fmda/fmda_rnn_serial.ipynb

   1 {
   2  "cells": [
   3   {
   4    "cell_type": "markdown",
   5    "id": "83b774b3-ef55-480a-b999-506676e49145",
   6    "metadata": {},
   7    "source": [
   8     "# v2.1 run RNN strategy serial by Location\n",
   9     "\n",
  10     "This version of the RNN runs the model on each location separately, one at a time. Two main runs:\n",
  11     "1. Run separate model at each location - training and prediction at least location independently - training mode periods 0:train_ind (was 0:h2), then prediction in test_ind:end. Validation data, if any, are from train_ind:test_ind\n",
  12     "2. Run same model with multiple fitting calls 0:train_ind at different locations, compare prediction accuracy in test_ind:end  at for all location. \n"
  13    ]
  14   },
  15   {
  16    "cell_type": "code",
  17    "execution_count": null,
  18    "id": "83cc1dc4-3dcb-4325-9263-58101a3dc378",
  19    "metadata": {},
  20    "outputs": [],
  21    "source": [
  22     "import numpy as np\n",
  23     "from utils import print_dict_summary, print_first, str2time, logging_setup\n",
  24     "import pickle\n",
  25     "import logging\n",
  26     "import os.path as osp\n",
  27     "from moisture_rnn_pkl import pkl2train\n",
  28     "from moisture_rnn import RNNParams, RNNData, RNN, create_rnn_data2 \n",
  29     "from utils import hash2, read_yml, read_pkl, retrieve_url\n",
  30     "from moisture_rnn import RNN\n",
  31     "import reproducibility\n",
  32     "from data_funcs import rmse, to_json\n",
  33     "from moisture_models import run_augmented_kf\n",
  34     "import copy\n",
  35     "import pandas as pd\n",
  36     "import matplotlib.pyplot as plt\n",
  37     "import yaml"
  38    ]
  39   },
  40   {
  41    "cell_type": "code",
  42    "execution_count": null,
  43    "id": "17db9b90-a931-4674-a447-5b8ffbcdc86a",
  44    "metadata": {},
  45    "outputs": [],
  46    "source": [
  47     "logging_setup()"
  48    ]
  49   },
  50   {
  51    "cell_type": "code",
  52    "execution_count": null,
  53    "id": "35319c1c-7849-4b8c-8262-f5aa6656e0c7",
  54    "metadata": {},
  55    "outputs": [],
  56    "source": [
  57     "retrieve_url(\n",
  58     "    url = \"https://demo.openwfm.org/web/data/fmda/dicts/test_CA_202401.pkl\", \n",
  59     "    dest_path = \"data/test_CA_202401.pkl\")"
  60    ]
  61   },
  62   {
  63    "cell_type": "code",
  64    "execution_count": null,
  65    "id": "eabdbd9c-07d9-4bae-9851-cca79f321895",
  66    "metadata": {},
  67    "outputs": [],
  68    "source": [
  69     "repro_file = \"data/reproducibility_dict_v2_TEST.pkl\"\n",
  70     "file_names=['fmda_nw_202401-05_f05.pkl']\n",
  71     "file_dir='data'\n",
  72     "file_paths = [osp.join(file_dir,file_name) for file_name in file_names]"
  73    ]
  74   },
  75   {
  76    "cell_type": "code",
  77    "execution_count": null,
  78    "id": "dcca6185-e799-4dd1-8acb-87ad33c411d7",
  79    "metadata": {},
  80    "outputs": [],
  81    "source": [
  82     "# read/write control\n",
  83     "train_file='train.pkl'\n",
  84     "train_create=False   # if false, read\n",
  85     "train_write=False\n",
  86     "train_read=True"
  87    ]
  88   },
  89   {
  90    "cell_type": "code",
  91    "execution_count": null,
