stool", "storeroom", "table", "tatami", "throne", "window", "windowsill", "bathhouse", "chest_of_drawers", ],
+#屋外／室外
+"Outdoor"             : ["alley", "arch", "beach", "bridge", "bus_stop", "bush", "cave", "(city)", "city", "cliff", "crescent", "crosswalk", "day", "desert", "fence", "ferris_wheel", "field", "forest", "grass", "graveyard", "hill", "lake", "lamppost", "moon", "mountain", "night", "ocean", "onsen", "outdoors", "path", "pool", "poolside", "railing", "railroad", "river", "road", "rock", "sand", "shore", "sky", "smokestack", "snow", "snowball", "snowman", "street", "sun", "sunlight", "sunset", "tent", "torii", "town", "tree", "turret", "utility_pole", "valley", "village", "waterfall", ],
+#物品／物品
+"Objects"             : ["anchor", "android", "armchair", "(bottle)", "backpack", "bag", "ball", "balloon", "bandages", "bandaid", "bandaids", "banknote", "banner", "barcode", "barrel", "baseball", "basket", "basketball", "beachball", "bell", "bench", "binoculars", "board_game", "bone", "book", "bottle", "bowl", "box", "box_art", "briefcase", "broom", "bucket", "(chess)", "(computer)", "(computing)", "(container)", "cage", "calligraphy_brush", "camera", "can", "candle", "candlestand", "cane", "card", "cartridge", "cellphone", "chain", "chandelier", "chess", "chess_piece", "choko_(cup)", "chopsticks", "cigar", "clipboard", "clock", "clothesline", "coin", "comb", "computer", "condom", "controller", "cosmetics", "couch", "cowbell", "crazy_straw", "cup", "cutting_board", "dice", "digital_media_player", "doll", "drawing_tablet", "drinking_straw", "easel", "electric_fan", "emblem", "envelope", "eraser", "feathers", "figure", "fire", "fishing_rod", "flag", "flask", "folding_fan", "fork", "frying_pan", "(gemstone)", "game_console", "gears", "gemstone", "gift", "glass", "glowstick", "gold", "handbag", "handcuffs", "handheld_game_console", "hose", "id_card", "innertube", "iphone", "jack-o'-lantern", "jar", "joystick", "key", "keychain", "kiseru", "ladder", "ladle", "lamp", "lantern", "laptop", "letter", "letterboxed", "lifebuoy", "lipstick", "liquid", "lock", "lotion", "_machine", "map", "marker", "model_kit", "money", "monitor", "mop", "mug", "needle", "newspaper", "nintendo", "nintendo_switch", "notebook", "(object)", "ofuda", "orb", "origami", "(playing_card)", "pack", "paddle", "paintbrush", "pan", "paper", "parasol", "patch", "pc", "pen", "pencil", "pencil", "pendant_watch", "phone", "pill", "pinwheel", "plate", "playstation", "pocket_watch", "pointer", "poke_ball", "pole", "quill", "racket", "randoseru", "remote_control", "ring", "rope", "sack", "saddle", "sakazuki", "satchel", "saucer", "scissors", "scroll", "seashell", "seatbelt", "shell", "shide", "shopping_cart", "shovel", "shower_head", "silk", "sketchbook", "smartphone", "soap", "sparkler", "spatula", "speaker", "spoon", "statue", "stethoscope", "stick", "sticker", "stopwatch", "string", "stuffed_", "stylus", "suction_cups", "suitcase", "surfboard", "syringe", "talisman", "tanzaku", "tape", "teacup", "teapot", "teddy_bear", "television", "test_tube", "tiles", "tokkuri", "tombstone", "torch", "towel", "toy", "traffic_cone", "tray", "treasure_chest", "uchiwa", "umbrella", "vase", "vial", "video_game", "viewfinder", "volleyball", "wallet", "watch", "watch", "whisk", "whiteboard", "wreath", "wrench", "wristwatch", "yunomi", "ace_of_hearts", "inkwell", "compass", "ipod", "sunscreen", "rocket", "cobblestone", ],
+#キャラクター設定／角色設定
+"Character Design"    : ["+boys", "+girls", "1other", "39", "_boys", "_challenge", "_connection", "_female", "_fur", "_girls", "_interface", "_male", "_man", "_person", "abyssal_ship", "age_difference", "aged_down", "aged_up", "albino", "alien", "alternate_muscle_size", "ambiguous_gender", "amputee", "androgynous", "angel", "animalization", "ass-to-ass", "assault_visor", "au_ra", "baby", "bartender", "beak", "bishounen", "borrowed_character", "boxers", "boy", "breast_envy", "breathing_fire", "bride", "broken", "brother_and_sister", "brothers", "camouflage", "cheating_(relationship)", "cheerleader", "chibi", "child", "clone", "command_spell", "comparison", "contemporary", "corpse