Upload folder using huggingface_hub

README.md

@@ -0,0 +1,2 @@
+# Haptix_image_dataset
+물리적 정보와 감각정 정보를 연결하고 라벨링한 데이터셋

README_IMAGE_ANALYSIS.md

@@ -0,0 +1,180 @@
+# 이미지 데이터셋 상태 분석 시스템
+
+## 개요
+이 노트북 파일(`iamge_status.ipynb`)은 Haptix 이미지 데이터셋의 라벨 분포를 자동으로 분석하고 시각화하는 종합 관리 시스템입니다.
+
+## 주요 기능
+
+### 1. **자동 라벨 인식**
+   - `_`(언더스코어)로 구분된 라벨을 자동으로 추출
+   - 새로운 폴더 추가 시 자동으로 인식
+   - 폴더명 형식: `LABEL1_LABEL2_LABEL3...`
+
+### 2. **다중 소스 지원**
+   - EmoSet_images
+   - Midjourney_images
+   - unsplash_images
+   - 새로운 소스 폴더 추가 시 자동 처리
+
+### 3. **생성 파일**
+
+#### 📊 시각화 파일
+- **`image_distribution_overview.png`**: 전체 라벨 분포 종합 분석
+  - 전체 라벨별 이미지 개수 (막대 그래프)
+  - 데이터 소스별 이미지 개수 비교
+  - 라벨 분포 비율 (파이 차트)
+  - 통계 정보 표시
+
+- **`image_distribution_by_source.png`**: 소스별 상세 분석
+  - EmoSet_images: 라벨별 분포
+  - Midjourney_images: 라벨별 분포
+  - unsplash_images: 라벨별 분포
+
+#### 📋 데이터 파일
+- **`image_label_distribution.csv`**: 라벨별 이미지 개수 (Excel/Google Sheets에서 열 수 있음)
+  - 전체 라벨 목록
+  - 각 소스별 이미지 개수
+  - Excel 등에서 추가 분석 가능
+
+## 분석 결과 요약
+
+### 📈 현재 상태 (2025-11-14)
+
+| 항목 | 값 |
+|------|-----|
+| **총 라벨 종류** | 22개 |
+| **총 이미지 개수** | 14,070개 |
+| **라벨당 평균** | 636.4개 |
+| **라벨당 최대값** | 1,575개 (P5Static) |
+| **라벨당 최소값** | 2개 (E5St, messy) |
+| **표준편차** | 470.0개 |
+
+### ⚠️ 주의 - 불균형 라벨
+
+평균의 **50% 이상 부족**한 라벨:
+- `P1Soft`: 293개 (평균의 46.0%)
+- `P1Hard`: 243개 (평균의 38.2%)
+- `energy`: 74개 (평균의 11.6%)
+- `pleasure`: 55개 (평균의 8.6%)
+- `unpleasnat`: 14개 (평균의 2.2%)
+- `E5St`: 2개 (평균의 0.3%)
+- `messy`: 2개 (평균의 0.3%)
+
+### 💡 권장사항
+
+**균형도 지표: 787.5x (최대값/최소값)**
+- ⚠️ **심각한 불균형이 존재합니다**
+
+#### 해결 방안:
+1. **데이터 보강 (Data Augmentation)**
+   - 부족한 라벨에 대해 이미지 회전, 확대/축소, 색상 변환 등 적용
+   - 특히 `E5St`, `messy`, `unpleasnat`, `pleasure`, `energy` 라벨 우선
+
+2. **모델 학습 시 클래스 가중치 조정**
+   ```python
+   class_weights = {
+       'P5Static': 1.0,      # 충분함
+       'E3N-Chaotic': 1.0,
+       # ... 
