civit_caption_prepper.ipynb

{"cells":[{"cell_type":"code","source":["from google.colab import drive\n","drive.mount('/content/drive')"],"metadata":{"id":"hDzdSOe90dAa"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["import os\n","import io\n","import random\n","import zipfile\n","from google.colab import drive\n","from PIL import Image\n","import cv2\n","import numpy as np\n","import gc\n","\n","# ────────────────────────────────────────────────\n","# Configuration\n","# ────────────────────────────────────────────────\n","num_frames = 1000  #@param {type:\"slider\", min:1, max:1500, step:1}\n","split_images_horizontally = False  #@param {type:\"boolean\"}\n","use_manual_upload = False  #@param {type:\"boolean\"}\n","\n","save_to_google_drive = True   #@param {type:\"boolean\"}\n","drive_folder_name = \"vertical_slices_output\"  #@param {type:\"string\"}\n","\n","#@markdown ---\n","#@markdown **ZIP MODE** (when `use_manual_upload` is OFF)\n","#@markdown Paste full path to your zip file\n","zip_file_path = \"/content/drive/MyDrive/fetch.zip\"  #@param {type:\"string\"}\n","\n","# ────────────────────────────────────────────────\n","# Mount Google Drive\n","# ────────────────────────────────────────────────\n","drive_mounted = False\n","drive_output_dir = None\n","\n","if save_to_google_drive:\n","    try:\n","        drive.mount('/content/drive', force_remount=False)\n","        drive_mounted = True\n","        drive_base = \"/content/drive/MyDrive\"\n","        drive_output_dir = os.path.join(drive_base, drive_folder_name)\n","        os.makedirs(drive_output_dir, exist_ok=True)\n","        print(f\"→ Google Drive mounted. Output → {drive_output_dir}/vertical_slices.zip\")\n","    except Exception as e:\n","        print(f\"Drive mount failed: {e}\")\n","        save_to_google_drive = False\n","        print(\"→ No persistent saving possible.\")\n","else:\n","    print(\"Google Drive saving disabled → output only temporary!\")\n","\n","# ────────────────────────────────────────────────\n","# Settings\n","# ────────────────────────────────────────────────\n","FRAME_SIZE = 1024\n","BORDER_PX  = 14\n","INNER_SIZE = FRAME_SIZE - 2 * BORDER_PX\n","num_vertical_panels = 4  #@param {type:\"slider\", min:1, max:12}\n","NUM_SLICES = num_vertical_panels\n","SLICE_WIDTH = INNER_SIZE // NUM_SLICES\n","target_h = INNER_SIZE\n","target_w = SLICE_WIDTH\n","\n","BORDER_COLOR = (24, 24, 24)\n","\n","extract_root = \"/content/extracted_images\"\n","output_dir   = \"/content/vertical_slice_frames\"\n","\n","os.makedirs(extract_root, exist_ok=True)\n","os.makedirs(output_dir,   exist_ok=True)\n","\n","# ────────────────────────────────────────────────\n","# Helper – extract slices from single image\n","# ────────────────────────────────────────────────\n","def extract_vertical_slices(photo, target_w, target_h, border_color):\n","    slices = []\n","    try:\n","        w, h = photo.size\n","        if h <= 0 or w <= 0:\n","            return slices\n","\n","        scale = target_h / h\n","        new_w = int(w * scale + 0.5)\n","        if new_w < 1:\n","            return slices\n","\n","        resized = photo.resize((new_w, target_h), Image.LANCZOS)\n","\n","        if new_w < target_w:\n","            sl = Image.new('RGB', (target_w, target_h), border_color)\n","            sl.paste(resized, ((target_w - new_w)//2, 0))\n","            slices.append(sl)\n","        else:\n","            num = max(1, new_w // target_w)\n","            used = num * target_w\n","            start = (new_w - used) // 2\n","            for i in range(num):\n","                left = start + i * target_w\n","                crop = resized.crop((left, 0, left + target_w, target_h))\n","                slices.append(crop)\n","\n","        del resized\n","    except Exception as e:\n","        print(f\"Error extracting slices: {e}\")\n","    return slices\n","\n","# ────────────────────────────────────────────────\n","# Load images → extract slices immediately (low memory)\n","# ────────────────────────────────────────────────\n","all_slices = []\n","\n","print(\"Loading images and extracting slices...