  92    "id": "bc0a775b-b587-42ef-8576-e36dc0be3a75",
  93    "metadata": {
  94     "scrolled": true
  95    },
  96    "outputs": [],
  97    "source": [
  98     "repro = read_pkl(repro_file)\n",
  99     "\n",
 100     "if train_create:\n",
 101     "    logging.info('creating the training cases from files %s',file_paths)\n",
 102     "    # osp.join works on windows too, joins paths using \\ or /\n",
 103     "    train = pkl2train(file_paths)\n",
 104     "if train_write:\n",
 105     "    with open(train_file, 'wb') as file:\n",
 106     "        logging.info('Writing the rain cases into file %s',train_file)\n",
 107     "        pickle.dump(train, file)\n",
 108     "if train_read:\n",
 109     "    logging.info('Reading the train cases from file %s',train_file)\n",
 110     "    train = read_pkl(train_file)"
 111    ]
 112   },
 113   {
 114    "cell_type": "code",
 115    "execution_count": null,
 116    "id": "211a1c2f-ba8d-40b8-b29c-daa38af97a26",
 117    "metadata": {},
 118    "outputs": [],
 119    "source": [
 120     "params_all = read_yml(\"params.yaml\")\n",
 121     "print(params_all.keys())"
 122    ]
 123   },
 124   {
 125    "cell_type": "code",
 126    "execution_count": null,
 127    "id": "698df86b-8550-4135-81df-45dbf503dd4e",
 128    "metadata": {},
 129    "outputs": [],
 130    "source": [
 131     "# from module_param_sets import param_sets"
 132    ]
 133   },
 134   {
 135    "cell_type": "code",
 136    "execution_count": null,
 137    "id": "4b0c9a9b-dd02-4251-aa4a-2acc1101e153",
 138    "metadata": {},
 139    "outputs": [],
 140    "source": [
 141     "param_sets_keys=['rnn']\n",
 142     "cases=list(train.keys())[0:100]\n",
 143     "# cases=list(train.keys())\n",
 144     "# cases.remove('reproducibility')\n",
 145     "cases"
 146    ]
 147   },
 148   {
 149    "cell_type": "code",
 150    "execution_count": null,
 151    "id": "dd22baf2-59d2-460e-8c47-b20116dd5982",
 152    "metadata": {},
 153    "outputs": [],
 154    "source": [
 155     "logging.info('Running over parameter sets %s',param_sets_keys)\n",
 156     "logging.info('Running over cases %s',cases)"
 157    ]
 158   },
 159   {
 160    "cell_type": "markdown",
 161    "id": "802f3eef-1702-4478-b6e3-2288a6edae24",
 162    "metadata": {},
 163    "source": [
 164     "## Run Reproducibility Case"
 165    ]
 166   },
 167   {
 168    "cell_type": "code",
 169    "execution_count": null,
 170    "id": "69a3adb9-39fd-4c0c-9c9b-aaa2a9a3af40",
 171    "metadata": {},
 172    "outputs": [],
 173    "source": [
 174     "params = repro['repro_info']['params']\n",
 175     "print(type(params))\n",
 176     "print(params)\n",
 177     "\n",
 178     "# Set up input data\n",
 179     "rnn_dat = RNNData(repro, scaler = params['scaler'], features_list = params['features_list'])\n",
 180     "rnn_dat.train_test_split(\n",
 181     "    train_frac = params['train_frac'],\n",
 182     "    val_frac = params['val_frac']\n",
 183     ")\n",
 184     "rnn_dat.scale_data()"
 185    ]
 186   },
 187   {
 188    "cell_type": "code",
 189    "execution_count": null,
 190    "id": "855703c4-d7a9-4579-bca7-7c737a81d0de",
 191    "metadata": {},
 192    "outputs": [],
 193    "source": [
 194     "reproducibility.set_seed(123)\n",