+       'E5St': 318.75,       # 매우 부족 (636.4 / 2)
+       'messy': 318.75,
+   }
+   ```
+
+3. **샘플링 전략**
+   - 언더샘플링(undersampling): 많은 라벨의 데이터 일부만 사용
+   - 오버샘플링(oversampling): 부족한 라벨의 데이터 반복 사용
+
+## 사용 방법
+
+### 1. 노트북 실행
+```bash
+# VS Code에서 Jupyter Notebook 커널 선택 후
+# 모든 셀을 순서대로 실행
+Ctrl + Shift + Enter  (현재 셀)
+또는
+셀 > 모두 실행 (Run All)
+```
+
+### 2. 새 이미지 소스 추가
+```bash
+# image 폴더에 새 폴더 추가 (예: new_source/)
+c:\Users\EL081\Desktop\backup\Haptix_image_dataset\image\new_source\
+    ├── LABEL1_LABEL2\
+    │   ├── image1.jpg
+    │   └── image2.png
+    └── LABEL3_LABEL4\
+        └── image3.jpg
+
+# 노트북을 다시 실행하면 자동으로 분석됨
+```
+
+### 3. 결과 해석
+- **빨간색 막대**: 평균 이하의 라벨 (주의 필요)
+- **파란색 막대**: 평균 이상의 라벨 (충분함)
+- **초록색 점선**: 평균값
+
+## 기술 사항
+
+### 지원 이미지 형식
+- JPG, JPEG, PNG, GIF, BMP, WEBP
+
+### 폴더 구조 자동 처리
+- 라벨 폴더 내 하위 폴더도 재귀적으로 검색
+- 파일명 무시 (라벨은 폴더명에서만 추출)
+
+### 확장성
+- 새 소스 추가 시 자동 인식
+- 새 라벨 추가 시 자동 처리
+- 코드 수정 불필요
+
+## 주의사항
+
+1. **파일명 규칙**
+   - 이미지가 포함된 폴더명은 반드시 `_`로 라벨을 구분해야 함
+   - 예: ✅ `P1Hard_E1N-Risky` / ❌ `P1Hard E1N-Risky`
+
+2. **이미지 개수**
+   - 폴더 내 모든 파일을 카운팅
+   - 비이미지 파일(텍스트, 문서)은 무시됨
+
+3. **성능**
+   - 수천 개 이미지 분석 시 약 15-20초 소요
+   - 처음 로드 시간이 가장 오래 걸림
+
+## 문제 해결
+
+### Q: 새 데이터를 추가했는데 반영이 안 됨
+A: 노트북의 **모든 셀을 다시 실행**해주세요. (`Ctrl + Shift + Enter`)
+
+### Q: CSV 파일이 Excel에서 한글이 깨짐
+A: CSV 파일을 메모장으로 열어 `UTF-8 BOM` 인코딩으로 저장 후 Excel에서 열기
+
+### Q: 특정 라벨만 분석하고 싶음
+A: 노트북의 첫 번째 셀에서 `base_path`를 수정하면 됨
+
+## 파일 목록
+
+```
+backup/
+├── iamge_status.ipynb                    ← 메인 분석 노트북
+├── image_distribution_overview.png       ← 전체 분석 그래프
+├── image_distribution_by_source.png      ← 소스별 분석 그래프
+├── image_label_distribution.csv          ← 라벨별 통계 데이터
+├── README_IMAGE_ANALYSIS.md              ← 이 파일
+└── Haptix_image_dataset/
+    ├── image/
+    │   ├── EmoSet_images/        (3,143개)
+    │   ├── Midjourney_images/    (17개)
+    │   └── unsplash_images/      (1,910개)
+    └── ...
+```
+
+## 라이선스 및 지원
+
+작성 날짜: 2025-11-14
+최종 업데이트: 2025-11-14
+
+---
+
+**질문이나 개선 사항이 있으면 노트북의 코드를 수정하거나 새로운 분석 셀을 추가할 수 있습니다.**

aaa.ipynb

@@ -0,0 +1,608 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "ac1c7423",
+   "metadata": {},
+   "source": [
+    "# Upload Image Dataset to Hugging Face (Preserve Folders)\n",
+    "\n",
+    "This minimal workflow uploads your local folder to a Hugging Face dataset repo, keeping the directory structure intact."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "277a6040",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Ready: huggingface_hub installed and imports ok.\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Install and import\n",
+    "import sys, subprocess, os\n",
+    "subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"-q\", \"huggingface_hub\", \"hf_transfer\"])\n",
+    "from huggingface_hub import HfApi\n",
+    "os.environ[\"HF_HUB_ENABLE_HF_TRANSFER\"] = \"1\"\n",
+    "print(\"Ready: huggingface_hub installed and imports ok.\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "32833fb5",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Token OK for: Smilesjs\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Set or prompt for HF token (safe)\n",
+    "import os\n",
+    "from getpass import getpass\n",
+    "from huggingface_hub import HfApi\n",
+    "\n",
+    "if not os.getenv(\"HF_TOKEN\"):\n",
+    "    token = getpass(\"Enter your HF_TOKEN (input hidden): \")\n",
+    "    os.environ[\"HF_TOKEN\"] = token\n",
+    "\n",
+    "# Optional: validate token early\n",
+    "try:\n",
+    "    _api = HfApi(token=os.environ[\"HF_TOKEN\"])\n",
+    "    who = _api.whoami()\n",
+    "    owner = who.get(\"name\") or who.get(\"email\") or who.get(\"id\")\n",
+    "    print(f\"Token OK for: {owner}\")\n",
+    "except Exception as e:\n",
+    "    raise RuntimeError(\"HF token seems invalid or network issue. Double-check the token.\") from e"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "id": "fe562d45",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Repo ready: https://huggingface.co/datasets/Smilesjs/Haptix_dataset\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Configure local path and repo\n",
+    "from pathlib import Path\n",
+    "import os\n",
+    "from huggingface_hub import HfApi\n",
+    "\n",
+    "LOCAL_FOLDER = Path(r\"c:\\\\Users\\\\EL081\\\\Desktop\\\\local_backup\\\\image\").resolve()\n",