\\n\")\n","\n","if use_manual_upload:\n","    from google.colab import files\n","    print(\"Upload images (jpg/jpeg/png/webp)\")\n","    uploaded = files.upload()\n","\n","    for filename, filedata in uploaded.items():\n","        fname_lower = filename.lower()\n","        if fname_lower.startswith('._') or '__macosx' in fname_lower:\n","            continue\n","        if fname_lower.endswith(('.jpg','.jpeg','.png','.webp')):\n","            try:\n","                img = Image.open(io.BytesIO(filedata)).convert('RGB')\n","                temp_slices = extract_vertical_slices(img, target_w, target_h, BORDER_COLOR)\n","                all_slices.extend(temp_slices)\n","                del img, temp_slices\n","                gc.collect()\n","            except Exception as e:\n","                print(f\"Skip {filename}: {e}\")\n","\n","else:\n","    if not zip_file_path or not os.path.isfile(zip_file_path):\n","        print(\"Invalid or missing zip path.\")\n","    else:\n","        try:\n","            with zipfile.ZipFile(zip_file_path, 'r') as zf:\n","                zf.extractall(extract_root)\n","            print(\"Extraction finished.\")\n","        except Exception as e:\n","            print(f\"Zip extraction failed: {e}\")\n","            raise\n","\n","        valid_exts = ('.jpg', '.jpeg', '.png', '.webp')\n","        count = 0\n","        for root_dir, _, files in os.walk(extract_root):\n","            if '__MACOSX' in root_dir:\n","                continue\n","            for fname in files:\n","                if fname.startswith('._'):\n","                    continue\n","                if fname.lower().endswith(valid_exts):\n","                    path = os.path.join(root_dir, fname)\n","                    try:\n","                        # Quick verify\n","                        with Image.open(path) as im:\n","                            im.verify()\n","                        img = Image.open(path).convert('RGB')\n","                        temp_slices = extract_vertical_slices(img, target_w, target_h, BORDER_COLOR)\n","                        all_slices.extend(temp_slices)\n","                        del img, temp_slices\n","                        count += 1\n","                        if count % 20 == 0:\n","                            print(f\"Processed {count} images...\")\n","                            gc.collect()\n","                    except Exception as e:\n","                        print(f\"Skip {fname}: {str(e)}\")\n","\n","print(f\"\\nTotal vertical slices extracted: {len(all_slices)}\")\n","\n","# ────────────────────────────────────────────────\n","# Generate frames\n","# ────────────────────────────────────────────────\n","if len(all_slices) == 0:\n","    print(\"No slices available — stopping.\")\n","else:\n","    print(f\"Generating {num_frames} frames...