 195     "rnn = RNN(params)\n",
 196     "m, errs = rnn.run_model(rnn_dat, reproducibility_run=True)"
 197    ]
 198   },
 199   {
 200    "cell_type": "markdown",
 201    "id": "49e31fdd-4c14-4a81-9e2b-4c6ba94d1f83",
 202    "metadata": {},
 203    "source": [
 204     "## Separate Models by Location"
 205    ]
 206   },
 207   {
 208    "cell_type": "code",
 209    "execution_count": null,
 210    "id": "e11e7c83-183f-48ba-abd8-a6aedff66090",
 211    "metadata": {},
 212    "outputs": [],
 213    "source": [
 214     "# Set up output dictionaries\n",
 215     "outputs_kf = {}\n",
 216     "outputs_rnn = {}"
 217    ]
 218   },
 219   {
 220    "cell_type": "code",
 221    "execution_count": null,
 222    "id": "dc5b47bd-4fbc-44b8-b2dd-d118e068b450",
 223    "metadata": {},
 224    "outputs": [],
 225    "source": [
 226     "\n",
 227     "for k in param_sets_keys:\n",
 228     "    params = RNNParams(params_all[k])\n",
 229     "    print(\"~\"*80)\n",
 230     "    print(\"Running with params:\")\n",
 231     "    print(params)\n",
 232     "    # Increase Val Frac so no errors, TODO fix validation\n",
 233     "    params.update({\n",
 234     "        'train_frac': .9,\n",
 235     "        'val_frac': .05,\n",
 236     "        'activation': ['relu', 'relu'],\n",
 237     "        'epochs': 10,\n",
 238     "        'dense_units': 10,\n",
 239     "        'rnn_layers': 2,\n",
 240     "        'batch_reset': 20        \n",
 241     "    })\n",
 242     "    for case in cases:\n",
 243     "        print(\"~\"*50)\n",
 244     "        logging.info('Processing case %s',case)\n",
 245     "        print_dict_summary(train[case])\n",
 246     "        # Format data & Run Model\n",
 247     "        # rnn_dat = create_rnn_data2(train[case], params)\n",
 248     "        rnn_dat = RNNData(train[case], scaler = params['scaler'], features_list = params['features_list'])\n",
 249     "        rnn_dat.train_test_split(\n",
 250     "            train_frac = params['train_frac'],\n",
 251     "            val_frac = params['val_frac']\n",
 252     "        )\n",
 253     "        rnn_dat.scale_data()\n",
 254     "        reproducibility.set_seed()\n",
 255     "        rnn = RNN(params)\n",
 256     "        m, errs = rnn.run_model(rnn_dat)\n",
 257     "        # Add model output to case\n",
 258     "        train[case]['m']=m\n",
 259     "        # Get RMSE Prediction Error\n",
 260     "        print(f\"RMSE: {errs}\")\n",
 261     "        outputs_rnn[case] = {'case':case, 'errs': errs.copy()}\n",
 262     "        \n",
 263     "        # Run Augmented KF\n",
 264     "        print('Running Augmented KF')\n",
 265     "        train[case]['h2'] = train[case]['hours'] // 2\n",
 266     "        train[case]['scale_fm'] = 1\n",
 267     "        m, Ec = run_augmented_kf(train[case])\n",
 268     "        m = m*rnn_dat['scale_fm']\n",
 269     "        y = rnn_dat['y']*rnn_dat['scale_fm']          \n",
 270     "        train[case]['m'] = m\n",
 271     "        print(f\"KF RMSE: {rmse(m,y)}\")\n",
 272     "        outputs_kf[case] = {'case':case, 'errs': rmse(m,y)}\n",
 273     "\n",
 274     "        # Save Outputs \n",
 275     "        to_json(outputs_rnn, \"rnn_errs.json\")\n",
 276     "        to_json(outputs_kf, \"kf_errs.json\")"
