Diff of
/experiments/Foresight MIMIC Final -- Prepare data.ipynb
[000000]
..
[20be63]
Switch to unified view
| a | b/experiments/Foresight MIMIC Final -- Prepare data.ipynb | ||
|---|---|---|---|
| 1 | { |
||
| 2 | "cells": [ |
||
| 3 | { |
||
| 4 | "cell_type": "code", |
||
| 5 | "execution_count": null, |
||
| 6 | "id": "01ae6e4a", |
||
| 7 | "metadata": {}, |
||
| 8 | "outputs": [], |
||
| 9 | "source": [ |
||
| 10 | "import torch\n", |
||
| 11 | "import os\n", |
||
| 12 | "import datasets\n", |
||
| 13 | "import numpy as np\n", |
||
| 14 | "from collections import defaultdict\n", |
||
| 15 | "from medcat.cat import CAT\n", |
||
| 16 | "from foresight.datasets import patient_concept_stream\n", |
||
| 17 | "from foresight.datasets.filters import filter_by_count, filter_by_type\n", |
||
| 18 | "from foresight.datasets.utils import get_embeddings_for_tokens, stream_to_separate_examples, add_to_stream, \\\n", |
||
| 19 | " remove_parents_from_stream, bucket_concepts, cleanup_stream, \\\n", |
||
| 20 | " split_stream, add_age, get_all_splits, add_ttd, add_position_ids\n", |
||
| 21 | "from foresight.utils import pickle\n", |
||
| 22 | "from foresight.utils.cdb_utils import get_parents_map \n", |
||
| 23 | "from foresight.utils.stream_utils import docs2stream, calculate_counts\n", |
||
| 24 | "from foresight.tokenizers.simple_map_tokenizer import SimpleMapTokenizer\n", |
||
| 25 | "from foresight.metrics.next_concept_prediction import precision, metrics_data2df, ComputePrecisionHF\n", |
||
| 26 | "from medcat.cdb import CDB\n", |
||
| 27 | "from foresight.utils import pickle\n", |
||
| 28 | "import plotly.express as px" |
||
| 29 | ] |
||
| 30 | }, |
||
| 31 | { |
||
| 32 | "cell_type": "code", |
||
| 33 | "execution_count": null, |
||
| 34 | "id": "9f57b67c", |
||
| 35 | "metadata": {}, |
||
| 36 | "outputs": [], |
||
| 37 | "source": [ |
||
| 38 | "DAYS = 1 # Do: 1, 14, 30\n", |
||
| 39 | "MAX_SEQ_LEN = 256\n", |
||
| 40 | "#TYPES = ['T-45', 'T-55', 'T-26', 'T-29', 'T-40', 'T-9', 'T-27', 'T-11', 'T-39', 'T-18']\n", |
||
| 41 | "TYPES = ['ALL_TYPES']\n", |
||
| 42 | "#TYPES = ['T-11']\n", |
||
| 43 | "#TYPES = ['T-11', 'T-18']" |
||
| 44 | ] |
||
| 45 | }, |
||
| 46 | { |
||
| 47 | "cell_type": "code", |
||
| 48 | "execution_count": null, |
||
| 49 | "id": "5ac3ee70", |
||
| 50 | "metadata": {}, |
||
| 51 | "outputs": [], |
||
| 52 | "source": [ |
||
| 53 | "BASE_NAME = 'annotated_february_2022'\n", |
||
| 54 | "DATASET_NAME = 'annotations_stream_phase2_v1'\n", |
||
| 55 | "RUN_NAME = f'{DATASET_NAME}_{DAYS}d_{MAX_SEQ_LEN}_{\"_\".join(TYPES)}'" |
||
| 56 | ] |
||
| 57 | }, |
||
| 58 | { |
||
| 59 | "cell_type": "code", |
||
| 60 | "execution_count": null, |
||
| 61 | "id": "c865ce1e", |
||
| 62 | "metadata": {}, |
||
| 63 | "outputs": [], |
||
| 64 | "source": [ |
||
| 65 | "ds_info = open(\"dataset-info/\" + RUN_NAME + '.txt', 'w')\n", |
||
| 66 | "def fprint(*texts):\n", |
||
| 67 | " for text in texts:\n", |
||
| 68 | " print(text)\n", |
||
| 69 | " ds_info.write(str(text) + \"\\n\")" |
||
| 70 | ] |
||
| 71 | }, |
||
| 72 | { |
||
| 73 | "cell_type": "code", |
||
| 74 | "execution_count": null, |
||
| 75 | "id": "922abc5f", |
||
| 76 | "metadata": {}, |
||
| 77 | "outputs": [], |
||
| 78 | "source": [ |
||
| 79 | "FROM_BASE = False\n", |
||
| 80 | "BASE_TOKENIZER_PATH = ''" |
||
| 81 | ] |
||
| 82 | }, |
||
| 83 | { |
||
| 84 | "cell_type": "code", |
||
| 85 | "execution_count": null, |
||
| 86 | "id": "c32b0b92", |
||
| 87 | "metadata": {}, |
||
| 88 | "outputs": [], |
||
| 89 | "source": [ |
||
| 90 | "TYPES = set(TYPES)" |
||
| 91 | ] |
||
| 92 | }, |
||
| 93 | { |
||
| 94 | "cell_type": "code", |
||
| 95 | "execution_count": null, |
||
| 96 | "id": "6efddbcc", |
||
| 97 | "metadata": {}, |
||
| 98 | "outputs": [], |
||
| 99 | "source": [ |
||
| 100 | "DATA_PATH = f\"./data/timecat/mimic/{BASE_NAME}/{DATASET_NAME}.pickle\"\n", |
||
| 101 | "DATA_PATH_SPLITS = f\"./data/timecat/mimic/{BASE_NAME}/{DATASET_NAME}_split/\"\n", |
||
| 102 | "TOKENIZER_PATH = f\"./data/timecat/models/gpt/tokenizer_{RUN_NAME}.pickle\"\n", |
||
| 103 | "ALMOST_PREPARED_DATASET_SPLIT_PATH = f\"./data/timecat/mimic/{BASE_NAME}/{RUN_NAME}_almost_prepared_split/\"\n", |
||
| 104 | "PREPARED_DATASET_SPLIT_PATH = f\"./data/timecat/mimic/{BASE_NAME}/{RUN_NAME}_prepared_split/\"\n", |
||
| 105 | "JUST_BEFORE_ENCODING_DATASET_SPLIT_PATH = f\"./data/timecat/mimic/{BASE_NAME}/{RUN_NAME}_just_before_encoding/\"\n", |
||
| 106 | "CAT_PATH = \"./data/models/modelpacks/mc_modelpack_phase2_snomed_190k_february_2022.zip\"\n", |
||
| 107 | "PT_DOB_PATH = \"./data/mimic/pt2dob_datetime.pickle\"\n", |
||
| 108 | "PT_DOD_PATH = \"./data/mimic/pt2dod_timestamp.pickle\"\n", |
||
| 109 | "PT_SEX_PATH = \"./data/mimic/pt2sex.pickle\"\n", |
||
| 110 | "PT_LNS_PATH = f\"./data/timecat/mimic/{BASE_NAME}/lns_{DATASET_NAME}.pickle\"\n", |
||
| 111 | "PT_CNTS_PATH = f\"./data/timecat/mimic/{BASE_NAME}/cnts_{DATASET_NAME}.pickle\"\n", |
||
| 112 | "PT_ETHNICITY_PATH = \"./data/mimic/pt2ethnicity.pickle\"\n", |
||
| 113 | "TOKEN_TYPES_PATH = f'./data/timecat/mimic/{BASE_NAME}/types_{DATASET_NAME}.pickle'\n", |
||
| 114 | "\n", |
||
| 115 | "BATCH_SIZE = 200\n", |
||
| 116 | "DEVICE = torch.device('cuda')\n", |
||
| 117 | "NUM_PROC = 16\n", |
||
| 118 | "MIN_COUNT = 2 # 3\n", |
||
| 119 | "MIN_GLOBAL_COUNT = 100 # 1000" |
||
| 120 | ] |
||
| 121 | }, |
||
| 122 | { |
||
| 123 | "cell_type": "code", |
||
| 124 | "execution_count": null, |
||
| 125 | "id": "a8eddb07", |
||
| 126 | "metadata": { |
||
| 127 | "scrolled": true |
||
| 128 | }, |
||
| 129 | "outputs": [], |
||
| 130 | "source": [ |
||
| 131 | "cat = CAT.load_model_pack(CAT_PATH, meta_cat_config_dict={'general': {'device': 'cpu'}})\n", |
