{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.11.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"gpu","dataSources":[{"sourceId":71549,"databundleVersionId":8561470,"sourceType":"competition"},{"sourceId":25568107,"sourceType":"kernelVersion"},{"sourceId":107247349,"sourceType":"kernelVersion"},{"sourceId":193161758,"sourceType":"kernelVersion"},{"sourceId":194501104,"sourceType":"kernelVersion"},{"sourceId":194526416,"sourceType":"kernelVersion"}],"dockerImageVersionId":31090,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import os\nimport pandas as pd\nimport numpy as np\nimport pydicom\nfrom tqdm.auto import tqdm\nimport matplotlib.pyplot as plt\nimport cv2\nimport glob\n\n!pip install iterative-stratification\n!pip install pgzip\n!pip install git+https://github.com/ultralytics/ultralytics.git@main\n!unzip -q /kaggle/input/lsdc-gen-yolo-data-scs/data_fold0.zip\nimport iterstrat\n#ls\n\n#data_path = \"/kaggle/input/rsna-2024-lumbar-spine-degenerative-classification/\"\n\ndf = pd.read_csv('/kaggle/input/rsna-2024-lumbar-spine-degenerative-classification/train.csv')\n#df = train_data\ndf = df.fillna('Unknown')\nfrom iterstrat.ml_stratifiers import MultilabelStratifiedKFold\nmskf = MultilabelStratifiedKFold(n_splits=5, shuffle=True, random_state=0)\n\nfold = 0\nfor train_index, test_index in mskf.split(df, df.iloc[:,1:]):\n    df.loc[test_index, 'fold'] = fold\n    fold += 1\n\ndf['fold'] = df['fold'].astype(int)\ndf[['study_id', 'fold']].to_csv('5folds.csv', index=False)\n\n\nROOT_DIR=  \"/kaggle/input/rsna-2024-lumbar-spine-degenerative-classification/\"\nIMG_DIR = \"/kaggle/input/rsna-2024-lumbar-spine-degenerative-classification/train_images\"\nFOLD = 0\nOD_INPUT_SIZE = 384\nSTD_BOX_SIZE = 20\nBATCH_SIZE = 32\nEPOCHS = 75\n\nSAMPLE = None\nCONDITIONS = ['Left Subarticular Stenosis', 'Right Subarticular Stenosis']\nSEVERITIES = ['Normal/Mild', 'Moderate', 'Severe']\nLEVELS = ['l1_l2', 'l2_l3', 'l3_l4', 'l4_l5', 'l5_s1']\n\nDATA_DIR = f'data_fold{FOLD}'\n\ntrain_val_df = pd.read_csv('/kaggle/input/rsna-2024-lumbar-spine-degenerative-classification/train.csv')\ntrain_xy = pd.read_csv('/kaggle/input/rsna-2024-lumbar-spine-degenerative-classification/train_label_coordinates.csv')\ntrain_des = pd.read_csv('/kaggle/input/rsna-2024-lumbar-spine-degenerative-classification/train_series_descriptions.csv')\n\nif SAMPLE:\n    train_val_df = train_val_df.sample(SAMPLE, random_state=2698)\n\nfold_df = pd.read_csv('5folds.csv')\ntest_df = fold_df[fold_df.fold == FOLD]\n\ntrain_xy.head(3)\n\ndef get_level(text):\n    for lev in ['l1_l2', 'l2_l3', 'l3_l4', 'l4_l5', 'l5_s1']:\n        if lev in text:\n            split = lev.split('_')\n            split[0] = split[0].capitalize()\n            split[1] = split[1].capitalize()\n            return '/'.join(split)\n    raise ValueError('Level not found '+ lev)\n    \ndef get_condition(text):\n    split = text.split('_')\n    for i in range(len(split)):\n        split[i] = split[i].capitalize()\n    