 277    ]
 278   },
 279   {
 280    "cell_type": "code",
 281    "execution_count": null,
 282    "id": "15384e4d-b8ec-4700-bdc2-83b0433d11c9",
 283    "metadata": {},
 284    "outputs": [],
 285    "source": [
 286     "logging.info('fmda_rnn_serial.ipynb done')"
 287    ]
 288   },
 289   {
 290    "cell_type": "code",
 291    "execution_count": null,
 292    "id": "d0e78fb3-b501-49d6-81a9-1a13da0134a0",
 293    "metadata": {},
 294    "outputs": [],
 295    "source": [
 296     "import importlib\n",
 297     "import moisture_rnn\n",
 298     "importlib.reload(moisture_rnn)\n",
 299     "from moisture_rnn import RNN"
 300    ]
 301   },
 302   {
 303    "cell_type": "code",
 304    "execution_count": null,
 305    "id": "37053436-8dfe-4c40-8614-811817e83782",
 306    "metadata": {},
 307    "outputs": [],
 308    "source": [
 309     "for k in outputs_rnn:\n",
 310     "    print(\"~\"*50)\n",
 311     "    print(outputs_rnn[k]['case'])\n",
 312     "    print(outputs_rnn[k]['errs']['prediction'])"
 313    ]
 314   },
 315   {
 316    "cell_type": "code",
 317    "execution_count": null,
 318    "id": "9154d5f7-015f-4ef7-af45-020410a1ea65",
 319    "metadata": {},
 320    "outputs": [],
 321    "source": [
 322     "for k in outputs_kf:\n",
 323     "    print(\"~\"*50)\n",
 324     "    print(outputs_kf[k]['case'])\n",
 325     "    print(outputs_kf[k]['errs'])"
 326    ]
 327   },
 328   {
 329    "cell_type": "markdown",
 330    "id": "f3c1c299-1655-4c64-a458-c7723db6ea6d",
 331    "metadata": {},
 332    "source": [
 333     "### TODO: FIX SCALING"
 334    ]
 335   },
 336   {
 337    "cell_type": "markdown",
 338    "id": "0c0c3470-30f5-4915-98a7-dcdf5760d482",
 339    "metadata": {},
 340    "source": [
 341     "## Training at Multiple Locations\n",
 342     "\n",
 343     "Still sequential"
 344    ]
 345   },
 346   {
 347    "cell_type": "code",
 348    "execution_count": null,
 349    "id": "dd1aca73-7279-473e-b2a3-95aa1db7b1a8",
 350    "metadata": {},
 351    "outputs": [],
 352    "source": [
 353     "params = RNNParams(params_all['rnn'])\n",
 354     "params.update({\n",
 355     "    'epochs': 1, # less epochs since it is per location\n",
 356     "    'activation': ['relu', 'relu'],\n",
 357     "    'train_frac': .9,\n",
 358     "    'val_frac': .05,    \n",
 359     "    'dense_units': 10,\n",
 360     "    'rnn_layers': 2,\n",
 361     "    'batch_reset': 20 \n",
 362     "})\n",
 363     "\n",
 364     "# rnn_dat = create_rnn_data2(train[cases[0]], params)\n",
 365     "rnn_dat = RNNData(train[cases[0]], params['scaler'], params['features_list'])\n",
 366     "rnn_dat.train_test_split(\n",
 367     "    train_frac = params['train_frac'],\n",
 368     "    val_frac = params['val_frac']\n",
 369     ")\n",
 370     "rnn_dat.scale_data()"
 371    ]
 372   },
 373   {
 374    "cell_type": "code",
 375    "execution_count": null,
 376    "id": "65b2f9a3-a8f2-4ac1-8e4d-ba38a86eaf40",
 377    "metadata": {},
 378    "outputs": [],
 379    "source": [
 380     "reproducibility.set_seed()\n",
 381     "rnn = RNN(params)"
 382    ]
 383   },
 384   {
 385    "cell_type": "code",
 386    "execution_count": null,
 387    "id": "47a85ef2-8145-4de8-9f2e-86622306ffd8",
 388    "metadata": {
 389     "scrolled": true