||
| 132 | "cdb = cat.cdb" |
||
| 133 | ] |
||
| 134 | }, |
||
| 135 | { |
||
| 136 | "cell_type": "markdown", |
||
| 137 | "id": "c5b51800", |
||
| 138 | "metadata": {}, |
||
| 139 | "source": [ |
||
| 140 | "# Convert docs.pickle into patient stream used by HF datasets" |
||
| 141 | ] |
||
| 142 | }, |
||
| 143 | { |
||
| 144 | "cell_type": "code", |
||
| 145 | "execution_count": null, |
||
| 146 | "id": "9445cd0d", |
||
| 147 | "metadata": {}, |
||
| 148 | "outputs": [], |
||
| 149 | "source": [ |
||
| 150 | "doc2pt = pickle.load(\"./data/timecat/mimic/doc2pt.pickle\")" |
||
| 151 | ] |
||
| 152 | }, |
||
| 153 | { |
||
| 154 | "cell_type": "code", |
||
| 155 | "execution_count": null, |
||
| 156 | "id": "41568f81", |
||
| 157 | "metadata": {}, |
||
| 158 | "outputs": [], |
||
| 159 | "source": [ |
||
| 160 | "doc2pt = {str(k):v for k,v in doc2pt.items()}" |
||
| 161 | ] |
||
| 162 | }, |
||
| 163 | { |
||
| 164 | "cell_type": "markdown", |
||
| 165 | "id": "e5bdc483", |
||
| 166 | "metadata": {}, |
||
| 167 | "source": [ |
||
| 168 | "### Get counts" |
||
| 169 | ] |
||
| 170 | }, |
||
| 171 | { |
||
| 172 | "cell_type": "code", |
||
| 173 | "execution_count": null, |
||
| 174 | "id": "ad82350f", |
||
| 175 | "metadata": {}, |
||
| 176 | "outputs": [], |
||
| 177 | "source": [ |
||
| 178 | "pt2cui2cnt = None\n", |
||
| 179 | "base_path = './data/timecat/mimic/annotated_february_2022/'\n", |
||
| 180 | "doc_paths = os.listdir(base_path)\n", |
||
| 181 | "doc_paths = [path for path in doc_paths if path.startswith(\"part_\")] # So we keep only annotations data\n", |
||
| 182 | "\n", |
||
| 183 | "for path in doc_paths:\n", |
||
| 184 | " docs = pickle.load(os.path.join(base_path, path))\n", |
||
| 185 | " \n", |
||
| 186 | " pt2cui2cnt = calculate_counts(docs=docs,\n", |
||
| 187 | " doc2pt=doc2pt,\n", |
||
| 188 | " pt2cui2cnt=pt2cui2cnt,\n", |
||
| 189 | " meta_requirements={'Presence': 'True', 'Subject': 'Patient'})" |
||
| 190 | ] |
||
| 191 | }, |
||
| 192 | { |
||
| 193 | "cell_type": "code", |
||
| 194 | "execution_count": null, |
||
| 195 | "id": "063bf5b0", |
||
| 196 | "metadata": {}, |
||
| 197 | "outputs": [], |
||
| 198 | "source": [ |
||
| 199 | "pickle.dump(dict(pt2cui2cnt), f\"./data/timecat/mimic/{BASE_NAME}/pt2cui2cnt.pickle\")" |
||
| 200 | ] |
||
| 201 | }, |
||
| 202 | { |
||
| 203 | "cell_type": "code", |
||
| 204 | "execution_count": null, |
||
| 205 | "id": "d0355f5e", |
||
| 206 | "metadata": {}, |
||
| 207 | "outputs": [], |
||
| 208 | "source": [ |
||
| 209 | "pt2cui2cnt = pickle.load(f\"./data/timecat/mimic/{BASE_NAME}/pt2cui2cnt.pickle\")" |
||
| 210 | ] |
||
| 211 | }, |
||
| 212 | { |
||
| 213 | "cell_type": "code", |
||
| 214 | "execution_count": null, |
||
| 215 | "id": "355925a7", |
||
| 216 | "metadata": {}, |
||
| 217 | "outputs": [], |
||
| 218 | "source": [ |
||
| 219 | "# Total number of annotations per type\n", |
||
| 220 | "cnt_per_type = {}\n", |
||
| 221 | "other_cnt = 0\n", |
||
| 222 | "for pt in pt2cui2cnt:\n", |
||
| 223 | " for cui in pt2cui2cnt[pt]:\n", |
||
| 224 | " if cat.cdb.cui2type_ids[cui]:\n", |
||
| 225 | " t = list(cat.cdb.cui2type_ids[cui])[0]\n", |
||
| 226 | " cnt_per_type[t] = cnt_per_type.get(t, 0) + pt2cui2cnt[pt][cui]\n", |
||
| 227 | " else:\n", |
||
| 228 | " other_cnt += 1" |
||
| 229 | ] |
||
| 230 | }, |
||
| 231 | { |
||
| 232 | "cell_type": "code", |
||
| 233 | "execution_count": null, |
||
| 234 | "id": "79c7b309", |
||
| 235 | "metadata": {}, |
||
| 236 | "outputs": [], |
||
| 237 | "source": [ |
||
| 238 | "fprint(\"Total number of annotations per type: \")\n", |
||
| 239 | "for t in cnt_per_type:\n", |
||
| 240 | " fprint(\"{:30}: {}\".format(cat.cdb.addl_info['type_id2name'][t].title(), cnt_per_type[t]))\n", |
||
| 241 | "fprint(\"\")" |
||
| 242 | ] |
||
| 243 | }, |
||
| 244 | { |
||
| 245 | "cell_type": "code", |
||
| 246 | "execution_count": null, |
||
| 247 | "id": "c513bf0c", |
||
| 248 | "metadata": {}, |
||
| 249 | "outputs": [], |
||
| 250 | "source": [ |
||
| 251 | "fprint(\"Total number of annotations: \", sum([x for x in cnt_per_type.values()]))\n", |
||
| 252 | "fprint(\"\")" |
||
| 253 | ] |
||
| 254 | }, |
||
| 255 | { |
||
| 256 | "cell_type": "code", |
||
| 257 | "execution_count": null, |
||
| 258 | "id": "6a0375ab", |
||
| 259 | "metadata": {}, |
||
| 260 | "outputs": [], |
||
| 261 | "source": [ |
||
| 262 | "# Get total number of different concepts\n", |
||
| 263 | "all_cuis = set()\n", |
||
| 264 | "for pt in pt2cui2cnt.keys():\n", |
||
| 265 | " for cui in pt2cui2cnt[pt]: \n", |
||
| 266 | " all_cuis.add(cui)\n", |
||
| 267 | "fprint(\"Total number of different concepts: \", len(all_cuis))\n", |
||
| 268 | "fprint(\"\")" |
||
| 269 | ] |
||
| 270 | }, |
||
| 271 | { |
||
| 272 | "cell_type": "code", |
||
| 273 | "execution_count": null, |
||
| 274 | "id": "5e5e6c4f", |
||
| 275 | "metadata": {}, |
||
| 276 | "outputs": [], |
||
| 277 | "source": [ |
||
| 278 | "# Total number of patients\n", |
||
| 279 | "fprint(\"Total number of patients: \", len(pt2cui2cnt))\n", |
||
| 280 | "fprint(\"\")" |
||
| 281 | ] |
||
| 282 | }, |
||
| 283 | { |
||
| 284 | "cell_type": "markdown", |
||
| 285 | "id": "8e3471c9", |
||
| 286 | "metadata": {}, |
||
| 287 | "source": [ |
||
| 288 | "### Get pt2stream" |
||
| 289 | ] |
||
| 290 | }, |
||
| 291 | { |
||
| 292 | "cell_type": "code", |
||
| 293 | "execution_count": null, |
||
| 294 | "id": "deb8bf33", |
||
| 295 | "metadata": {}, |
||
| 296 | "outputs": [], |
||
| 297 | "source": [ |
||
| 298 | "base_path = f'./data/timecat/mimic/{BASE_NAME}/'\n", |
||
| 299 | "doc_paths = os.listdir(base_path)\n", |
||
