{"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":"none","dataSources":[{"sourceId":96164,"databundleVersionId":12993472,"sourceType":"competition"},{"sourceId":291790313,"sourceType":"kernelVersion"},{"sourceId":291791094,"sourceType":"kernelVersion"},{"sourceId":291826676,"sourceType":"kernelVersion"}],"dockerImageVersionId":31234,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import pandas as pd\ndef iBlend(path_to_ds, file_short_names, sls):\n\n    def tida(sls):\n        \n        def read_subm(sls,i):\n            tnm = sls[\"subm\"][i][\"name\"]\n            FiN = sls[\"path\"] + tnm + \".csv\"\n            return pd.read_csv(FiN).rename(columns={'target':tnm, sls[\"target\"]:tnm})\n        \n        dfs_subm = [read_subm(sls,i) for i in range(len(sls[\"subm\"]))]\n        df_subms = pd.merge(dfs_subm[0],  dfs_subm[1], on=['ID'])\n        \n        for i in range(2, len(sls[\"subm\"])): \n            df_subms = pd.merge(df_subms, dfs_subm[i], on=['ID'])\n            \n        cols = [col for col in df_subms.columns if col != \"ID\"]\n        short_name_cols = [c.replace(sls[\"prefix\"], '') for c in cols]\n        corrects = [wt for wt in sls[\"subwts\"]]\n        weights = [subm['weight'] for subm in sls[\"subm\"]]\n        \n        def alls(x, cs=cols):\n            tes = {c: x[c] for c in cs}.items()\n            subms_sorted = [\n              t[0].replace(sls[\"prefix\"], '')\n              for t in sorted(tes,key=lambda k:k[1],reverse=True if sls[\"sort\"]=='desc' else False)]\n            return subms_sorted\n        \n        def correct(x, cs=cols, w=weights, cw=corrects):\n            ic = [x['alls'].index(c) for c in short_name_cols]\n            cS = [x[cols[j]] * (w[j] + cw[ic[j]]) for j in range(len(cols))]\n            return sum(cS)\n        \n        df_subms['alls']        = df_subms.apply(lambda x: alls   (x), axis=1)\n        df_subms[sls[\"target\"]] = df_subms.apply(lambda x: correct(x), axis=1)\n        \n        schema_rename = { old_nc:new_shnc for old_nc, new_shnc in zip(cols, short_name_cols) }\n        \n        df_subms = df_subms.rename(columns=schema_rename)\n        df_subms = df_subms.rename(columns={sls[\"target\"]:\"ensemble\"})\n        \n        df_subms.insert(loc=1, column=' _ ', value=['   '] * sls[\"q_rows\"])\n        \n        df_subms[' _ '] = df_subms[' _ '].astype(str)\n        pd.set_option('display.max_rows',100)\n        pd.set_option('display.float_format', '{:.4f}'.format)\n        vcols = ['ID'] + [' _ '] + short_name_cols + [' _ '] + ['alls'] + [' _ '] + ['ensemble']\n        df_subms = df_subms[vcols]\n        display(df_subms.head(7))\n        pd.set_option('display.float_format', '{:.7f}'.format)\n        df_subms = df_subms.rename(columns={\"ensemble\":sls[\"target\"]})\n        \n        return df_subms\n        \n\n    sample_subm = pd.read_csv(path_to_ds + file_short_names[1] + \".csv\")\n\n    \n    def ensemble_tida(sls,submission=sample_subm):   \n        sls['sort'] = 'desc'\n        dfs = tida(sls)\n        dfD = dfs[['ID', sls['target']]]\n        dfD.to_csv(f'tida_desc.csv', index=False)\n        sls['sort'] = 'asc'\n        dfs = tida(sls)\n        dfA = dfs[['ID', sls['target']]]\n        dfA.to_csv(f'tida_asc.csv',  index=False)\n        target,d,a = sls['target'],sls['desc'],sls['asc']\n        submission[target] = dfD[target] * d + a * dfA[target]\n        return submission\n\n    submission = ensemble_tida(sls)\n    \n    return submission","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-01-14T13:09:41.597247Z","iopub.execute_input":"2026-01-14T13:09:41.597653Z","iopub.status.idle":"2026-01-14T13:09:44.287061Z","shell.execute_reply.started":"2026-01-14T13:09:41.597620Z","shell.execute_reply":"2026-01-14T13:09:44.285727Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"!cp /kaggle/input/catboost/submission.csv    /kaggle/working/subm_catboost.csv\n!cp /kaggle/input/liear-base/submission.csv  /kaggle/working/subm_linear.csv\n!cp /kaggle/input/mlp-base/submission.csv    /kaggle/working/subm_mlp.csv\npath_to_ds = \"/kaggle/working/\"\nfile_short_names = [\n    \"subm_catboost\",\n    \"subm_linear\",\n    \"subm_mlp\"\n]\nparams = {\n    'path'  : path_to_ds,\n    'sort'  : \"dynamic\",\n    'target': \"prediction\",\n    'q_rows': 538_150,          # 改成你真实 test 行数\n    'prefix': \"subm_\",\n\n    # 升序 / 降序混合比例\n    'desc'  : 0,\n    'asc'   : 1,\n\n    # 排名修正（最重要的旋钮）\n    'subwts': [+0.15, 0.00, -0.15],\n\n    # 模型基础权重\n    'subm'  : [\n        { 'name': \"subm_catboost\", 'weight': 0.25 },\n        { 'name': \"subm_linear\",   'weight': 0.35 },\n        { 'name': \"subm_mlp\",      'weight': 0.4 },\n    ]\n}\ndf = iBlend(path_to_ds, file_short_names, params)\ndf.to_csv(\"submission.csv\", index=False)\ndisplay(df.head())\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2026-01-14T13:10:56.124717Z","iopub.execute_input":"2026-01-14T13:10:56.125067Z","iopub.status.idle":"2026-01-14T13:11:39.324985Z","shell.execute_reply.started":"2026-01-14T13:10:56.125043Z","shell.execute_reply":"2026-01-14T13:11:39.323696Z"}},"outputs":[],"execution_count":null}]}