+    "REPO_ID = \"Smilesjs/Haptix_dataset\"  # change if needed\n",
+    "REPO_TYPE = \"dataset\"\n",
+    "PRIVATE = True\n",
+    "\n",
+    "HF_TOKEN = os.getenv(\"HF_TOKEN\")\n",
+    "if not HF_TOKEN:\n",
+    "    raise ValueError(\"HF_TOKEN is not set. In PowerShell: $env:HF_TOKEN='hf_...'\")\n",
+    "\n",
+    "if not LOCAL_FOLDER.exists():\n",
+    "    raise FileNotFoundError(f\"Local folder not found: {LOCAL_FOLDER}\")\n",
+    "\n",
+    "api = HfApi(token=HF_TOKEN)\n",
+    "api.create_repo(repo_id=REPO_ID, repo_type=REPO_TYPE, private=PRIVATE, exist_ok=True)\n",
+    "print(f\"Repo ready: https://huggingface.co/datasets/{REPO_ID}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 34,
+   "id": "9da1f410",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Starting resumable upload to Smilesjs/Haptix_dataset with upload_large_folder... (this can take time)\n",
+      "Local: C:\\Users\\EL081\\Desktop\\local_backup\\image\n"
+     ]
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "2d2dc4acd4ed4ffb8d0ad51d3efda3f9",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Recovering from metadata files:   0%|          | 0/6920 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "\n",
+      "\n",
+      "---------- 2025-11-14 19:10:53 (0:00:00) ----------\n",
+      "Files:   hashed 6920/6920 (1.2G/1.2G) | pre-uploaded: 6918/6918 (1.2G/1.2G) | committed: 6920/6920 (1.2G/1.2G) | ignored: 0\n",
+      "Workers: hashing: 0 | get upload mode: 0 | pre-uploading: 0 | committing: 0 | waiting: 0\n",
+      "---------------------------------------------------\n",
+      "Upload complete: None\n",
+      "Upload complete: None\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Upload large folder preserving structure (resumable)\n",
+    "ignore_patterns = [\n",
+    "    \"**/.ipynb_checkpoints/**\",\n",
+    "    \"**/__pycache__/**\",\n",
+    "    \"**/*.tmp\",\n",
+    "    \"**/*.db\",\n",
+    "    \"**/Thumbs.db\",\n",
+    "    \"**/.DS_Store\",\n",
+    "]\n",
+    "\n",
+    "print(f\"Starting resumable upload to {REPO_ID} with upload_large_folder... (this can take time)\")\n",
+    "print(f\"Local: {LOCAL_FOLDER}\")\n",
+    "op = api.upload_large_folder(\n",
+    "    repo_id=REPO_ID,\n",
+    "    repo_type=REPO_TYPE,\n",
+    "    folder_path=str(LOCAL_FOLDER),\n",
+    "    # path_in_repo can be set to 'image' if you want a top-level 'image/' folder\n",
+    "    # path_in_repo=\"\",\n",
+    "    ignore_patterns=ignore_patterns,\n",
+    "    # allow_duplicates=False,  # uncomment to avoid duplicate files in future runs\n",
+    "    # commit_message=\"Upload dataset (resumable)\"\n",
+    ")\n",
+    "print(\"Upload complete:\", op)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 35,
+   "id": "21b788d2",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Scanning repo: https://huggingface.co/datasets/Smilesjs/Haptix_dataset\n",
+      "Cache invalidated for HfFileSystem.\n",
+      "Cache invalidated for HfFileSystem.\n",
+      "Found 3 candidate paths in 0.64s\n",
+      "API list_repo_files reports: 3 files\n",
+      "Fetched metadata for 3 files (0 errors) in 0.00s\n",
+      "\n",
+      "=== Summary ===\n",
+      "Total files: 3\n",
+      "Total size : 63.74 KB\n",
+      "\n",
+      "Top extensions by size:\n",
+      "  .png: 1 files, 61.33 KB\n",
+      "  (no ext): 1 files, 2.40 KB\n",
+      "  .txt: 1 files, 5.00 B\n",
+      "\n",
+      "Top-level folder counts:\n",
+      "  datasets: 3 files\n",
+      "\n",
+      "Top 20 largest files:\n",
+      "  61.33 KB  |  datasets/Smilesjs/Haptix_dataset/augmented_images/E3P-Harmonic_P3Rough_P2Cold_P5Static/E3P+P3R+P2Co+P5St_aug_001.png\n",
+      "  2.40 KB  |  datasets/Smilesjs/Haptix_dataset/.gitattributes\n",
+      "  5.00 B  |  datasets/Smilesjs/Haptix_dataset/test_probe.txt\n",
+      "Found 3 candidate paths in 0.64s\n",
+      "API list_repo_files reports: 3 files\n",
+      "Fetched metadata for 3 files (0 errors) in 0.00s\n",
+      "\n",
+      "=== Summary ===\n",
+      "Total files: 3\n",
+      "Total size : 63.74 KB\n",
+      "\n",
+      "Top extensions by size:\n",
+      "  .png: 1 files, 61.33 KB\n",
+      "  (no ext): 1 files, 2.40 KB\n",
+      "  .txt: 1 files, 5.00 B\n",
+      "\n",
+      "Top-level folder counts:\n",
+      "  datasets: 3 files\n",
+      "\n",
+      "Top 20 largest files:\n",