| 300 | "doc_paths = [path for path in doc_paths if path.startswith(\"part_\")] # So we keep only annotations data\n", |
||
| 301 | "pt2stream = None\n", |
||
| 302 | "doc2time = {str(k):v for k,v in pickle.load(\"./data/timecat/mimic/doc2time.pickle\").items()}\n", |
||
| 303 | "\n", |
||
| 304 | "for path in doc_paths:\n", |
||
| 305 | " docs = pickle.load(os.path.join(base_path, path))\n", |
||
| 306 | " pt2stream = docs2stream(docs,\n", |
||
| 307 | " doc2pt=doc2pt,\n", |
||
| 308 | " pt2cui2cnt=pt2cui2cnt,\n", |
||
| 309 | " doc2time=doc2time,\n", |
||
| 310 | " entity_type_column='type_ids',\n", |
||
| 311 | " meta_requirements={'Subject': 'Patient', 'Presence': 'True'},\n", |
||
| 312 | " historical_meta='Time',\n", |
||
| 313 | " historical_meta_value='Past',\n", |
||
| 314 | " old_pt2stream=pt2stream,\n", |
||
| 315 | " skip_cuis={'S-418023006', '17971005'},\n", |
||
| 316 | " require_time=True)" |
||
| 317 | ] |
||
| 318 | }, |
||
| 319 | { |
||
| 320 | "cell_type": "code", |
||
| 321 | "execution_count": null, |
||
| 322 | "id": "c89f0594", |
||
| 323 | "metadata": {}, |
||
| 324 | "outputs": [], |
||
| 325 | "source": [ |
||
| 326 | "pickle.dump(dict(pt2stream), DATA_PATH)" |
||
| 327 | ] |
||
| 328 | }, |
||
| 329 | { |
||
| 330 | "cell_type": "markdown", |
||
| 331 | "id": "6cfb0fa2", |
||
| 332 | "metadata": {}, |
||
| 333 | "source": [ |
||
| 334 | "# Load dataset" |
||
| 335 | ] |
||
| 336 | }, |
||
| 337 | { |
||
| 338 | "cell_type": "code", |
||
| 339 | "execution_count": null, |
||
| 340 | "id": "ba72e618", |
||
| 341 | "metadata": { |
||
| 342 | "scrolled": true |
||
| 343 | }, |
||
| 344 | "outputs": [], |
||
| 345 | "source": [ |
||
| 346 | "dataset = datasets.load_dataset(os.path.abspath(patient_concept_stream.__file__), \n", |
||
| 347 | " data_files={'train': DATA_PATH})['train']" |
||
| 348 | ] |
||
| 349 | }, |
||
| 350 | { |
||
| 351 | "cell_type": "markdown", |
||
| 352 | "id": "305eed65", |
||
| 353 | "metadata": {}, |
||
| 354 | "source": [ |
||
| 355 | "# Calculate counts" |
||
| 356 | ] |
||
| 357 | }, |
||
| 358 | { |
||
| 359 | "cell_type": "code", |
||
| 360 | "execution_count": null, |
||
| 361 | "id": "28bb5e02", |
||
| 362 | "metadata": {}, |
||
| 363 | "outputs": [], |
||
| 364 | "source": [ |
||
| 365 | "# Calculate counts for tokens\n", |
||
| 366 | "token_cnt = defaultdict(int)\n", |
||
| 367 | "for _dataset in get_all_splits(dataset):\n", |
||
| 368 | " for stream in _dataset['stream']:\n", |
||
| 369 | " unique_tokens = set([sample['token'] for sample in stream])\n", |
||
| 370 | " for token in unique_tokens:\n", |
||
| 371 | " token_cnt[token] += 1\n", |
||
| 372 | "token_cnt = dict(token_cnt)" |
||
| 373 | ] |
||
| 374 | }, |
||
| 375 | { |
||
| 376 | "cell_type": "code", |
||
| 377 | "execution_count": null, |
||
| 378 | "id": "d5fbd7c7", |
||
| 379 | "metadata": {}, |
||
| 380 | "outputs": [], |
||
| 381 | "source": [ |
||
| 382 | "pickle.dump(token_cnt, PT_CNTS_PATH)" |
||
| 383 | ] |
||
| 384 | }, |
||
| 385 | { |
||
| 386 | "cell_type": "code", |
||
| 387 | "execution_count": null, |
||
| 388 | "id": "9c928e97", |
||
| 389 | "metadata": {}, |
||
| 390 | "outputs": [], |
||
| 391 | "source": [ |
||
| 392 | "MIN_GLOBAL_COUNT = 100 # 1000" |
||
| 393 | ] |
||
| 394 | }, |
||
| 395 | { |
||
| 396 | "cell_type": "markdown", |
||
| 397 | "id": "3b6766cf", |
||
| 398 | "metadata": {}, |
||
| 399 | "source": [ |
||
| 400 | "# Load and filter by count" |
||
| 401 | ] |
||
| 402 | }, |
||
| 403 | { |
||
| 404 | "cell_type": "code", |
||
| 405 | "execution_count": null, |
||
| 406 | "id": "04085e31", |
||
| 407 | "metadata": {}, |
||
| 408 | "outputs": [], |
||
| 409 | "source": [ |
||
| 410 | "token_cnt = pickle.load(PT_CNTS_PATH)" |
||
| 411 | ] |
||
| 412 | }, |
||
| 413 | { |
||
| 414 | "cell_type": "code", |
||
| 415 | "execution_count": null, |
||
| 416 | "id": "049dfd9f", |
||
| 417 | "metadata": {}, |
||
| 418 | "outputs": [], |
||
| 419 | "source": [ |
||
| 420 | "dataset = filter_by_count(dataset, min_count=MIN_COUNT, min_count_global=MIN_GLOBAL_COUNT, min_length=5, max_length=-1, \n", |
||
| 421 | " num_proc=NUM_PROC, token_cnt=token_cnt)" |
||
| 422 | ] |
||
| 423 | }, |
||
| 424 | { |
||
| 425 | "cell_type": "markdown", |
||
| 426 | "id": "0ca8e8e8", |
||
| 427 | "metadata": {}, |
||
| 428 | "source": [ |
||
| 429 | "### Split and save" |
||
| 430 | ] |
||
| 431 | }, |
||
| 432 | { |
||
| 433 | "cell_type": "code", |
||
| 434 | "execution_count": null, |
||
| 435 | "id": "b5feea2f", |
||
| 436 | "metadata": {}, |
||
| 437 | "outputs": [], |
||
| 438 | "source": [ |
||
| 439 | "# Total number of annotations per type\n", |
||
| 440 | "cnt_per_type = {}\n", |
||
| 441 | "for cui in token_cnt:\n", |
||
| 442 | " if cat.cdb.cui2type_ids[cui]:\n", |
||
| 443 | " t = list(cat.cdb.cui2type_ids[cui])[0]\n", |
||
| 444 | " cnt_per_type[t] = cnt_per_type.get(t, 0) + token_cnt[cui]" |
||
| 445 | ] |
||
| 446 | }, |
||
| 447 | { |
||
| 448 | "cell_type": "code", |
||
| 449 | "execution_count": null, |
||
| 450 | "id": "5f743cfe", |
||
| 451 | "metadata": {}, |
||
| 452 | "outputs": [], |
||
| 453 | "source": [ |
||
| 454 | "dataset = dataset.train_test_split(test_size = 0.05)" |
||
| 455 | ] |
||
| 456 | }, |
||
| 457 | { |
||
| 458 | "cell_type": "code", |
||
| 459 | "execution_count": null, |
||
| 460 | "id": "49dffff3", |
||
| 461 | "metadata": {}, |
||
| 462 | "outputs": [], |
||
| 463 | "source": [ |
||
| 464 | "dataset.save_to_disk(DATA_PATH_SPLITS)" |
||
| 465 | ] |
||
| 466 | }, |
||
| 467 | { |
||
| 468 | "cell_type": "markdown", |
||
| 469 | "id": "2580b573", |
||
| 470 | "metadata": {}, |
||
| 471 | "source": [ |
||
| 472 | "# CONTINUE FROM HERE WHEN NOT THE FIRST RUN" |
||
| 473 | ] |
||