split = split[:-2]\n    return ' '.join(split)\n#     raise ValueError('Condition not found '+ lev)\ntrain_xy['condition'].unique()\n\n#train_df = train_val_df.dropna()\nlabel_df = {'study_id':[], 'condition': [], 'level':[], 'label':[]}\n\nfor i, row in train_val_df.iterrows():\n    study_id = row['study_id']\n    for k, label in row.iloc[1:].to_dict().items():\n        level = get_level(k)\n        condition = get_condition(k)\n        label_df['study_id'].append(study_id)\n        label_df['condition'].append(condition)\n        label_df['level'].append(level)\n        label_df['label'].append(label)\n#         break\n#     break\n\nlabel_df = pd.DataFrame(label_df)\nlabel_df = label_df.merge(fold_df, on='study_id')\ntrain_xy = train_xy.merge(train_des, how='inner', on=['study_id', 'series_id'])\nlabel_df = label_df.merge(train_xy, how='inner', on=['study_id', 'condition', 'level'])\ndef query_train_xy_row(study_id, series_id=None, instance_num=None):\n    if series_id is not None and instance_num is not None:\n        return label_df[(label_df.study_id==study_id) & (label_df.series_id==series_id) &\n            (label_df.instance_number==instance_num)]\n    elif series_id is None and instance_num is None:\n        return label_df[(label_df.study_id==study_id)]\n    else:\n        return label_df[(train_xy.study_id==study_id) & (label_df.series_id==series_id)]\ndef read_dcm(src_path):\n    dicom_data = pydicom.dcmread(src_path)\n    image = dicom_data.pixel_array\n    image = (image - image.min()) / (image.max() - image.min() +1e-6) * 255\n    image = np.stack([image]*3, axis=-1).astype('uint8')\n    return image\n\ndef get_accronym(text):\n    split = text.split(' ')\n    return ''.join([x[0] for x in split])\n# study_id = 4003253 \n# series_id = 2448190387\n# instance_num = 28\n\nex = label_df.sample(1).iloc[0]\nstudy_id = ex.study_id\nseries_id = ex.series_id\ninstance_num = ex.instance_number\n\nWIDTH = 10\n\npath = os.path.join(IMG_DIR, str(study_id), str(series_id), f'{instance_num}.dcm')\nimg = read_dcm(path)\n\ntmp_df = query_train_xy_row(study_id, series_id, instance_num)\nfor i, row in tmp_df.iterrows():\n    lbl = f\"{get_accronym(row['condition'])}_{row['level']}\"\n    x, y = row['x'], row['y']\n    x1 = int(x - WIDTH)\n    x2 = int(x + WIDTH)\n    y1 = int(y - WIDTH)\n    y2 = int(y + WIDTH)\n    color = None\n    if row['label'] == 'Normal/Mild':\n        color =  (0, 255, 0)\n    elif row['label'] == 'Moderate':\n        color = (255,255,0) \n    elif row['label'] == 'Severe':\n        color = (255,0,0)\n        \n    fontFace = cv2.FONT_HERSHEY_SIMPLEX\n    fontScale = 0.5\n    thickness = 1\n    cv2.rectangle(img, (x1,y1), (x2,y2), color, 2)\n    cv2.putText(img, lbl, (x1,y1), fontFace, fontScale, color, thickness, cv2.LINE_AA)\n\ntmp_df\n\nplt.imshow(img)\nplt.show()\n\n# label_df[['study_id', 'series_id']].drop_duplicates()\ndef read_dcm(src_path):\n    dicom_data = pydicom.dcmread(src_path)\n    image = dicom_data.pixel_array\n    