\")\n","    frame_idx = 0\n","\n","    while frame_idx < num_frames:\n","        k = min(NUM_SLICES, len(all_slices))\n","        group = random.sample(all_slices * 4, k)\n","        random.shuffle(group)\n","\n","        canvas = Image.new('RGB', (INNER_SIZE, INNER_SIZE), (0,0,0))\n","        for col in range(NUM_SLICES):\n","            crop = group[col % len(group)]\n","            canvas.paste(crop, (col * SLICE_WIDTH, 0))\n","\n","        final = Image.new('RGB', (FRAME_SIZE, FRAME_SIZE), BORDER_COLOR)\n","        final.paste(canvas, (BORDER_PX, BORDER_PX))\n","\n","        fname = f\"frame_{frame_idx+1:04d}.jpg\"\n","        final.save(os.path.join(output_dir, fname), \"JPEG\", quality=82)\n","\n","        del canvas, final\n","        frame_idx += 1\n","\n","        if frame_idx % 50 == 0:\n","            print(f\"Saved {frame_idx}/{num_frames} frames...\")\n","        if frame_idx % 100 == 0:\n","            gc.collect()\n","\n","    print(f\"Finished — created {frame_idx} frames\")\n","\n","    # ─── Save fixed-name zip directly to Drive (overwrite) ───\n","    if save_to_google_drive and drive_mounted:\n","        final_zip_name = \"vertical_slices.zip\"\n","        local_zip_path = f\"/content/{final_zip_name}\"\n","        drive_zip_path = os.path.join(drive_output_dir, final_zip_name)\n","\n","        print(\"Creating zip (flat structure)...\")\n","\n","        # Create zip with files at root level\n","        with zipfile.ZipFile(local_zip_path, 'w', zipfile.ZIP_DEFLATED, compresslevel=6) as zf:\n","            for fname in sorted(os.listdir(output_dir)):\n","                if fname.endswith('.jpg'):\n","                    full_path = os.path.join(output_dir, fname)\n","                    zf.write(full_path, arcname=fname)   # ← flat: no subfolder\n","\n","        print(\"Copying to Google Drive (will overwrite previous version)...\")\n","        !cp -f \"{local_zip_path}\" \"{drive_zip_path}\"\n","\n","        print(f\"\\nSuccess! Overwritten file:\")\n","        print(f\"→ {drive_zip_path}\")\n","\n","        # Optional cleanup (uncomment if you want to free Colab disk space)\n","        # !rm -rf \"{output_dir}\" \"{local_zip_path}\"\n","        print(\"Temporary files kept in /content — delete manually if needed.\")\n","    else:\n","        print(\"\\nNo Drive save → frames are only in /content/vertical_slice_frames\")"],"metadata":{"id":"qiXhuNfyhVtJ"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["#@title WD Tagger + TOS Filter + Caption Cleaner + Tag Spreading → Drive (single-line + comma spacing) { run: \"auto\" }\n","from google.colab import drive\n","import os\n","import zipfile\n","from pathlib import Path\n","import shutil\n","from tqdm.auto import tqdm\n","import re\n","!pip install timm pillow pandas requests nltk -q\n","\n","import timm\n","import torch\n","from PIL import Image\n","import torchvision.transforms as transforms\n","import pandas as pd\n","import requests\n","from io import StringIO\n","import nltk\n","from collections import defaultdict\n","\n","# Fix for recent NLTK versions\n","nltk.download('punkt', quiet=True)\n","nltk.download('punkt_tab', quiet=True)\n","\n","# ────────────────────────────────────────────────\n","#@markdown ### Settings\n","drive.mount('/content/drive', force_remount=False)\n","\n","zip_path = \"/content/drive/MyDrive/vertical_slices_output/vertical_slices.zip\"                        #@param {type:\"string\"}\n","output_zip_name = \"cleaned_tagged_dataset.zip\"  #@param {type:\"string\"}\n","output_folder_on_drive = \"/content/drive/MyDrive/Cleaned_Datasets\"  #@param {type:\"string\"}\n","\n","case_sensitive_loli_check = False     #@param {type:\"boolean\"}\n","tag_probability_threshold = 0.35     #@param {type:\"slider\", min:0.1, max:0.6, step:0.05}\n","\n","# ────────────────────────────────────────────────\n","if not zip_path