| 474 | }, |
||
| 475 | { |
||
| 476 | "cell_type": "markdown", |
||
| 477 | "id": "3c535b25", |
||
| 478 | "metadata": {}, |
||
| 479 | "source": [ |
||
| 480 | "### Load splits" |
||
| 481 | ] |
||
| 482 | }, |
||
| 483 | { |
||
| 484 | "cell_type": "code", |
||
| 485 | "execution_count": null, |
||
| 486 | "id": "acf777de", |
||
| 487 | "metadata": {}, |
||
| 488 | "outputs": [], |
||
| 489 | "source": [ |
||
| 490 | "token_cnt = pickle.load(PT_CNTS_PATH)" |
||
| 491 | ] |
||
| 492 | }, |
||
| 493 | { |
||
| 494 | "cell_type": "code", |
||
| 495 | "execution_count": null, |
||
| 496 | "id": "ab30780b", |
||
| 497 | "metadata": {}, |
||
| 498 | "outputs": [], |
||
| 499 | "source": [ |
||
| 500 | "dataset = datasets.load_from_disk(DATA_PATH_SPLITS)" |
||
| 501 | ] |
||
| 502 | }, |
||
| 503 | { |
||
| 504 | "cell_type": "code", |
||
| 505 | "execution_count": null, |
||
| 506 | "id": "3929e28a", |
||
| 507 | "metadata": {}, |
||
| 508 | "outputs": [], |
||
| 509 | "source": [ |
||
| 510 | "dataset" |
||
| 511 | ] |
||
| 512 | }, |
||
| 513 | { |
||
| 514 | "cell_type": "code", |
||
| 515 | "execution_count": null, |
||
| 516 | "id": "6fe3b048", |
||
| 517 | "metadata": {}, |
||
| 518 | "outputs": [], |
||
| 519 | "source": [ |
||
| 520 | "fprint(\"Total number of pts in train/test: {}/{}\".format(len(dataset['train']), len(dataset['test'])))" |
||
| 521 | ] |
||
| 522 | }, |
||
| 523 | { |
||
| 524 | "cell_type": "markdown", |
||
| 525 | "id": "2d7964a1", |
||
| 526 | "metadata": {}, |
||
| 527 | "source": [ |
||
| 528 | "# Filter to required type" |
||
| 529 | ] |
||
| 530 | }, |
||
| 531 | { |
||
| 532 | "cell_type": "code", |
||
| 533 | "execution_count": null, |
||
| 534 | "id": "1067d23e", |
||
| 535 | "metadata": {}, |
||
| 536 | "outputs": [], |
||
| 537 | "source": [ |
||
| 538 | "if \"ALL_TYPES\" not in TYPES:\n", |
||
| 539 | " print(\"FILTERING\")\n", |
||
| 540 | " dataset = filter_by_type(dataset, types_to_keep=TYPES, num_proc=NUM_PROC)" |
||
| 541 | ] |
||
| 542 | }, |
||
| 543 | { |
||
| 544 | "cell_type": "markdown", |
||
| 545 | "id": "58033ab4", |
||
| 546 | "metadata": {}, |
||
| 547 | "source": [ |
||
| 548 | "# Add Death token" |
||
| 549 | ] |
||
| 550 | }, |
||
| 551 | { |
||
| 552 | "cell_type": "code", |
||
| 553 | "execution_count": null, |
||
| 554 | "id": "9db46905", |
||
| 555 | "metadata": {}, |
||
| 556 | "outputs": [], |
||
| 557 | "source": [ |
||
| 558 | "pt2dod_timestamp = {str(k):v for k,v in pickle.load(PT_DOD_PATH).items()}\n", |
||
| 559 | "pt2death = {k:\"The patient has died\" for k in pt2dod_timestamp.keys()}\n", |
||
| 560 | "dataset = dataset.map(\n", |
||
| 561 | " lambda examples: add_to_stream(examples, pt2death, last=True, prefix=None, token_type='death'),\n", |
||
| 562 | " batched=True,\n", |
||
| 563 | " load_from_cache_file=False,\n", |
||
| 564 | " num_proc=NUM_PROC)" |
||
| 565 | ] |
||
| 566 | }, |
||
| 567 | { |
||
| 568 | "cell_type": "markdown", |
||
| 569 | "id": "b5c995e2", |
||
| 570 | "metadata": {}, |
||
| 571 | "source": [ |
||
| 572 | "# Bucket and split" |
||
| 573 | ] |
||
| 574 | }, |
||
| 575 | { |
||
| 576 | "cell_type": "code", |
||
| 577 | "execution_count": null, |
||
| 578 | "id": "3c019959", |
||
| 579 | "metadata": { |
||
| 580 | "scrolled": true |
||
| 581 | }, |
||
| 582 | "outputs": [], |
||
| 583 | "source": [ |
||
| 584 | "dataset = dataset.map(\n", |
||
| 585 | " lambda examples: bucket_concepts(examples, bucket_size_seconds=DAYS*24*60*60, duration_separator=False),\n", |
||
| 586 | " batched=True,\n", |
||
| 587 | " load_from_cache_file=False,\n", |
||
| 588 | " num_proc=NUM_PROC)" |
||
| 589 | ] |
||
| 590 | }, |
||
| 591 | { |
||
| 592 | "cell_type": "code", |
||
| 593 | "execution_count": null, |
||
| 594 | "id": "18804746", |
||
| 595 | "metadata": {}, |
||
| 596 | "outputs": [], |
||
| 597 | "source": [ |
||
| 598 | "dataset" |
||
| 599 | ] |
||
| 600 | }, |
||
| 601 | { |
||
| 602 | "cell_type": "markdown", |
||
| 603 | "id": "27383ac4", |
||
| 604 | "metadata": {}, |
||
| 605 | "source": [ |
||
| 606 | "## Trim long streams" |
||
| 607 | ] |
||
| 608 | }, |
||
| 609 | { |
||
| 610 | "cell_type": "code", |
||
| 611 | "execution_count": null, |
||
| 612 | "id": "460f779e", |
||
| 613 | "metadata": {}, |
||
| 614 | "outputs": [], |
||
| 615 | "source": [ |
||
| 616 | "from collections import defaultdict\n", |
||
| 617 | "lns = []\n", |
||
| 618 | "for _dataset in get_all_splits(dataset):\n", |
||
| 619 | " for stream in _dataset['stream']:\n", |
||
| 620 | " lns.append(len(stream))\n", |
||
| 621 | "pickle.dump(lns, PT_LNS_PATH)" |
||
| 622 | ] |
||
| 623 | }, |
||
| 624 | { |
||
| 625 | "cell_type": "code", |
||
| 626 | "execution_count": null, |
||
| 627 | "id": "e2c18011", |
||
| 628 | "metadata": {}, |
||
| 629 | "outputs": [], |
||
| 630 | "source": [ |
||
| 631 | "lns = pickle.load(PT_LNS_PATH)" |
||
| 632 | ] |
||
| 633 | }, |
||
| 634 | { |
||
| 635 | "cell_type": "code", |
||
| 636 | "execution_count": null, |
||
| 637 | "id": "f44e23e0", |
||
| 638 | "metadata": {}, |
||
| 639 | "outputs": [], |
||
| 640 | "source": [ |
||
| 641 | "len(lns)" |
||
| 642 | ] |
||
| 643 | }, |
||
| 644 | { |
||
| 645 | "cell_type": "code", |
||
| 646 | "execution_count": null, |
||
| 647 | "id": "9ebf1af4", |
||
| 648 | "metadata": {}, |
||
| 649 | "outputs": [], |
||
| 650 | "source": [ |
||
| 651 | "max(lns)" |
||
| 652 | ] |
||
| 653 | }, |
||
| 654 | { |
||
| 655 | "cell_type": "code", |
||
| 656 | "execution_count": null, |
||
| 657 | "id": "78d8b228", |
||
| 658 | "metadata": {}, |
||
| 659 | "outputs": [], |
||
| 660 | "source": [ |
||
| 661 | "max_len = int(np.percentile(lns, 95))\n", |
||
| 662 | "max_len" |
||
| 663 | ] |