image = (image - image.min()) / (image.max() - image.min() +1e-6) * 255\n    image = np.stack([image]*3, axis=-1).astype('uint8')\n    return image\nfiltered_df = label_df[label_df.condition.map(lambda x: x in CONDITIONS)]\nlabel2id = {}\nid2label = {}\ni = 0\nfor cond in CONDITIONS:\n    for level in LEVELS:\n        for severity in SEVERITIES:\n            cls_ = f\"{cond.lower().replace(' ', '_')}_{level}_{severity.lower()}\"\n            label2id[cls_] = i\n            id2label[i] = cls_\n            i+=1\nid2label\n\ntrain_df = filtered_df[filtered_df.fold != FOLD]\nval_df = filtered_df[filtered_df.fold == FOLD]\n\n_IM_DIR = f'{DATA_DIR}/images/train'\n_ANN_DIR = f'{DATA_DIR}/labels/train'\nname = np.random.choice(os.listdir(_IM_DIR))[:-4]\n\nim = plt.imread(os.path.join(_IM_DIR, name+'.jpg')).copy()\nH,W = im.shape[:2]\nanns = np.loadtxt(os.path.join(_ANN_DIR, name+'.txt')).reshape(-1, 5)\n\nfor _cls, x,y,w,h in anns.tolist():\n    x *= W\n    y *= H\n    w *= W\n    h *= H\n    x1 = int(x-w/2)\n    x2 = int(x+w/2)\n    y1 = int(y-h/2)\n    y2 = int(y+h/2)\n    label = id2label[_cls]\n    \n#     if _cls == 0:\n#         c = (255,0,0)\n#     elif _cls == 1:\n#         c = (0,255,0)\n#     else:\n#         c = (255,255,0)\n    c = (0,255,255)\n\n    im = cv2.rectangle(im, (x1,y1), (x2,y2), c, 2)\n    cv2.putText(im, label, (x1,y1), fontFace, 0.3, c, 1, cv2.LINE_AA)\n\n\nplt.imshow(im)\n\n# ls data_fold0/labels/val\n# os.path.join(_ANN_DIR, name+'.txt')\n# cat 'train_fold0/labels/404602713_1230697721_12.txt'\n# Install the ultralytics package from GitHub\n!pip install git+https://github.com/ultralytics/ultralytics.git@main\n\nfor k, v in id2label.items():\n    print(f'{k}: {v}')\n\n#ls\n\n\n#%%writefile yolo_ss.yaml\n#path: \"/kaggle/working/data_fold0\" # dataset root dir\n#train: \"images/train\"  \n#val: \"images/val\" \n#test: \"images/val\" \nwith open(\"yolo_ss.yaml\", \"w\") as f:\n    f.write('path: \"/kaggle/working/data_fold0\"\\n')\n    f.write('train: \"images/train\"\\n')\n    f.write('val: \"images/val\"\\n')\n    f.write('test: \"images/val\"\\n')\n    f.write(\"names:\\n\")\n    f.write(\"    0: spinal_canal_stenosis_l1_l2_normal/mild\\n\")\n    f.write(\"    1: spinal_canal_stenosis_l1_l2_moderate\\n\")\n    f.write(\"    2: spinal_canal_stenosis_l1_l2_severe\\n\")\n    f.write(\"    3: spinal_canal_stenosis_l2_l3_normal/mild\\n\")\n    f.write(\"    4: spinal_canal_stenosis_l2_l3_moderate\\n\")\n    f.write(\"    5: spinal_canal_stenosis_l2_l3_severe\\n\")\n    f.write(\"    6: spinal_canal_stenosis_l3_l4_normal/mild\\n\")\n    f.write(\"    7: spinal_canal_stenosis_l3_l4_moderate\\n\")\n    f.write(\"    8: spinal_canal_stenosis_l3_l4_severe\\n\")\n    f.write(\"    9: spinal_canal_stenosis_l4_l5_normal/mild\\n\")\n    f.write(\"    10: spinal_canal_stenosis_l4_l5_moderate\\n\")\n    f.write(\"    11: spinal_canal_stenosis_l4_l5_severe\\n\")\n    f.write(\"    12: spinal_canal_stenosis_l5_s1_normal/mild\\n\")\n    