+      "  61.33 KB  |  datasets/Smilesjs/Haptix_dataset/augmented_images/E3P-Harmonic_P3Rough_P2Cold_P5Static/E3P+P3R+P2Co+P5St_aug_001.png\n",
+      "  2.40 KB  |  datasets/Smilesjs/Haptix_dataset/.gitattributes\n",
+      "  5.00 B  |  datasets/Smilesjs/Haptix_dataset/test_probe.txt\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Analyze remote repo file status (counts, sizes, top files)\n",
+    "from huggingface_hub import HfFileSystem\n",
+    "from concurrent.futures import ThreadPoolExecutor, as_completed\n",
+    "from collections import Counter, defaultdict\n",
+    "from pathlib import Path\n",
+    "import os, math, time\n",
+    "\n",
+    "if 'REPO_ID' not in globals():\n",
+    "    raise RuntimeError(\"REPO_ID is not defined. Run the setup cells first.\")\n",
+    "if 'HF_TOKEN' not in globals() or not HF_TOKEN:\n",
+    "    raise RuntimeError(\"HF_TOKEN is not available. Run the token cell first.\")\n",
+    "\n",
+    "def human_bytes(n: int) -> str:\n",
+    "    if n is None:\n",
+    "        return \"?\"\n",
+    "    units = ['B','KB','MB','GB','TB']\n",
+    "    i = 0\n",
+    "    f = float(n)\n",
+    "    while f >= 1024 and i < len(units)-1:\n",
+    "        f /= 1024.0\n",
+    "        i += 1\n",
+    "    return f\"{f:.2f} {units[i]}\"\n",
+    "\n",
+    "fs = HfFileSystem(token=HF_TOKEN)\n",
+    "repo_url = f\"https://huggingface.co/datasets/{REPO_ID}\"\n",
+    "root = f\"hf://datasets/{REPO_ID}\"\n",
+    "print(f\"Scanning repo: {repo_url}\")\n",
+    "\n",
+    "# Try to invalidate any local cache to avoid stale listings\n",
+    "if hasattr(fs, \"invalidate_cache\"):\n",
+    "    try:\n",
+    "        fs.invalidate_cache()\n",
+    "        # Also try root-specific invalidation if supported\n",
+    "        try:\n",
+    "            fs.invalidate_cache(root)\n",
+    "        except Exception:\n",
+    "            pass\n",
+    "        print(\"Cache invalidated for HfFileSystem.\")\n",
+    "    except Exception as _e:\n",
+    "        print(\"Cache invalidation skipped:\", type(_e).__name__)\n",
+    "\n",
+    "# List all files recursively\n",
+    "t0 = time.time()\n",
+    "paths = fs.find(root)  # returns file paths recursively\n",
+    "if not isinstance(paths, list):\n",
+    "    paths = list(paths)\n",
+    "elapsed = time.time() - t0\n",
+    "print(f\"Found {len(paths):,} candidate paths in {elapsed:.2f}s\")\n",
+    "\n",
+    "# Fallback cross-check with API listing to detect propagation delays\n",
+    "try:\n",
+    "    api_files = api.list_repo_files(repo_id=REPO_ID, repo_type=REPO_TYPE)\n",
+    "    print(f\"API list_repo_files reports: {len(api_files):,} files\")\n",
+    "    if len(paths) < len(api_files):\n",
+    "        print(\"Note: HfFileSystem may be stale; UI/CDN can lag for ~1-2 minutes.\")\n",
+    "except Exception as _e:\n",
+    "    print(\"API list_repo_files failed:\", type(_e).__name__)\n",
+    "\n",
+    "# Fetch sizes concurrently\n",
+    "def _info(pth: str):\n",
+    "    try:\n",
+    "        info = fs.info(pth)\n",
+    "        if info.get('type') == 'file':\n",
+    "            return pth, int(info.get('size') or 0)\n",
+    "    except Exception:\n",
+    "        return pth, None\n",
+    "    return pth, None\n",
+    "\n",
+    "sizes = {}\n",
+    "errors = 0\n",
+    "t0 = time.time()\n",
+    "with ThreadPoolExecutor(max_workers=24) as ex:\n",
+    "    futs = [ex.submit(_info, p) for p in paths]\n",
+    "    for fut in as_completed(futs):\n",
+    "        p, sz = fut.result()\n",
+    "        if sz is None:\n",
+    "            errors += 1\n",
+    "        else:\n",
+    "            sizes[p] = sz\n",
+    "elapsed = time.time() - t0\n",
+    "print(f\"Fetched metadata for {len(sizes):,} files ({errors} errors) in {elapsed:.2f}s\")\n",
+    "\n",
+    "# Summaries\n",
+    "total_files = len(sizes)\n",
+    "total_bytes = sum(sizes.values())\n",
+    "by_ext_count = Counter()\n",
+    "by_ext_bytes = defaultdict(int)\n",
+    "for p, sz in sizes.items():\n",
+    "    ext = os.path.splitext(p)[1].lower() or \"\"\n",
+    "    by_ext_count[ext] += 1\n",
+    "    by_ext_bytes[ext] += sz\n",
+    "\n",
+    "print(\"\\n=== Summary ===\")\n",
+    "print(f\"Total files: {total_files:,}\")\n",
+    "print(f\"Total size : {human_bytes(total_bytes)}\")\n",
+    "\n",