||
| 664 | }, |
||
| 665 | { |
||
| 666 | "cell_type": "code", |
||
| 667 | "execution_count": null, |
||
| 668 | "id": "f50eaf63", |
||
| 669 | "metadata": { |
||
| 670 | "scrolled": true |
||
| 671 | }, |
||
| 672 | "outputs": [], |
||
| 673 | "source": [ |
||
| 674 | "fig = px.histogram(x=[x for x in lns if x < max_len and x > 5], labels={'x': 'length'})" |
||
| 675 | ] |
||
| 676 | }, |
||
| 677 | { |
||
| 678 | "cell_type": "code", |
||
| 679 | "execution_count": null, |
||
| 680 | "id": "67531287", |
||
| 681 | "metadata": {}, |
||
| 682 | "outputs": [], |
||
| 683 | "source": [ |
||
| 684 | "fig.write_html(\"./dataset-info/\" + RUN_NAME + \".html\")" |
||
| 685 | ] |
||
| 686 | }, |
||
| 687 | { |
||
| 688 | "cell_type": "code", |
||
| 689 | "execution_count": null, |
||
| 690 | "id": "5426144f", |
||
| 691 | "metadata": {}, |
||
| 692 | "outputs": [], |
||
| 693 | "source": [ |
||
| 694 | "dataset = filter_by_count(dataset, min_count=0, min_count_global=0, min_length=10, max_length=max_len, \n", |
||
| 695 | " num_proc=NUM_PROC, token_cnt=token_cnt)" |
||
| 696 | ] |
||
| 697 | }, |
||
| 698 | { |
||
| 699 | "cell_type": "code", |
||
| 700 | "execution_count": null, |
||
| 701 | "id": "71c14b2b", |
||
| 702 | "metadata": {}, |
||
| 703 | "outputs": [], |
||
| 704 | "source": [ |
||
| 705 | "dataset" |
||
| 706 | ] |
||
| 707 | }, |
||
| 708 | { |
||
| 709 | "cell_type": "markdown", |
||
| 710 | "id": "fb87c646", |
||
| 711 | "metadata": {}, |
||
| 712 | "source": [ |
||
| 713 | "## Split to max len" |
||
| 714 | ] |
||
| 715 | }, |
||
| 716 | { |
||
| 717 | "cell_type": "code", |
||
| 718 | "execution_count": null, |
||
| 719 | "id": "0ffc1da8", |
||
| 720 | "metadata": { |
||
| 721 | "scrolled": true |
||
| 722 | }, |
||
| 723 | "outputs": [], |
||
| 724 | "source": [ |
||
| 725 | "dataset = dataset.map(\n", |
||
| 726 | " lambda examples: split_stream(examples, max_seq_len=MAX_SEQ_LEN-32),\n", |
||
| 727 | " batched=True,\n", |
||
| 728 | " load_from_cache_file=False,\n", |
||
| 729 | " num_proc=NUM_PROC)" |
||
| 730 | ] |
||
| 731 | }, |
||
| 732 | { |
||
| 733 | "cell_type": "markdown", |
||
| 734 | "id": "577beded", |
||
| 735 | "metadata": {}, |
||
| 736 | "source": [ |
||
| 737 | "## Save again" |
||
| 738 | ] |
||
| 739 | }, |
||
| 740 | { |
||
| 741 | "cell_type": "code", |
||
| 742 | "execution_count": null, |
||
| 743 | "id": "9e18f654", |
||
| 744 | "metadata": {}, |
||
| 745 | "outputs": [], |
||
| 746 | "source": [ |
||
| 747 | "dataset.save_to_disk(ALMOST_PREPARED_DATASET_SPLIT_PATH)" |
||
| 748 | ] |
||
| 749 | }, |
||
| 750 | { |
||
| 751 | "cell_type": "code", |
||
| 752 | "execution_count": null, |
||
| 753 | "id": "da54723d", |
||
| 754 | "metadata": {}, |
||
| 755 | "outputs": [], |
||
| 756 | "source": [ |
||
| 757 | "dataset = datasets.load_from_disk(ALMOST_PREPARED_DATASET_SPLIT_PATH)" |
||
| 758 | ] |
||
| 759 | }, |
||
| 760 | { |
||
| 761 | "cell_type": "code", |
||
| 762 | "execution_count": null, |
||
| 763 | "id": "2b10396a", |
||
| 764 | "metadata": {}, |
||
| 765 | "outputs": [], |
||
| 766 | "source": [ |
||
| 767 | "dataset" |
||
| 768 | ] |
||
| 769 | }, |
||
| 770 | { |
||
| 771 | "cell_type": "markdown", |
||
| 772 | "id": "60b7eb32", |
||
| 773 | "metadata": {}, |
||
| 774 | "source": [ |
||
| 775 | "# Add DOD and TTD" |
||
| 776 | ] |
||
| 777 | }, |
||
| 778 | { |
||
| 779 | "cell_type": "code", |
||
| 780 | "execution_count": null, |
||
| 781 | "id": "d04702ae", |
||
| 782 | "metadata": { |
||
| 783 | "scrolled": true |
||
| 784 | }, |
||
| 785 | "outputs": [], |
||
| 786 | "source": [ |
||
| 787 | "pt2dob_timestamp = {str(k):v for k,v in pickle.load(PT_DOB_PATH).items()}\n", |
||
| 788 | "dataset = dataset.map(\n", |
||
| 789 | " lambda examples: add_age(examples, pt2dob_timestamp=pt2dob_timestamp),\n", |
||
| 790 | " batched=True,\n", |
||
| 791 | " load_from_cache_file=False,\n", |
||
| 792 | " num_proc=NUM_PROC)" |
||
| 793 | ] |
||
| 794 | }, |
||
| 795 | { |
||
| 796 | "cell_type": "code", |
||
| 797 | "execution_count": null, |
||
| 798 | "id": "79d57596", |
||
| 799 | "metadata": {}, |
||
| 800 | "outputs": [], |
||
| 801 | "source": [ |
||
| 802 | "\"\"\"\n", |
||
| 803 | "pt2dod_timestamp = {str(k):v for k,v in pickle.load(PT_DOD_PATH).items()}\n", |
||
| 804 | "# ADD time to die\n", |
||
| 805 | "dataset = dataset.map(\n", |
||
| 806 | " lambda examples: add_ttd(examples, pt2dod_timestamp=pt2dod_timestamp, ttd_normalizer=14 * 24 * 60 * 60),\n", |
||
| 807 | " batched=True,\n", |
||
| 808 | " load_from_cache_file=False,\n", |
||
| 809 | " num_proc=NUM_PROC)\n", |
||
| 810 | "\"\"\"" |
||
| 811 | ] |
||
| 812 | }, |
||
| 813 | { |
||
| 814 | "cell_type": "markdown", |
||
| 815 | "id": "312a702f", |
||
| 816 | "metadata": {}, |
||
| 817 | "source": [ |
||
| 818 | "### Another way for TTD" |
||
| 819 | ] |
||
| 820 | }, |
||
| 821 | { |
||
| 822 | "cell_type": "code", |
||
| 823 | "execution_count": null, |
||
| 824 | "id": "88aca6b0", |
||
| 825 | "metadata": {}, |
||
| 826 | "outputs": [], |
||
| 827 | "source": [ |
||
| 828 | "\"\"\"\n", |
||
| 829 | "# ADD time to die\n", |
||
| 830 | "dataset['train'] = dataset['train'].map(\n", |
||
| 831 | " lambda examples: add_ttd(examples, pt2dod_timestamp=pt2dod_timestamp, ttd_normalizer=14 * 24 * 60 * 60),\n", |
||
| 832 | " batched=True,\n", |
||
| 833 | " load_from_cache_file=False,\n", |
||
| 834 | " num_proc=NUM_PROC)\n", |
||
| 835 | "\n", |
||
| 836 | "dataset['test'] = dataset['test'].map(\n", |
||
| 837 | " lambda examples: add_ttd(examples, pt2dod_timestamp=pt2dod_timestamp, ttd_normalizer=14 * 24 * 60 * 60,\n", |
||
| 838 | " max_nttd=10, ttd_prob=1, duplicate_streams=True),\n", |
||