or not os.path.isfile(zip_path):\n","    print(\"❌ Please provide a valid zip file path\")\n","    raise SystemExit\n","\n","print(f\"📦 Input zip:  {zip_path}\")\n","print(f\"📤 Will save:  {output_folder_on_drive}/{output_zip_name}\\n\")\n","\n","# ────────────────────────────────────────────────\n","extract_dir = Path(\"/content/extracted\")\n","cleaned_dir = Path(\"/content/cleaned_dataset\")\n","\n","shutil.rmtree(extract_dir, ignore_errors=True)\n","shutil.rmtree(cleaned_dir, ignore_errors=True)\n","extract_dir.mkdir(exist_ok=True, parents=True)\n","cleaned_dir.mkdir(exist_ok=True, parents=True)\n","\n","print(\"📂 Extracting archive...\")\n","with zipfile.ZipFile(zip_path, 'r') as zf:\n","    zf.extractall(extract_dir)\n","\n","# ────────────────────────────────────────────────\n","# Load WD tagger\n","print(\"🔧 Loading WD tagger model...\")\n","tags_url = \"https://huggingface.co/SmilingWolf/wd-vit-tagger-v3/resolve/main/selected_tags.csv\"\n","tags_df = pd.read_csv(StringIO(requests.get(tags_url).text))\n","tags = tags_df['name'].tolist()\n","\n","model = timm.create_model(\"hf_hub:SmilingWolf/wd-vit-tagger-v3\", pretrained=True)\n","\n","device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n","model = model.eval().to(device)\n","\n","preprocess = transforms.Compose([\n","    transforms.Resize((448, 448)),\n","    transforms.ToTensor(),\n","    transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),\n","])\n","\n","def get_wd_tags(img_path):\n","    try:\n","        img = Image.open(img_path).convert(\"RGB\")\n","        x = preprocess(img).unsqueeze(0).to(device)\n","        with torch.no_grad():\n","            logits = model(x)\n","        probs = torch.sigmoid(logits).squeeze(0).cpu().numpy()\n","        selected = [tags[i] for i, p in enumerate(probs) if p > tag_probability_threshold]\n","        return selected\n","    except Exception as e:\n","        print(f\"  tagging failed: {img_path.name} → {str(e)}\")\n","        return []\n","\n","# ────────────────────────────────────────────────\n","def is_junk_file(path: Path) -> bool:\n","    \"\"\"Skip macOS metadata files and common junk\"\"\"\n","    name = path.name\n","    path_str = str(path)\n","    return (\n","        name.startswith('._') or\n","        '__MACOSX' in path_str or\n","        name in {'.DS_Store', 'Thumbs.db', 'desktop.ini', '.Spotlight-V100', '.Trashes'}\n","    )\n","\n","def normalize_skin_tags(tag_list: list) -> list:\n","    \"\"\"Replace blue_skin / colored_skin with plain 'skin' (only once)\"\"\"\n","    new_list = []\n","    has_skin = False\n","    for t in tag_list:\n","        if t in (\"blue_skin\", \"colored_skin\"):\n","            if not has_skin:\n","                new_list.append(\"skin\")\n","                has_skin = True\n","        else:\n","            new_list.append(t)\n","    return new_list\n","\n","def clean_caption(text: str) -> str:\n","    if not text.strip():\n","        return \"\"\n","\n","    # Remove unwanted characters & patterns\n","    text = text.replace('`', '')\n","    text = text.replace('(', '').replace(')', '')\n","    text = text.replace('*', '')\n","\n","    # Remove word 'small' (standalone, case insensitive)\n","    text = re.sub(r'\\bsmall\\b', '', text, flags=re.IGNORECASE)\n","\n","    # Collapse all whitespace including newlines, tabs, etc.