 390    },
 391    "outputs": [],
 392    "source": [
 393     "print(\"~\"*80)\n",
 394     "print(\"Running with params:\")\n",
 395     "print(params)\n",
 396     "\n",
 397     "for case in cases[0:10]:\n",
 398     "    print(\"~\"*50)\n",
 399     "    logging.info('Processing case %s',case)\n",
 400     "    print_dict_summary(train[case])\n",
 401     "    rnn_dat_temp = RNNData(train[case], params['scaler'], params['features_list'])\n",
 402     "    rnn_dat_temp.train_test_split(\n",
 403     "        train_frac = params['train_frac'],\n",
 404     "        val_frac = params['val_frac']\n",
 405     "    )\n",
 406     "    rnn_dat_temp.scale_data()\n",
 407     "    rnn.fit(rnn_dat_temp['X_train'], rnn_dat_temp['y_train'],\n",
 408     "           validation_data=(rnn_dat_temp['X_val'], rnn_dat_temp['y_val']))\n",
 409     "    # run_rnn_pkl(train[case],param_sets[i])"
 410    ]
 411   },
 412   {
 413    "cell_type": "markdown",
 414    "id": "a0421b8d-49aa-4409-8cbf-7732f1137838",
 415    "metadata": {},
 416    "source": [
 417     "### Predict "
 418    ]
 419   },
 420   {
 421    "cell_type": "code",
 422    "execution_count": null,
 423    "id": "63d7854a-94f7-425c-9561-4fe518e044bb",
 424    "metadata": {
 425     "scrolled": true
 426    },
 427    "outputs": [],
 428    "source": [
 429     "# Predict Cases Used in Training\n",
 430     "rmses = []\n",
 431     "inds = np.arange(0,10)\n",
 432     "train_keys = list(train.keys())\n",
 433     "for i in inds:\n",
 434     "    print(\"~\"*50)\n",
 435     "    case = train_keys[i]\n",
 436     "    print(f\"Predicting case {case}\")\n",
 437     "    # rnn_dat = create_rnn_data2(train[case], params)\n",
 438     "    rnn_dat_temp = RNNData(train[case], params['scaler'], params['features_list'])\n",
 439     "    rnn_dat_temp.train_test_split(\n",
 440     "        train_frac = params['train_frac'],\n",
 441     "        val_frac = params['val_frac']\n",
 442     "    )\n",
 443     "    rnn_dat_temp.scale_data()\n",
 444     "    X_temp = np.concatenate((rnn_dat_temp.X_train, rnn_dat_temp.X_val, rnn_dat_temp.X_test), axis=0)\n",
 445     "    m = rnn.predict(X_temp)\n",
 446     "    test_ind = rnn_dat['test_ind']\n",
 447     "    rmses.append(rmse(m[test_ind:], rnn_dat['y_test'].flatten()))"
 448    ]
 449   },
 450   {
 451    "cell_type": "code",
 452    "execution_count": null,
 453    "id": "2a5423e0-778b-4f69-9ed0-f0082a1fefe5",
 454    "metadata": {},
 455    "outputs": [],
 456    "source": [
 457     "rmses"
 458    ]
 459   },
 460   {
 461    "cell_type": "code",
 462    "execution_count": null,
 463    "id": "45c9caae-7ced-4f21-aa05-c9b125e8fdcb",
 464    "metadata": {},
 465    "outputs": [],
 466    "source": [
 467     "pd.DataFrame({'Case': list(train.keys())[0:10], 'RMSE': rmses}).style.hide(axis=\"index\")"
 468    ]
 469   },
 470   {
 471    "cell_type": "code",
 472    "execution_count": null,
 473    "id": "f710f482-b600-4ea5-9a8a-823a13b4ec7a",
 474    "metadata": {
 475     "scrolled": true
 476    },
 477    "outputs": [],
 478    "source": [
 479     "# Predict New Locations\n",
 480     "rmses = []\n",
 481     "for i, case in enumerate(list(train.keys())[10:100]):\n",
 482     "    print(\"~\"*50)\n",
 483     "    print(f\"Predicting case {case}\")\n",