| 839 | " batched=True,\n", |
||
| 840 | " load_from_cache_file=False,\n", |
||
| 841 | " num_proc=NUM_PROC)\n", |
||
| 842 | "\"\"\"" |
||
| 843 | ] |
||
| 844 | }, |
||
| 845 | { |
||
| 846 | "cell_type": "markdown", |
||
| 847 | "id": "a520f26a", |
||
| 848 | "metadata": {}, |
||
| 849 | "source": [ |
||
| 850 | "# Add sex and ethnicity" |
||
| 851 | ] |
||
| 852 | }, |
||
| 853 | { |
||
| 854 | "cell_type": "code", |
||
| 855 | "execution_count": null, |
||
| 856 | "id": "fbcf3d29", |
||
| 857 | "metadata": { |
||
| 858 | "scrolled": true |
||
| 859 | }, |
||
| 860 | "outputs": [], |
||
| 861 | "source": [ |
||
| 862 | "# Add Sex\n", |
||
| 863 | "pt2sex = pickle.load(PT_SEX_PATH)\n", |
||
| 864 | "dataset = dataset.map(\n", |
||
| 865 | " lambda examples: add_to_stream(examples, pt2sex, last=False, prefix='<SEX>', token_type='sex'),\n", |
||
| 866 | " batched=True,\n", |
||
| 867 | " load_from_cache_file=False,\n", |
||
| 868 | " num_proc=NUM_PROC)" |
||
| 869 | ] |
||
| 870 | }, |
||
| 871 | { |
||
| 872 | "cell_type": "code", |
||
| 873 | "execution_count": null, |
||
| 874 | "id": "14005ffc", |
||
| 875 | "metadata": { |
||
| 876 | "scrolled": true |
||
| 877 | }, |
||
| 878 | "outputs": [], |
||
| 879 | "source": [ |
||
| 880 | "# Ethnicity\n", |
||
| 881 | "pt2ethnicity = pickle.load(PT_ETHNICITY_PATH)\n", |
||
| 882 | "dataset = dataset.map(\n", |
||
| 883 | " lambda examples: add_to_stream(examples, pt2ethnicity, last=False, prefix='<ETHNICITY>', token_type='ethnicity'),\n", |
||
| 884 | " batched=True,\n", |
||
| 885 | " load_from_cache_file=False,\n", |
||
| 886 | " num_proc=NUM_PROC)" |
||
| 887 | ] |
||
| 888 | }, |
||
| 889 | { |
||
| 890 | "cell_type": "markdown", |
||
| 891 | "id": "b0dd9e11", |
||
| 892 | "metadata": {}, |
||
| 893 | "source": [ |
||
| 894 | "# Final filter" |
||
| 895 | ] |
||
| 896 | }, |
||
| 897 | { |
||
| 898 | "cell_type": "code", |
||
| 899 | "execution_count": null, |
||
| 900 | "id": "bdb7f59b", |
||
| 901 | "metadata": {}, |
||
| 902 | "outputs": [], |
||
| 903 | "source": [ |
||
| 904 | "dataset = filter_by_count(dataset, min_count=None, min_count_global=None, min_length=10, num_proc=NUM_PROC)" |
||
| 905 | ] |
||
| 906 | }, |
||
| 907 | { |
||
| 908 | "cell_type": "markdown", |
||
| 909 | "id": "f1ce6130", |
||
| 910 | "metadata": {}, |
||
| 911 | "source": [ |
||
| 912 | "# Remove parents" |
||
| 913 | ] |
||
| 914 | }, |
||
| 915 | { |
||
| 916 | "cell_type": "code", |
||
| 917 | "execution_count": null, |
||
| 918 | "id": "47f86d59", |
||
| 919 | "metadata": {}, |
||
| 920 | "outputs": [], |
||
| 921 | "source": [ |
||
| 922 | "# Diseases\n", |
||
| 923 | "cuis = [token for token in cdb.config.linking['filters']['cuis'] if token in cdb.cui2names]\n", |
||
| 924 | "ch2parents = get_parents_map(cuis, cdb.addl_info['pt2ch'], depth=2)" |
||
| 925 | ] |
||
| 926 | }, |
||
| 927 | { |
||
| 928 | "cell_type": "code", |
||
| 929 | "execution_count": null, |
||
| 930 | "id": "64b4664f", |
||
| 931 | "metadata": {}, |
||
| 932 | "outputs": [], |
||
| 933 | "source": [ |
||
| 934 | "dataset = dataset.map(\n", |
||
| 935 | " lambda examples: remove_parents_from_stream(examples, ch2parents=ch2parents, separator='<SEP>'),\n", |
||
| 936 | " batched=True,\n", |
||
| 937 | " load_from_cache_file=False,\n", |
||
| 938 | " num_proc=NUM_PROC)" |
||
| 939 | ] |
||
| 940 | }, |
||
| 941 | { |
||
| 942 | "cell_type": "markdown", |
||
| 943 | "id": "0a3aa069", |
||
| 944 | "metadata": {}, |
||
| 945 | "source": [ |
||
| 946 | "## Add position IDs" |
||
| 947 | ] |
||
| 948 | }, |
||
| 949 | { |
||
| 950 | "cell_type": "code", |
||
| 951 | "execution_count": null, |
||
| 952 | "id": "0858f524", |
||
| 953 | "metadata": {}, |
||
| 954 | "outputs": [], |
||
| 955 | "source": [ |
||
| 956 | "dataset = dataset.map(\n", |
||
| 957 | " lambda examples: add_position_ids(examples, separators={'<SEP>', '<SEP-1>', '<SEP-7>' '<SEP-14>', '<SEP-30>', '<SEP-365>'}),\n", |
||
| 958 | " batched=True,\n", |
||
| 959 | " load_from_cache_file=False,\n", |
||
| 960 | " num_proc=NUM_PROC)" |
||
| 961 | ] |
||
| 962 | }, |
||
| 963 | { |
||
| 964 | "cell_type": "markdown", |
||
| 965 | "id": "cf641980", |
||
| 966 | "metadata": {}, |
||
| 967 | "source": [ |
||
| 968 | "# Get token_type2tokens" |
||
| 969 | ] |
||
| 970 | }, |
||
| 971 | { |
||
| 972 | "cell_type": "code", |
||
| 973 | "execution_count": null, |
||
| 974 | "id": "9bef3781", |
||
| 975 | "metadata": { |
||
| 976 | "scrolled": true |
||
| 977 | }, |
||
| 978 | "outputs": [], |
||
| 979 | "source": [ |
||
| 980 | "token_type2tokens = defaultdict(set)\n", |
||
| 981 | "total_cnt = 0\n", |
||
| 982 | "for _dataset in get_all_splits(dataset):\n", |
||
| 983 | " for stream in _dataset['stream']:\n", |
||
| 984 | " for example in stream:\n", |
||
| 985 | " token_type2tokens[example['token_type']].add(example['token'])\n", |
||
| 986 | " total_cnt += 1\n", |
||
| 987 | "token_type2tokens = dict(token_type2tokens)\n", |
||
| 988 | "pickle.dump(token_type2tokens, TOKEN_TYPES_PATH)\n", |
||
| 989 | "fprint(\"Total number of annotations: \", total_cnt)" |
||
| 990 | ] |
||
| 991 | }, |
||
| 992 | { |
||
| 993 | "cell_type": "code", |
||
| 994 | "execution_count": null, |
||
| 995 | "id": "8bb1f2a1", |
||
| 996 | "metadata": {}, |
||
| 997 | "outputs": [], |
||
| 998 | "source": [ |
||
| 999 | "pickle.dump(token_type2tokens, TOKEN_TYPES_PATH)\n", |
||
| 1000 | "fprint(\"Total number of annotations: \", total_cnt)" |
||
| 1001 | ] |
||
| 1002 | }, |
||
| 1003 | { |
||
| 1004 | "cell_type": "markdown", |
||
| 1005 | "id": "61c34003", |
||
| 1006 | "metadata": {}, |
||
| 1007 | "source": [ |
||
| 1008 | "# Cleanup stream and leave only what we need" |
||