\n","    text = re.sub(r'\\s+', ' ', text)\n","\n","    # Common safety / TOS replacements\n","    text = re.sub(r'\\byoung girl\\b', 'young woman', text, flags=re.IGNORECASE)\n","    text = re.sub(r'\\bswastika\\b', 'manji', text, flags=re.IGNORECASE)\n","    text = re.sub(r'\\byoung\\b', '', text, flags=re.IGNORECASE)\n","\n","    return text.strip()\n","\n","def spread_tags_into_caption(caption: str, new_tags: list) -> str:\n","    new_tags = normalize_skin_tags(new_tags)\n","\n","    if not new_tags:\n","        cleaned = clean_caption(caption)\n","        return ' , '.join(cleaned.split(',')).strip() if cleaned else ''\n","\n","    base = clean_caption(caption)\n","    if not base:\n","        return ' , '.join(new_tags)\n","\n","    sentences = nltk.sent_tokenize(base)\n","    if len(sentences) <= 1:\n","        combined = base + \" \" + \" , \".join(new_tags)\n","    else:\n","        # Distribute tags between sentences\n","        num_gaps = len(sentences) - 1\n","        tags_per_gap = max(1, len(new_tags) // num_gaps)\n","        extra = len(new_tags) % num_gaps\n","\n","        parts = []\n","        tag_idx = 0\n","        for i, sent in enumerate(sentences):\n","            parts.append(sent.strip())\n","            if i < num_gaps:\n","                cnt = tags_per_gap + (1 if i < extra else 0)\n","                if cnt > 0:\n","                    group = new_tags[tag_idx : tag_idx + cnt]\n","                    tag_idx += cnt\n","                    parts.append(\" , \".join(group))\n","\n","        # Remaining tags at the end\n","        if tag_idx < len(new_tags):\n","            parts.append(\" , \".join(new_tags[tag_idx:]))\n","\n","        combined = \" \".join(parts)\n","\n","    # Final cleanup: normalize comma spacing\n","    combined = re.sub(r'\\s*,\\s*', ' , ', combined)\n","    combined = re.sub(r'\\s+', ' ', combined).strip()\n","\n","    return combined\n","\n","# ────────────────────────────────────────────────\n","print(\"\\n🔍 Processing files...\\n\")\n","\n","removed = 0\n","kept = 0\n","\n","groups = defaultdict(list)\n","for f in extract_dir.rglob(\"*\"):\n","    if f.is_file() and not is_junk_file(f):\n","        groups[f.stem].append(f)\n","\n","for stem, files in tqdm(groups.items(), desc=\"Groups\"):\n","    # Filter again just in case\n","    valid_files = [f for f in files if not is_junk_file(f)]\n","\n","    imgs = [\n","        f for f in valid_files\n","        if f.suffix.lower() in {'.jpg', '.jpeg', '.png', '.webp', '.gif', '.bmp', '.tiff'}\n","    ]\n","    txts = [f for f in valid_files if f.suffix.lower() == '.txt']\n","\n","    if not imgs:\n","        continue\n","\n","    img = imgs[0]  # take the first valid image\n","    wd_tags = get_wd_tags(img)\n","\n","    # Loli check (case sensitive or not)\n","    joined_tags = ' '.join(wd_tags)\n","    has_loli = 'loli' in (joined_tags.lower() if not case_sensitive_loli_check else joined_tags)\n","\n","    if has_loli:\n","        removed += 1\n","        # Optional: remove files from temp dir (not strictly needed)\n","        # for f in valid_files: f.unlink(missing_ok=True)\n","        continue\n","\n","    kept += 1\n","\n","    # Read original caption if exists\n","    orig_caption = \"\"\n","    if txts:\n","        try:\n","            orig_caption = txts[0].read_text(encoding=\"utf-8\", errors=\"replace\").strip()\n","        except:\n","            pass\n","\n","    # Create final caption\n","    final_caption = spread_tags_into_caption(orig_caption, wd_tags)\n","\n","    # Copy only non-junk files to cleaned folder\n","    for f in valid_files:\n","        rel = f.relative_to(extract_dir)\n","        dst = cleaned_dir / rel\n","        dst.parent.mkdir(parents=True, exist_ok=True)\n","        shutil.copy2(f, dst)\n","\n","    # Write cleaned caption next to the image\n","    txt_name = img.stem + \".txt\"\n","    txt_rel = img.relative_to(extract_dir).parent / txt_name\n","    txt_dst = cleaned_dir / txt_rel\n","    txt_dst.parent.mkdir(parents=True, exist_ok=True)\n","    with open(txt_dst, \"w\", encoding=\"utf-8\") as fw:\n","        fw.write(final_caption)\n","\n","print(f\"\\n✅ Done processing\")\n","print(f\"   Removed (loli detected): {removed}\")\n","print(f\"   Kept & cleaned         : {kept}\")\n","\n","# ────────────────────────────────────────────────\n","print(\"\\n🗜️  Creating output zip...