 484     "    rnn_dat_temp = RNNData(train[case], params['scaler'], params['features_list'])\n",
 485     "    rnn_dat_temp.train_test_split(\n",
 486     "        train_frac = params['train_frac'],\n",
 487     "        val_frac = params['val_frac']\n",
 488     "    )\n",
 489     "    rnn_dat_temp.scale_data()\n",
 490     "    X = np.concatenate((rnn_dat_temp.X_train, rnn_dat_temp.X_val, rnn_dat_temp.X_test))\n",
 491     "    m = rnn.predict(X)\n",
 492     "    train[case]['m'] = m\n",
 493     "    test_ind = rnn_dat['test_ind']\n",
 494     "    rmses.append(rmse(m[test_ind:], rnn_dat.y_test.flatten()))\n",
 495     "\n",
 496     "df = pd.DataFrame({'Case': list(train.keys())[10:100], 'RMSE': rmses})"
 497    ]
 498   },
 499   {
 500    "cell_type": "code",
 501    "execution_count": null,
 502    "id": "d793ac87-d94b-4b16-a271-46cdc259b4fe",
 503    "metadata": {},
 504    "outputs": [],
 505    "source": [
 506     "df[0:5].style.hide(axis=\"index\")"
 507    ]
 508   },
 509   {
 510    "cell_type": "code",
 511    "execution_count": null,
 512    "id": "b99606d1-bd46-4041-8303-1bcbb196f6f4",
 513    "metadata": {},
 514    "outputs": [],
 515    "source": [
 516     "df"
 517    ]
 518   },
 519   {
 520    "cell_type": "code",
 521    "execution_count": null,
 522    "id": "52ec264d-d4b7-444c-b623-002d6383da30",
 523    "metadata": {},
 524    "outputs": [],
 525    "source": [
 526     "df.RMSE.mean()"
 527    ]
 528   },
 529   {
 530    "cell_type": "code",
 531    "execution_count": null,
 532    "id": "998922cd-46bb-4063-8284-0497e19c39b0",
 533    "metadata": {},
 534    "outputs": [],
 535    "source": [
 536     "plt.hist(df.RMSE)"
 537    ]
 538   },
 539   {
 540    "cell_type": "code",
 541    "execution_count": null,
 542    "id": "889f3bbb-9fb2-4621-9e93-1d0bc0f83e01",
 543    "metadata": {},
 544    "outputs": [],
 545    "source": []
 546   },
 547   {
 548    "cell_type": "code",
 549    "execution_count": null,
 550    "id": "fe407f61-15f2-4086-a386-7d7a5bb90d26",
 551    "metadata": {},
 552    "outputs": [],
 553    "source": []
 554   },
 555   {
 556    "cell_type": "code",
 557    "execution_count": null,
 558    "id": "2fdb63b3-68b8-4877-a7a2-f63257cb29d5",
 559    "metadata": {},
 560    "outputs": [],
 561    "source": []
 562   },
 563   {
 564    "cell_type": "code",
 565    "execution_count": null,
 566    "id": "5c7563c5-a880-45c7-8381-8ce4e1a44216",
 567    "metadata": {},
 568    "outputs": [],
 569    "source": []
 570   },
 571   {
 572    "cell_type": "code",
 573    "execution_count": null,
 574    "id": "ad5dae6c-1269-4674-a49e-2efe8b956911",
 575    "metadata": {},
 576    "outputs": [],
 577    "source": []
 578   }
 579  ],
 580  "metadata": {
 581   "kernelspec": {
 582    "display_name": "Python 3 (ipykernel)",
 583    "language": "python",
 584    "name": "python3"
 585   },
 586   "language_info": {
 587    "codemirror_mode": {
 588     "name": "ipython",
 589     "version": 3
 590    },
 591    "file_extension": ".py",
 592    "mimetype": "text/x-python",
 593    "name": "python",
 594    "nbconvert_exporter": "python",
 595    "pygments_lexer": "ipython3",
 596    "version": "3.12.5"
 597   }
 598  },
 599  "nbformat": 4,
 600  "nbformat_minor": 5
 601 }