+    "# Top extensions by bytes\n",
+    "top_ext = sorted(by_ext_bytes.items(), key=lambda x: x[1], reverse=True)[:10]\n",
+    "print(\"\\nTop extensions by size:\")\n",
+    "for ext, b in top_ext:\n",
+    "    print(f\"  {ext or '(no ext)'}: {by_ext_count[ext]:,} files, {human_bytes(b)}\")\n",
+    "\n",
+    "# Top-level folder distribution\n",
+    "top_level = Counter()\n",
+    "for p in sizes.keys():\n",
+    "    rel = p.replace(root.rstrip('/'), '').lstrip('/')\n",
+    "    parts = rel.split('/')\n",
+    "    top = parts[0] if parts and parts[0] else '/'\n",
+    "    top_level[top] += 1\n",
+    "print(\"\\nTop-level folder counts:\")\n",
+    "for name, cnt in top_level.most_common(15):\n",
+    "    print(f\"  {name}: {cnt:,} files\")\n",
+    "\n",
+    "# Largest files\n",
+    "top_k = 20\n",
+    "largest = sorted(sizes.items(), key=lambda x: x[1], reverse=True)[:top_k]\n",
+    "print(f\"\\nTop {top_k} largest files:\")\n",
+    "for p, sz in largest:\n",
+    "    rel = p.replace(root.rstrip('/'), '').lstrip('/')\n",
+    "    print(f\"  {human_bytes(sz)}  |  {rel}\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 30,
+   "id": "beb35ad2",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Repo: Smilesjs/Haptix_dataset type: dataset\n",
+      "list_repo_files count: 3\n",
+      "First 10: ['.gitattributes', 'augmented_images/E3P-Harmonic_P3Rough_P2Cold_P5Static/E3P+P3R+P2Co+P5St_aug_001.png', 'test_probe.txt']\n"
+     ]
+    },
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "No files have been modified since last commit. Skipping to prevent empty commit.\n",
+      "WARNING:huggingface_hub.hf_api:No files have been modified since last commit. Skipping to prevent empty commit.\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Uploaded marker: test_probe.txt\n",
+      "list_repo_files after: 3\n",
+      "Contains marker? True\n",
+      "\n",
+      "Recent commits:\n",
+      "- 8e57433 sample image upload\n",
+      "- 6a8d3d3 probe upload\n",
+      "- 16f395e initial commit\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Probe: list files and upload a tiny marker file to confirm branch\n",
+    "from io import BytesIO\n",
+    "print(\"Repo:\", REPO_ID, \"type:\", REPO_TYPE)\n",
+    "files = api.list_repo_files(repo_id=REPO_ID, repo_type=REPO_TYPE)\n",
+    "print(\"list_repo_files count:\", len(files))\n",
+    "print(\"First 10:\", files[:10])\n",
+    "\n",
+    "marker_path = \"test_probe.txt\"\n",
+    "api.upload_file(path_or_fileobj=BytesIO(b\"probe\"), path_in_repo=marker_path, repo_id=REPO_ID, repo_type=REPO_TYPE, commit_message=\"probe upload\")\n",
+    "print(\"Uploaded marker:\", marker_path)\n",
+    "files2 = api.list_repo_files(repo_id=REPO_ID, repo_type=REPO_TYPE)\n",
+    "print(\"list_repo_files after:\", len(files2))\n",
+    "print(\"Contains marker?\", marker_path in files2)\n",
+    "\n",
+    "# Show latest commits\n",
+    "commits = api.list_repo_commits(repo_id=REPO_ID, repo_type=REPO_TYPE)\n",
+    "print(\"\\nRecent commits:\")\n",
+    "for c in commits[:5]:\n",
+    "    print(\"-\", c.commit_id[:7], c.title)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 28,
+   "id": "dbc2c9de",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Uploading sample: C:\\Users\\EL081\\Desktop\\local_backup\\image\\augmented_images\\E3P-Harmonic_P3Rough_P2Cold_P5Static\\E3P+P3R+P2Co+P5St_aug_001.png\n"
+     ]
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "8b0ce6dca03e4dcfbd235e3e9c1573e7",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Processing Files (0 / 0): |          |  0.00B /  0.00B            "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "fb3e8bf2c1b5438aaf6334a4333e0c8a",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "New Data Upload: |          |  0.00B /  0.00B            "
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Uploaded to: augmented_images/E3P-Harmonic_P3Rough_P2Cold_P5Static/E3P+P3R+P2Co+P5St_aug_001.png\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Probe: upload one image file to confirm dataset accepts content\n",
+    "import os\n",
+    "from pathlib import Path\n",
+    "from itertools import chain\n",
+    "exts = {\".jpg\",\".jpeg\",\".png\"}\n",
+    "sample_path = None\n",
+    "for root_dir, dirs, files in os.walk(LOCAL_FOLDER):\n",
+    "    for fn in files:\n",