\")\n","\n","final_zip = Path(f\"/content/{output_zip_name}\")\n","\n","with zipfile.ZipFile(final_zip, \"w\", zipfile.ZIP_DEFLATED) as zf:\n","    for item in tqdm(cleaned_dir.rglob(\"*\"), desc=\"Zipping\"):\n","        if item.is_file() and not is_junk_file(item):\n","            arc = item.relative_to(cleaned_dir)\n","            zf.write(item, arc)\n","\n","# ────────────────────────────────────────────────\n","print(\"\\n💾 Copying to Drive...\")\n","os.makedirs(output_folder_on_drive, exist_ok=True)\n","drive_dest = Path(output_folder_on_drive) / output_zip_name\n","shutil.copy2(final_zip, drive_dest)\n","\n","size_mb = final_zip.stat().st_size / (1024 * 1024)\n","print(f\"→ Saved: {drive_dest}\")\n","print(f\"   Size: {size_mb:.1f} MiB\")\n","\n","# ────────────────────────────────────────────────\n","print(\"\\n🧹 Cleaning up temp folders...\")\n","shutil.rmtree(extract_dir, ignore_errors=True)\n","shutil.rmtree(cleaned_dir, ignore_errors=True)\n","\n","print(\"\\nAll finished ✓\")"],"metadata":{"id":"DCZpKdWX0UNZ"},"execution_count":null,"outputs":[]}],"metadata":{"colab":{"provenance":[{"file_id":"https://huggingface.co/datasets/codeShare/lora-training-data/blob/main/civit_caption_prepper.ipynb","timestamp":1773264797996},{"file_id":"https://huggingface.co/datasets/codeShare/lora-training-data/blob/main/civit_caption_prepper.ipynb","timestamp":1773163850245},{"file_id":"https://huggingface.co/datasets/codeShare/lora-training-data/blob/main/civit_caption_prepper.ipynb","timestamp":1773090196076},{"file_id":"https://huggingface.co/datasets/codeShare/lora-training-data/blob/main/civit_caption_prepper.ipynb","timestamp":1773089575687},{"file_id":"https://huggingface.co/datasets/codeShare/lora-training-data/blob/main/civit_caption_prepper.ipynb","timestamp":1773080355474},{"file_id":"https://huggingface.co/datasets/codeShare/lora-training-data/blob/main/Drive to WebP.ipynb","timestamp":1772998638620},{"file_id":"https://huggingface.co/datasets/codeShare/lora-training-data/blob/main/Drive to WebP.ipynb","timestamp":1763646205520},{"file_id":"https://huggingface.co/datasets/codeShare/lora-training-data/blob/main/Drive to WebP.ipynb","timestamp":1760993725927},{"file_id":"https://huggingface.co/datasets/codeShare/lora-training-data/blob/main/YT-playlist-to-mp3.ipynb","timestamp":1760450712160},{"file_id":"https://huggingface.co/datasets/codeShare/lora-training-data/blob/main/YT-playlist-to-mp3.ipynb","timestamp":1756712618300},{"file_id":"https://huggingface.co/codeShare/JupyterNotebooks/blob/main/YT-playlist-to-mp3.ipynb","timestamp":1747490904984},{"file_id":"https://huggingface.co/codeShare/JupyterNotebooks/blob/main/YT-playlist-to-mp3.ipynb","timestamp":1740037333374},{"file_id":"https://huggingface.co/codeShare/JupyterNotebooks/blob/main/YT-playlist-to-mp3.ipynb","timestamp":1736477078136},{"file_id":"https://huggingface.co/codeShare/JupyterNotebooks/blob/main/YT-playlist-to-mp3.ipynb","timestamp":1725365086834}],"gpuType":"T4"},"kernelspec":{"display_name":"Python 3","name":"python3"},"language_info":{"name":"python"},"accelerator":"GPU"},"nbformat":4,"nbformat_minor":0}