| 1009 | ] |
||
| 1010 | }, |
||
| 1011 | { |
||
| 1012 | "cell_type": "code", |
||
| 1013 | "execution_count": null, |
||
| 1014 | "id": "33d5fbdb", |
||
| 1015 | "metadata": { |
||
| 1016 | "scrolled": true |
||
| 1017 | }, |
||
| 1018 | "outputs": [], |
||
| 1019 | "source": [ |
||
| 1020 | "dataset = dataset.map(\n", |
||
| 1021 | " lambda examples: cleanup_stream(examples, keep_time=True, keep_type=True, keep_position_ids=True,\n", |
||
| 1022 | " keep_context_representation=False),\n", |
||
| 1023 | " batched=True,\n", |
||
| 1024 | " load_from_cache_file=False,\n", |
||
| 1025 | " num_proc=NUM_PROC)" |
||
| 1026 | ] |
||
| 1027 | }, |
||
| 1028 | { |
||
| 1029 | "cell_type": "markdown", |
||
| 1030 | "id": "71ad6184", |
||
| 1031 | "metadata": {}, |
||
| 1032 | "source": [ |
||
| 1033 | "### Save" |
||
| 1034 | ] |
||
| 1035 | }, |
||
| 1036 | { |
||
| 1037 | "cell_type": "code", |
||
| 1038 | "execution_count": null, |
||
| 1039 | "id": "fc98c4f0", |
||
| 1040 | "metadata": {}, |
||
| 1041 | "outputs": [], |
||
| 1042 | "source": [ |
||
| 1043 | "dataset.save_to_disk(JUST_BEFORE_ENCODING_DATASET_SPLIT_PATH)" |
||
| 1044 | ] |
||
| 1045 | }, |
||
| 1046 | { |
||
| 1047 | "cell_type": "code", |
||
| 1048 | "execution_count": null, |
||
| 1049 | "id": "c2b2c1e1", |
||
| 1050 | "metadata": {}, |
||
| 1051 | "outputs": [], |
||
| 1052 | "source": [ |
||
| 1053 | "dataset = datasets.load_from_disk(JUST_BEFORE_ENCODING_DATASET_SPLIT_PATH)" |
||
| 1054 | ] |
||
| 1055 | }, |
||
| 1056 | { |
||
| 1057 | "cell_type": "code", |
||
| 1058 | "execution_count": null, |
||
| 1059 | "id": "41ac5775", |
||
| 1060 | "metadata": {}, |
||
| 1061 | "outputs": [], |
||
| 1062 | "source": [ |
||
| 1063 | "JUST_BEFORE_ENCODING_DATASET_SPLIT_PATH" |
||
| 1064 | ] |
||
| 1065 | }, |
||
| 1066 | { |
||
| 1067 | "cell_type": "code", |
||
| 1068 | "execution_count": null, |
||
| 1069 | "id": "38c09fd5", |
||
| 1070 | "metadata": {}, |
||
| 1071 | "outputs": [], |
||
| 1072 | "source": [ |
||
| 1073 | "# Total number of patients fater intial filtering\n", |
||
| 1074 | "train_len = len(dataset['train'])\n", |
||
| 1075 | "test_len = len(dataset['test'])\n", |
||
| 1076 | "fprint(\"Total number of pts in train: \", train_len)\n", |
||
| 1077 | "fprint(\"Total number of pts in test: \", test_len)\n", |
||
| 1078 | "fprint(\"Total number of pts: \", train_len + test_len)" |
||
| 1079 | ] |
||
| 1080 | }, |
||
| 1081 | { |
||
| 1082 | "cell_type": "code", |
||
| 1083 | "execution_count": null, |
||
| 1084 | "id": "29bcc0c4", |
||
| 1085 | "metadata": {}, |
||
| 1086 | "outputs": [], |
||
| 1087 | "source": [ |
||
| 1088 | "# Total number of annotations per type after filtering\n", |
||
| 1089 | "cnt_per_type_after = {}\n", |
||
| 1090 | "for _dataset in get_all_splits(dataset):\n", |
||
| 1091 | " for stream in _dataset['stream']:\n", |
||
| 1092 | " for cui in stream:\n", |
||
| 1093 | " if cat.cdb.cui2type_ids.get(cui, None):\n", |
||
| 1094 | " t = list(cat.cdb.cui2type_ids[cui])[0]\n", |
||
| 1095 | " cnt_per_type_after[t] = cnt_per_type_after.get(t, 0) + 1" |
||
| 1096 | ] |
||
| 1097 | }, |
||
| 1098 | { |
||
| 1099 | "cell_type": "code", |
||
| 1100 | "execution_count": null, |
||
| 1101 | "id": "208c24fb", |
||
| 1102 | "metadata": {}, |
||
| 1103 | "outputs": [], |
||
| 1104 | "source": [ |
||
| 1105 | "fprint(\"Total number of annotations per type: \\n\")\n", |
||
| 1106 | "for t in cnt_per_type_after:\n", |
||
| 1107 | " fprint(\"{:30}: {}\".format(cat.cdb.addl_info['type_id2name'][t].title(), cnt_per_type_after[t]))" |
||
| 1108 | ] |
||
| 1109 | }, |
||
| 1110 | { |
||
| 1111 | "cell_type": "markdown", |
||
| 1112 | "id": "b2b4d62e", |
||
| 1113 | "metadata": {}, |
||
| 1114 | "source": [ |
||
| 1115 | "# Make tokenizer" |
||
| 1116 | ] |
||
| 1117 | }, |
||
| 1118 | { |
||
| 1119 | "cell_type": "code", |
||
| 1120 | "execution_count": null, |
||
| 1121 | "id": "8cf2ec1a", |
||
| 1122 | "metadata": {}, |
||
| 1123 | "outputs": [], |
||
| 1124 | "source": [ |
||
| 1125 | "extra_tokenizer = None\n", |
||
| 1126 | "#extra_tokenizer = SimpleMapTokenizer.load(\"./data/time/models/slam_tokenizer_annotations_stream_phase2_1d_200_ALL_TYPES.pickle\")" |
||
| 1127 | ] |
||
| 1128 | }, |
||
| 1129 | { |
||
| 1130 | "cell_type": "code", |
||
| 1131 | "execution_count": null, |
||
| 1132 | "id": "e150f156", |
||
| 1133 | "metadata": {}, |
||
| 1134 | "outputs": [], |
||
| 1135 | "source": [ |
||
| 1136 | "token_type2tokens = pickle.load(TOKEN_TYPES_PATH)\n", |
||
| 1137 | "extra_concepts = None\n", |
||
| 1138 | "if extra_tokenizer is not None:\n", |
||
| 1139 | " extra_concepts = list(extra_tokenizer.tkn2id.keys())\n", |
||
| 1140 | "\n", |
||
| 1141 | " for k,v in extra_tokenizer.token_type2tokens.items():\n", |
||
| 1142 | " if k in token_type2tokens:\n", |
||
| 1143 | " token_type2tokens[k].update(extra_tokenizer.token_type2tokens[k])\n", |
||
| 1144 | " else:\n", |
||
| 1145 | " token_type2tokens[k] = extra_tokenizer.token_type2tokens[k]" |
||
| 1146 | ] |
||
| 1147 | }, |
||
| 1148 | { |
||
| 1149 | "cell_type": "code", |
||
| 1150 | "execution_count": null, |
||
| 1151 | "id": "767d8549", |
||
| 1152 | "metadata": {}, |
||
| 1153 | "outputs": [], |
||
| 1154 | "source": [ |
||
| 1155 | "_types = list(cdb.addl_info['type_id2name'].keys()) + list(token_type2tokens.keys())\n", |
||
| 1156 | "embeddings, tkn2id, id2tkn, = get_embeddings_for_tokens(dataset, cdb, context_type='xlong', types=_types,\n", |
||
| 1157 | " concepts=extra_concepts)" |
||
| 1158 | ] |
||
| 1159 | }, |
||
| 1160 | { |
||
| 1161 | "cell_type": "code", |
||
| 1162 | "execution_count": null, |
||
| 1163 | "id": "e3b9946e", |
||
| 1164 | "metadata": {}, |
||