", "corruption", "cosplay", "couple", "creature_and_personification", "crossdressing", "crossover", "cyberpunk", "cyborg", "cyclops", "damaged", "dancer", "danmaku", "darkness", "death", "defeat", "demon", "disembodied_", "draph", "drone", "duel", "dwarf", "egyptian", "electricity", "elezen", "elf", "enmaided", "erune", "everyone", "evolutionary_line", "expressions", "fairy", "family", "fangs", "fantasy", "fashion", "fat", "father_and_daughter", "father_and_son", "fewer_digits", "fins", "flashback", "fluffy", "fumo_(doll)", "furry", "fusion", "fuuin_no_tsue", "gameplay_mechanics", "genderswap", "ghost", "giant", "giantess", "gibson_les_paul", "girl", "goblin", "groom", "guro", "gyaru", "habit", "harem", "harpy", "harvin", "heads_together", "health_bar", "height_difference", "hitodama", "horror_(theme)", "humanization", "husband_and_wife", "hydrokinesis", "hypnosis", "hyur", "idol", "insignia", "instant_loss", "interracial", "interspecies", "japari_bun", "jeweled_branch_of_hourai", "jiangshi", "jirai_kei", "joints", "karakasa_obake", "keyhole", "kitsune", "knight", "kodona", "kogal", "kyuubi", "lamia", "left-handed", "loli", "lolita", "look-alike", "machinery", "magic", "male_focus", "manly", "matching_outfits", "mature_female", "mecha", "mermaid", "meta", "miko", "milestone_celebration", "military", "mind_control", "miniboy", "minigirl", "miqo'te", "monster", "monsterification", "mother_and_daughter", "mother_and_son", "multiple_others", "muscular", "nanodesu_(phrase)", "narrow_waist", "nekomata", "netorare", "ninja", "no_humans", "nontraditional", "nun", "nurse", "object_namesake", "obliques", "office_lady", "old", "on_body", "onee-shota", "oni", "orc", "others", "otoko_no_ko", "oversized_object", "paint_splatter", "pantyshot", "pawpads", "persona", "personality", "personification", "pet_play", "petite", "pirate", "playboy_bunny", "player_2", "plugsuit", "plump", "poi", "pokemon", "police", "policewoman", "pom_pom_(cheerleading)", "princess", "prosthesis", "pun", "puppet", "race_queen", "radio_antenna", "real_life_insert", "redesign", "reverse_trap", "rigging", "robot", "rod_of_remorse", "sailor", "salaryman", "samurai", "sangvis_ferri", "scales", "scene_reference", "school", "sheikah", "shota", "shrine", "siblings", "side-by-side", "sidesaddle", "sisters", "size_difference", "skeleton", "skinny", "slave", "slime_(substance)", "soldier", "spiked_shell", "spokencharacter", "steampunk", "streetwear", "striker_unit", "strongman", "submerged", "suggestive", "super_saiyan", "superhero", "surreal", "take_your_pick", "tall", "talons", "taur", "teacher", "team_rocket", "three-dimensional_maneuver_gear", "time_paradox", "tomboy", "traditional_youkai", "transformation", "trick_or_treat", "tusks", "twins", "ufo", "under_covers", "v-fin", "v-fin", "vampire", "virtual_youtuber", "waitress", "watching_television", "wedding", "what", "when_you_see_it", "wife_and_wife", "wing", "wings", "witch", "world_war_ii", "yandere", "year_of", "yes", "yin_yang", "yordle", "you're_doing_it_wrong", "you_gonna_get_raped", "yukkuri_shiteitte_ne", "yuri", "zombie", "(alice_in_wonderland)", "(arknights)", "(blue_archive)", "(cosplay)", "(creature)", "(emblem)", "(evangelion)", "(fate)", "(fate/stay_night)", "(ff11)", "(fire_emblem)", "(genshin_impact)", "(grimm)", "(houseki_no_kuni)", "(hyouka)", "(idolmaster)", "(jojo)", "(kancolle)", "(kantai_collection)", "(kill_la_kill)", "(league_of_legends)", "(legends)", "(lyomsnpmp)", "(machimazo)", "(madoka_magica)", "(mecha)", "(meme)", "(nier:automata)", "(organ)", "(overwatch)", "(pokemon)", "(project_moon)", "(project_sekai)", "(sao)", "(senran_kagura)", "(splatoon)", "(touhou)", "(tsukumo_sana)", "(youkai_watch)", "(yu-gi-oh!_gx)", "(zelda)", "sextuplets", "imperial_japanese_army", "extra_faces", "_miku", ], 
+#構図／構圖