+    "        if Path(fn).suffix.lower() in exts:\n",
+    "            sample_path = Path(root_dir) / fn\n",
+    "            break\n",
+    "    if sample_path:\n",
+    "        break\n",
+    "if not sample_path:\n",
+    "    raise RuntimeError(\"No image file found under LOCAL_FOLDER.\")\n",
+    "rel = sample_path.relative_to(LOCAL_FOLDER).as_posix()\n",
+    "dest = rel  # or f\"image/{rel}\" to nest under image/\n",
+    "print(\"Uploading sample:\", sample_path)\n",
+    "api.upload_file(path_or_fileobj=str(sample_path), path_in_repo=dest, repo_id=REPO_ID, repo_type=REPO_TYPE, commit_message=\"sample image upload\")\n",
+    "print(\"Uploaded to:\", dest)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 31,
+   "id": "5a5057d4",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "No .hf_transfer metadata directories found under LOCAL_FOLDER.\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Optional: clear hf_transfer cached metadata under LOCAL_FOLDER to avoid stale plans\n",
+    "import os, shutil\n",
+    "found = []\n",
+    "for root_dir, dirs, files in os.walk(LOCAL_FOLDER):\n",
+    "    if \".hf_transfer\" in dirs:\n",
+    "        found.append(os.path.join(root_dir, \".hf_transfer\"))\n",
+    "if found:\n",
+    "    print(\"Found hf_transfer metadata dirs:\")\n",
+    "    for d in found:\n",
+    "        print(\" -\", d)\n",
+    "    # Uncomment next lines to clear the cached metadata that may point to another repo\n",
+    "    # for d in found:\n",
+    "    #     shutil.rmtree(d, ignore_errors=True)\n",
+    "    # print(\"Cleared hf_transfer metadata.\")\n",
+    "else:\n",
+    "    print(\"No .hf_transfer metadata directories found under LOCAL_FOLDER.\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 32,
+   "id": "1668f9a0",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "upload_large_folder signature: (repo_id: 'str', folder_path: 'Union[str, Path]', *, repo_type: 'str', revision: 'Optional[str]' = None, private: 'Optional[bool]' = None, allow_patterns: 'Optional[Union[list[str], str]]' = None, ignore_patterns: 'Optional[Union[list[str], str]]' = None, num_workers: 'Optional[int]' = None, print_report: 'bool' = True, print_report_every: 'int' = 60) -> 'None'\n",
+      "\n",
+      "Doc (first 40 lines):\n",
+      " Upload a large folder to the Hub in the most resilient way possible.\n",
+      "\n",
+      "Several workers are started to upload files in an optimized way. Before being committed to a repo, files must be\n",
+      "hashed and be pre-uploaded if they are LFS files. Workers will perform these tasks for each file in the folder.\n",
+      "At each step, some metadata information about the upload process is saved in the folder under `.cache/.huggingface/`\n",
+      "to be able to resume the process if interrupted. The whole process might result in several commits.\n",
+      "\n",
+      "Args:\n",
+      "    repo_id (`str`):\n",
+      "        The repository to which the file will be uploaded.\n",
+      "        E.g. `\"HuggingFaceTB/smollm-corpus\"`.\n",
+      "    folder_path (`str` or `Path`):\n",
+      "        Path to the folder to upload on the local file system.\n",
+      "    repo_type (`str`):\n",
+      "        Type of the repository. Must be one of `\"model\"`, `\"dataset\"` or `\"space\"`.\n",
+      "        Unlike in all other `HfApi` methods, `repo_type` is explicitly required here. This is to avoid\n",
+      "        any mistake when uploading a large folder to the Hub, and therefore prevent from having to re-upload\n",
+      "        everything.\n",
+      "    revision (`str`, `optional`):\n",
+      "        The branch to commit to. If not provided, the `main` branch will be used.\n",
+      "    private (`bool`, `optional`):\n",
+      "        Whether the repository should be private.\n",
+      "        If `None` (default), the repo will be public unless the organization's default is private.\n",
+      "    allow_patterns (`list[str]` or `str`, *optional*):\n",
+      "        If provided, only files matching at least one pattern are uploaded.\n",
+      "    ignore_patterns (`list[str]` or `str`, *optional*):\n",
+      "        If provided, files matching any of the patterns are not uploaded.\n",
+      "    num_workers (`int`, *optional*):\n",
+      "        Number of workers to start. Defaults to `os.cpu_count() - 2` (minimum 2).\n",