f.write(\"    13: spinal_canal_stenosis_l5_s1_moderate\\n\")\n    f.write(\"    14: spinal_canal_stenosis_l5_s1_severe\\n\")\n   \nwith open('yolo_ss.yaml', 'r') as f:\n    print(f.read())\n\nimport wandb\nfrom wandb.integration.ultralytics import add_wandb_callback\n\nfrom kaggle_secrets import UserSecretsClient\nuser_secrets = UserSecretsClient()\nsecret_value_0 = user_secrets.get_secret(\"wandbkey\")\nwandb.login(key=secret_value_0)\n\n\n\n# Initialize W&B run\nwandb.init(\n    project=\"lsdc_yolov8scs\",\n#     name=f\"Demo_fold0\",\n#     tags=[\"baseline\", \"search-lr\", ],\n    group=\";\".join(CONDITIONS),\n#     config={\n#         \"lr\": LR,\n#         \"model-name\":\"xtremedistill-trim\",\n#         \"dataset\": [\n#             \"raw_compettion\",\n#             \"MPWare\",\n#             \"Nicholas\"\n#         ]\n#     }\n)\n\nfrom ultralytics import YOLO\n\n\n# Initialize YOLO Model\nmodel = YOLO(\"yolov8s.pt\")\n\n# Add W&B callback for Ultralytics\nadd_wandb_callback(model, enable_model_checkpointing=True)\n\n# Train/fine-tune your model\n# At the end of each epoch, predictions on validation batches are logged\n# to a W&B table with insightful and interactive overlays for\n# computer vision tasks\nmodel.train(project=\"lsdc_yolov8scs\", data=\"yolo_ss.yaml\", \n            epochs=EPOCHS, imgsz=OD_INPUT_SIZE, batch=BATCH_SIZE)\n\n# Finish the W&B run\nwandb.finish()\n\nfrom collections import defaultdict\n# # test generated annotations\n\n_IM_DIR = f'{DATA_DIR}/images/val'\n_ANN_DIR = f'{DATA_DIR}/labels/val'\nname = np.random.choice(os.listdir(_IM_DIR))[:-4]\n\npath = os.path.join(_IM_DIR, name+'.jpg')\n\nim = plt.imread(path).copy()\nH,W = im.shape[:2]\nanns = np.loadtxt(os.path.join(_ANN_DIR, name+'.txt')).reshape(-1, 5)\n\nfor _cls, x,y,w,h in anns.tolist():\n    x *= W\n    y *= H\n    w *= W\n    h *= H\n    x1 = int(x-w/2)\n    x2 = int(x+w/2)\n    y1 = int(y-h/2)\n    y2 = int(y+h/2)\n    label = id2label[_cls]\n    print(label)\n    \n#     if _cls == 0:\n#         c = (255,0,0)\n#     elif _cls == 1:\n#         c = (0,255,0)\n#     else:\n#         c = (255,255,0)\n    c = (0,255,255)\n\n    im = cv2.rectangle(im, (x1,y1), (x2,y2), c, 2)\n    cv2.putText(im, label, (x1,y1), fontFace, 0.3, c, 1, cv2.LINE_AA)\n\n\nplt.imshow(im)\n\n\n# Initialize YOLO Model\nmodel = YOLO(glob.glob(\"lsdc_yolov8scs/*/weights/best.pt\")[0])\n\n# Add W&B callback for Ultralytics\n# add_wandb_callback(model, enable_model_checkpointing=True)\n\n# Perform prediction which automatically logs to a W&B Table\n# with interactive overlays for bounding boxes, segmentation masks\nout = model.predict([path], save=True, conf=0.2)\n\n# Finish the W&B run\nwandb.finish()\n\nim = plt.imread(glob.glob(f'{out[0].save_dir}/*.jpg')[0])\nplt.imshow(im)\n","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"execution":{"iopub.status.busy":"2025-08-30T13:40:14.032669Z","iopub.execute_input":"2025-08-30T13:40:14.032867Z"}},"outputs":[],"execution_count":null}]}