+"Composition"         : ["abstract", "anime_coloring", "animification", "back-to-back", "bad_anatomy", "blurry", "border", "bound", "cameo", "cheek-to-cheek", "chromatic_aberration", "close-up", "collage", "color_guide", "colorful", "comic", "contrapposto", "cover", "cowboy_shot", "crosshatching", "depth_of_field", "dominatrix", "dutch_angle", "_focus", "face-to-face", "fake_screenshot", "film_grain", "fisheye", "flat_color", "foreshortening", "from_above", "from_behind", "from_below", "from_side", "full_body", "glitch", "greyscale", "halftone", "head_only", "heads-up_display", "high_contrast", "horizon", "_inset", "inset", "jaggy_lines", "1koma", "2koma", "3koma", "4koma", "5koma", "leaning", "leaning_forward", "leaning_to_the_side", "left-to-right_manga", "lens_flare", "limited_palette", "lineart", "lineup", "lower_body", "(medium)", "marker_(medium)", "meme", "mixed_media", "monochrome", "multiple_views", "muted_color", "oekaki", "on_side", "out_of_frame", "outline", "painting", "parody", "partially_colored", "partially_underwater_shot", "perspective", "photorealistic", "picture_frame", "pillarboxed", "portrait", "poster_(object)", "product_placement", "profile", "realistic", "recording", "retro_artstyle", "(style)", "_style", "sandwiched", "science_fiction", "sepia", "shikishi", "side-by-side", "sideways", "sideways_glance", "silhouette", "sketch", "spot_color", "still_life", "straight-on", "symmetry", "(texture)", "tachi-e", "taking_picture", "tegaki", "too_many", "traditional_media", "turnaround", "underwater", "upper_body", "upside-down", "upskirt", "variations", "wide_shot", "_design", "symbolism", "rounded_corners", "surrounded", ],
+#季節／季節
+"Season"              : ["akeome", "anniversary", "autumn", "birthday", "christmas", "_day", "festival", "halloween", "kotoyoro", "nengajou", "new_year", "spring_(season)", "summer", "tanabata", "valentine", "winter", ],
+#背景／背景
+"Background"          : ["_background", "backlighting", "bloom", "bokeh", "brick_wall", "bubble", "cable", "caustics", "cityscape", "cloud", "confetti", "constellation", "contrail", "crowd", "crystal", "dark", "debris", "dusk", "dust", "egasumi", "embers", "emphasis_lines", "energy", "evening", "explosion", "fireworks", "fog", "footprints", "glint", "graffiti", "ice", "industrial_pipe", "landscape", "light", "light_particles", "light_rays", "lightning", "lights", "moonlight", "motion_blur", "motion_lines", "mountainous_horizon", "nature", "(planet)", "pagoda", "people", "pillar", "planet", "power_lines", "puddle", "rain", "rainbow", "reflection", "ripples", "rubble", "ruins", "scenery", "shade", "shooting_star", "sidelighting", "smoke", "snowflakes", "snowing", "space", "sparkle", "sparks", "speed_lines", "spider_web", "spotlight", "star_(sky)", "stone_wall", "sunbeam", "sunburst", "sunrise", "_theme", "tile_wall", "twilight", "wall_clock", "wall_of_text", "water", "waves", "wind", "wire", "wooden_wall", "lighthouse", ],
+# パターン／圖案
+"Patterns"            : ["arrow", "bass_clef", "blank_censor", "circle", "cube", "heart", "hexagon", "hexagram", "light_censor", "(pattern)", "pattern", "pentagram", "roman_numeral", "(shape)", "(symbol)", "shape", "sign", "symbol", "tally", "treble_clef", "triangle", "tube", "yagasuri", ],
+#検閲／審查
+"Censorship"          : ["blur_censor", "_censor", "_censoring", "censored", "character_censor", "convenient", "hair_censor", "heart_censor", "identity_censor", "maebari", "novelty_censor", "soap_censor", "steam_censor", "tail_censor", "uncensored", ],
+#その他／其他
+"Others"              : ["2007", "2008", "2009", "2010", "2011", "2012", "2013", "2014", "2015", "2016", "2017", "2018", "2019", "2020", "2021", "2022", "2023", "2024", "artist", "artist_name", "artistic_error", "asian", "(company)", "character_name", "content_rating", "copyright", "cover_page", "dated", "english_text", "japan", "layer", "logo", "name", "numbered", "page_number", "pixiv_id", "ranguage", "reference_sheet", "signature", "speech_bubble", "subtitled", "text", "thank_you", "typo", "username", "wallpaper", "watermark", "web_address", "screwdriver", "translated", ],
+"Quality Tags"              : ["masterpiece", "_quality", "highres", "absurdres", "ultra-detailed", "lowres", ],
+}
+
+reversed_categories = {value: key for key, values in categories.items() for value in values}
+
+# Precompute keyword lengths
+keyword_lengths = {keyword: len(keyword) for keyword in reversed_categories}
+
+# Trie for efficient keyword matching