+      "        A higher number of workers may speed up the process if your machine allows it. However, on machines with a\n",
+      "        slower connection, it is recommended to keep the number of workers low to ensure better resumability.\n",
+      "        Indeed, partially uploaded files will have to be completely re-uploaded if the process is interrupted.\n",
+      "    print_report (`bool`, *optional*):\n",
+      "        Whether to print a report of the upload progress. Defaults to True.\n",
+      "        Report is printed to `sys.stdout` every X seconds (60 by defaults) and overwrites the previous report.\n",
+      "    print_report_every (`int`, *optional*):\n",
+      "        Frequency at which the report is printed. Defaults to 60 seconds.\n",
+      "\n",
+      "> [!TIP]\n",
+      "> A few things to keep in mind:\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Inspect upload_large_folder signature\n",
+    "import inspect, textwrap\n",
+    "sig = inspect.signature(api.upload_large_folder)\n",
+    "print(\"upload_large_folder signature:\", sig)\n",
+    "doc = inspect.getdoc(api.upload_large_folder)\n",
+    "print(\"\\nDoc (first 40 lines):\\n\", \"\\n\".join(doc.splitlines()[:40]))"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.8"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

check.ipynb

@@ -0,0 +1,219 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "afc7ef8c",
+   "metadata": {},
+   "source": [
+    "# Image Folder Analysis\n",
+    "이 노트북은 `image` 폴더 내의 모든 이미지 파일 수를 분석합니다."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "7a5442aa",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "============================================================\n",
+      "이미지 폴더 분석 결과\n",
+      "============================================================\n",
+      "\n",
+      "전체 파일 수: 13,841\n",
+      "이미지 파일 수: 6,918\n",
+      "기타 파일 수: 6,923\n",
+      "\n",
+      "============================================================\n",
+      "메인 폴더별 이미지 수\n",
+      "============================================================\n"
+     ]
+    }
+   ],
+   "source": [
+    "import os\n",
+    "import pandas as pd\n",
+    "from pathlib import Path\n",
+    "from collections import defaultdict\n",
+    "\n",
+    "# 이미지 폴더 경로\n",
+    "image_folder = r\"C:\\Users\\EL081\\Desktop\\local_backup\\image\"\n",
+    "\n",
+    "# 이미지 파일 확장자\n",
+    "image_extensions = {'.jpg', '.jpeg', '.png', '.gif', '.bmp', '.webp', '.tiff', '.tif', '.ico', '.svg'}\n",
+    "\n",
+    "# 전체 파일 수 계산\n",
+    "total_files = 0\n",
+    "total_images = 0\n",
+    "folder_stats = defaultdict(lambda: {'total': 0, 'images': 0})\n",
+    "\n",
+    "for root, dirs, files in os.walk(image_folder):\n",
+    "    for file in files:\n",
+    "        total_files += 1\n",
+    "        ext = Path(file).suffix.lower()\n",
+    "        if ext in image_extensions:\n",
+    "            total_images += 1\n",
+    "        \n",
+    "        # 메인 폴더별 통계\n",
+    "        relative_path = os.path.relpath(root, image_folder)\n",
+    "        main_folder = relative_path.split(os.sep)[0]\n",
+    "        folder_stats[main_folder]['total'] += 1\n",
+    "        if ext in image_extensions:\n",
+    "            folder_stats[main_folder]['images'] += 1\n",
+    "\n",
+    "print(\"=\" * 60)\n",
+    "print(\"이미지 폴더 분석 결과\")\n",
+    "print(\"=\" * 60)\n",
+    "print(f\"\\n전체 파일 수: {total_files:,}\")\n",
+    "print(f\"이미지 파일 수: {total_images:,}\")\n",
+    "print(f\"기타 파일 수: {total_files - total_images:,}\")\n",
+    "print(\"\\n\" + \"=\" * 60)\n",
+    "print(\"메인 폴더별 이미지 수\")\n",
+    "print(\"=\" * 60)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "be675b07",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      ".cache                                   이미지:      0   전체:  6,921\n",
+      "EmoSet_images                            이미지:  2,888   전체:  2,888\n",
+      "Midjourney_images                        이미지:     17   전체:     17\n",
+      "augmented_images                         이미지:  1,313   전체:  1,315\n",
+      "generated_images                         이미지:    795   전체:    795\n",
+      "unsplash_images                          이미지:  1,905   전체:  1,905\n",
+      "\n",
+      "============================================================\n",
+      "메인 폴더별 이미지 수 (상세)\n",
+      "============================================================\n",
+      "              폴더명  이미지수  전체파일수\n",
+      "    EmoSet_images  2888   2888\n",
+      "  unsplash_images  1905   1905\n",
+      " augmented_images  1313   1315\n",
+      " generated_images   795    795\n",
+      "Midjourney_images    17     17\n",
+      "           .cache     0   6921\n"
+     ]
+    }
+   ],
+   "source": [
+    "# 메인 폴더별 상세 통계\n",
+    "for folder in sorted(folder_stats.keys()):\n",
+    "    stats = folder_stats[folder]\n",
+    "    print(f\"{folder:<40} 이미지: {stats['images']:>6,}   전체: {stats['total']:>6,}\")\n",
+    "\n",
+    "# 데이터프레임으로 정렬된 결과 표시\n",
+    "df_stats = pd.DataFrame([\n",
+    "    {'폴더명': folder, '이미지수': stats['images'], '전체파일수': stats['total']}\n",
+    "    for folder, stats in folder_stats.items()\n",
+    "]).sort_values('이미지수', ascending=False)\n",
+    "\n",
+    "print(\"\\n\" + \"=\" * 60)\n",
+    "print(\"메인 폴더별 이미지 수 (상세)\")\n",
+    "print(\"=\" * 60)\n",
+    "print(df_stats.to_string(index=False))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "84374e6a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 파일 타입별 분석\n",
+    "extension_count = defaultdict(int)\n",
+    "\n",
+    "for root, dirs, files in os.walk(image_folder):\n",
+    "    for file in files:\n",
+    "        ext = Path(file).suffix.lower()\n",
+    "        if ext:\n",
+    "            extension_count[ext] += 1\n",
+    "\n",
+    "print(\"\\n\" + \"=\" * 60)\n",
+    "print(\"파일 형식별 수\")\n",
+    "print(\"=\" * 60)\n",
+    "\n",
+    "df_ext = pd.DataFrame([\n",
+    "    {'파일형식': ext, '수': count}\n",
+    "    for ext, count in extension_count.items()\n",
+    "]).sort_values('수', ascending=False)\n",
+    "\n",
+    "print(df_ext.to_string(index=False))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "d37eae36",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 시각화\n",
+    "import matplotlib.pyplot as plt\n",
+    "plt.figure(figsize=(12, 6))\n",
+    "\n",
+    "# 1. 메인 폴더별 이미지 수 차트\n",
+    "plt.subplot(1, 2, 1)\n",
+    "df_sorted = df_stats.sort_values('이미지수', ascending=True)\n",
+    "plt.barh(df_sorted['폴더명'], df_sorted['이미지수'], color='steelblue')\n",
+    "plt.xlabel('이미지 수')\n",
+    "plt.title('메인 폴더별 이미지 수')\n",
+    "plt.grid(axis='x', alpha=0.3)\n",
+    "\n",
+    "# 2. 상위 10개 파일 형식 차트\n",
+    "plt.subplot(1, 2, 2)\n",
+    "df_ext_top = df_ext.head(10)\n",
+    "plt.bar(df_ext_top['파일형식'], df_ext_top['수'], color='coral')\n",
+    "plt.xlabel('파일 형식')\n",
+    "plt.ylabel('수')\n",
+    "plt.title('상위 10개 파일 형식')\n",
+    "plt.xticks(rotation=45)\n",
+    "plt.grid(axis='y', alpha=0.3)\n",
+    "\n",
+    "plt.tight_layout()\n",
+    "plt.show()\n",
+    "\n",
+    "print(\"\\n✓ 분석 완료!\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "b91d4a75",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.8"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}

hug.ipynb

@@ -0,0 +1,79 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "ae64f9ce",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from huggingface_hub import HfApi"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "167062d2",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "application/vnd.jupyter.widget-view+json": {
+       "model_id": "2f64faaf886b47e483da57cac528c88f",
+       "version_major": 2,
+       "version_minor": 0
+      },
+      "text/plain": [
+       "Recovering from metadata files:   0%|          | 0/6920 [00:00<?, ?it/s]"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "\n",
+      "\n",
+      "\n",
+      "---------- 2025-11-14 21:22:14 (0:00:00) ----------\n",
+      "Files:   hashed 6920/6920 (1.2G/1.2G) | pre-uploaded: 6918/6918 (1.2G/1.2G) | committed: 6920/6920 (1.2G/1.2G) | ignored: 0\n",
+      "Workers: hashing: 0 | get upload mode: 0 | pre-uploading: 0 | committing: 0 | waiting: 0\n",
+      "---------------------------------------------------\n"
+     ]
+    }
+   ],
+   "source": [
+    "api = HfApi()\n",
+    "\n",
+    "HfApi().upload_large_folder(\n",
+    "    folder_path=\"C:/Users/EL081/Desktop/local_backup/image\",\n",
+    "    repo_id=\"Smilesjs/Haptix_dataset\",\n",
+    "    repo_type=\"dataset\",\n",
+    ")"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.8"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}