+class TrieNode:
+    def __init__(self):
+        self.children = {}
+        self.category = None
+
+def build_trie(keywords):
+    root = TrieNode()
+    for keyword, category in reversed_categories.items():
+        node = root
+        for char in keyword:
+            if char not in node.children:
+                node.children[char] = TrieNode()
+            node = node.children[char]
+        node.category = category
+    return root
+
+trie_root = build_trie(reversed_categories)
+
+def find_category(trie_root, tag):
+    node = trie_root
+    for char in tag:
+        if char in node.children:
+            node = node.children[char]
+            if node.category:
+                return node.category
+        else:
+            break
+    return None
+
+def classify_tags(tags: list[str], local_test: bool = False):
+    # Dictionary for automatic classification
+    classified_tags: defaultdict[str, list] = defaultdict(list)
+    fuzzy_match_tags: defaultdict[str, list] = defaultdict(list)
+    unclassified_tags: list[str] = []
+
+    # Logic for automatic grouping
+    for tag in tags:
+        classified = False
+        tag_new = tag.replace(" ", "_").replace("-", "_").replace("\\(", "(").replace("\\)", ")")  # Replace spaces in source tags with underscores
+
+        # Exact match using the trie
+        category = find_category(trie_root, tag_new)
+        if category:
+            classified = True
+        else:
+            # Fuzzy match
+            tag_parts = tag_new.split("_")
+            for keyword, keyword_length in keyword_lengths.items():
+                if keyword in tag_new and keyword_length > 3:  # Adjust the threshold if needed
+                    classified = True
+                    category = reversed_categories[keyword]
+                    break
+
+        if classified and tag not in classified_tags[category]:  # Avoid duplicates
+            classified_tags[category].append(tag)
+        elif not classified and tag not in unclassified_tags:
+            unclassified_tags.append(tag)  # Unclassified tags
+
+    if local_test:
+        # Output the grouping result
+        for category, tags in classified_tags.items():
+            print(f"{category}:")
+            print(", ".join(tags))
+            print()
+
+        print()
+        print("Fuzzy match:")
+        for category, tags in fuzzy_match_tags.items():
+            print(f"{category}:")
+            print(", ".join(tags))
+            print()
+        print()
+
+        if len(unclassified_tags) > 0:
+            print(f"\nUnclassified tags: {len(unclassified_tags)}")
+            print(f"{unclassified_tags[:200]}")  # Display some unclassified tags
+
+    return classified_tags, unclassified_tags
+
+# Code for "Tag Categorizer" tab
+def process_tags(input_tags: str):
+    # Split tags using regex to handle both commas and question marks
+    tags = []
+    for tag in re.split(r'\?|,|\n', input_tags):
+        tag = tag.strip()
+        if tag:
+            # Remove numbers at the end of tags
+            tag = re.sub(r'\b\d+\b', '', tag).strip()
+            
+            # Replace underscores with spaces
+            tag = tag.replace('_', ' ')
+            
+            # Escape parentheses (handle both escaped and unescaped)
+            if '(' in tag or ')' in tag:
+                # First, replace existing backslashes to handle properly
+                tag = tag.replace('\\', '')
+                # Replace parentheses with escaped versions
+                tag = tag.replace('(', r'\(').replace(')', r'\)')
+            
+            if tag:  # Only add if tag is not empty after processing
+                tags.append(tag)
+    
+    # Classify the cleaned tags
+    classified_tags, unclassified_tags = classify_tags(tags)
+    
+    # Create the outputs
+    categorized_string = ', '.join([tag for category in classified_tags.values() for tag in category])
+    categorized_json = {category: tags for category, tags in classified_tags.items()}
+    
+    return categorized_string, categorized_json, ""  # Initialize enhanced_prompt as empty