| 1165 | "outputs": [], |
||
| 1166 | "source": [ |
||
| 1167 | "tkn2name = {tkn:cdb.get_name(tkn) for tkn in tkn2id.keys()}\n", |
||
| 1168 | "tokenizer = SimpleMapTokenizer(tkn2id=tkn2id, pad_id=tkn2id['<PAD>'], tkn2name=tkn2name,\n", |
||
| 1169 | " token_type2tokens=token_type2tokens, embeddings=embeddings,\n", |
||
| 1170 | " global_token_cnt=token_cnt, max_len=MAX_SEQ_LEN)" |
||
| 1171 | ] |
||
| 1172 | }, |
||
| 1173 | { |
||
| 1174 | "cell_type": "code", |
||
| 1175 | "execution_count": null, |
||
| 1176 | "id": "51b9cce4", |
||
| 1177 | "metadata": {}, |
||
| 1178 | "outputs": [], |
||
| 1179 | "source": [ |
||
| 1180 | "assert len(tokenizer.tkn2id) == len(tokenizer.id2tkn)\n", |
||
| 1181 | "assert len(tokenizer.embeddings) == len(tokenizer.id2tkn)\n", |
||
| 1182 | "assert len(tokenizer.tkn2name) == len(tokenizer.id2tkn)\n", |
||
| 1183 | "fprint(tokenizer.pad_id, tokenizer.id2tkn[tokenizer.pad_id])" |
||
| 1184 | ] |
||
| 1185 | }, |
||
| 1186 | { |
||
| 1187 | "cell_type": "code", |
||
| 1188 | "execution_count": null, |
||
| 1189 | "id": "031b224b", |
||
| 1190 | "metadata": {}, |
||
| 1191 | "outputs": [], |
||
| 1192 | "source": [ |
||
| 1193 | "len(tokenizer.tkn2name)" |
||
| 1194 | ] |
||
| 1195 | }, |
||
| 1196 | { |
||
| 1197 | "cell_type": "code", |
||
| 1198 | "execution_count": null, |
||
| 1199 | "id": "0295fe2d", |
||
| 1200 | "metadata": {}, |
||
| 1201 | "outputs": [], |
||
| 1202 | "source": [ |
||
| 1203 | "# save\n", |
||
| 1204 | "tokenizer.save(TOKENIZER_PATH)" |
||
| 1205 | ] |
||
| 1206 | }, |
||
| 1207 | { |
||
| 1208 | "cell_type": "code", |
||
| 1209 | "execution_count": null, |
||
| 1210 | "id": "f6cd9f04", |
||
| 1211 | "metadata": {}, |
||
| 1212 | "outputs": [], |
||
| 1213 | "source": [ |
||
| 1214 | "# Total number of different concepts after all filtering\n", |
||
| 1215 | "fprint(\"Total number of concepts after filtering: \", len(tokenizer.tkn2id))\n", |
||
| 1216 | "fprint(\"\")" |
||
| 1217 | ] |
||
| 1218 | }, |
||
| 1219 | { |
||
| 1220 | "cell_type": "code", |
||
| 1221 | "execution_count": null, |
||
| 1222 | "id": "b9f996f0", |
||
| 1223 | "metadata": {}, |
||
| 1224 | "outputs": [], |
||
| 1225 | "source": [ |
||
| 1226 | "# Total number annotations after all filtering\n", |
||
| 1227 | "fprint(\"Total number of annotations after filtering: \", sum([x for x in cnt_per_type_after.values()]))\n", |
||
| 1228 | "fprint(\"\")" |
||
| 1229 | ] |
||
| 1230 | }, |
||
| 1231 | { |
||
| 1232 | "cell_type": "markdown", |
||
| 1233 | "id": "862166a9", |
||
| 1234 | "metadata": {}, |
||
| 1235 | "source": [ |
||
| 1236 | "# Print number of different concepts per type after filtering" |
||
| 1237 | ] |
||
| 1238 | }, |
||
| 1239 | { |
||
| 1240 | "cell_type": "code", |
||
| 1241 | "execution_count": null, |
||
| 1242 | "id": "82ebd72c", |
||
| 1243 | "metadata": {}, |
||
| 1244 | "outputs": [], |
||
| 1245 | "source": [ |
||
| 1246 | "cnt_per_type = {}\n", |
||
| 1247 | "for cui in tkn2id:\n", |
||
| 1248 | " if cat.cdb.cui2type_ids.get(cui, ['Other']):\n", |
||
| 1249 | " t = list(cat.cdb.cui2type_ids.get(cui, ['Other']))[0]\n", |
||
| 1250 | " cnt_per_type[t] = cnt_per_type.get(t, 0) + 1\n", |
||
| 1251 | "fprint(\"Total number of <<different>> concepts per type after filtering\")\n", |
||
| 1252 | "for t in cnt_per_type:\n", |
||
| 1253 | " fprint(\"{:30}: {}\".format(cat.cdb.addl_info['type_id2name'].get(t, t).title(), cnt_per_type[t]))\n", |
||
| 1254 | "fprint(\"\")" |
||
| 1255 | ] |
||
| 1256 | }, |
||
| 1257 | { |
||
| 1258 | "cell_type": "markdown", |
||
| 1259 | "id": "19cf388d", |
||
| 1260 | "metadata": {}, |
||
| 1261 | "source": [ |
||
| 1262 | "# Create global tokenizer" |
||
| 1263 | ] |
||
| 1264 | }, |
||
| 1265 | { |
||
| 1266 | "cell_type": "code", |
||
| 1267 | "execution_count": null, |
||
| 1268 | "id": "ac9c9839", |
||
| 1269 | "metadata": {}, |
||
| 1270 | "outputs": [], |
||
| 1271 | "source": [ |
||
| 1272 | "_types = list(cdb.addl_info['type_id2name'].keys()) + list(token_type2tokens.keys())\n", |
||
| 1273 | "concepts = list(cat.config.linking['filters']['cuis'])\n", |
||
| 1274 | "embeddings, tkn2id, id2tkn, = get_embeddings_for_tokens(dataset, cdb, context_type='xlong', types=_types, concepts=concepts)" |
||
| 1275 | ] |
||
| 1276 | }, |
||
| 1277 | { |
||
| 1278 | "cell_type": "code", |
||
| 1279 | "execution_count": null, |
||
| 1280 | "id": "eabbbf88", |
||
| 1281 | "metadata": {}, |
||
| 1282 | "outputs": [], |
||
| 1283 | "source": [ |
||
| 1284 | "tkn2name = {tkn:cdb.get_name(tkn) for tkn in tkn2id.keys()}\n", |
||
| 1285 | "tokenizer = SimpleMapTokenizer(tkn2id=tkn2id, pad_id=tkn2id['<PAD>'], tkn2name=tkn2name,\n", |
||
| 1286 | " token_type2tokens=token_type2tokens, embeddings=embeddings,\n", |
||
| 1287 | " global_token_cnt=token_cnt, max_len=MAX_SEQ_LEN)" |
||
| 1288 | ] |
||
| 1289 | }, |
||
| 1290 | { |
||
| 1291 | "cell_type": "code", |
||
| 1292 | "execution_count": null, |
||
| 1293 | "id": "22833eec", |
||
| 1294 | "metadata": {}, |
||
| 1295 | "outputs": [], |
||
| 1296 | "source": [ |
||
| 1297 | "tokenizer.save(BASE_TOKENIZER_PATH)" |
||
| 1298 | ] |
||
| 1299 | }, |
||
| 1300 | { |
||
| 1301 | "cell_type": "markdown", |
||
| 1302 | "id": "4c06d09b", |
||
| 1303 | "metadata": {}, |
||
| 1304 | "source": [ |
||
| 1305 | "# Convert tokens to IDs" |
||
| 1306 | ] |
||
| 1307 | }, |
||
| 1308 | { |
||
| 1309 | "cell_type": "code", |
||
| 1310 | "execution_count": null, |
||
| 1311 | "id": "864df946", |
||
| 1312 | "metadata": {}, |
||
| 1313 | "outputs": [], |
||
| 1314 | "source": [ |
||
| 1315 | "if FROM_BASE:\n", |
||
| 1316 | " print(\"USING BASE TOKENIZER\")\n", |
||
| 1317 | " TOKENIZER_PATH = BASE_TOKENIZER_PATH" |
||
| 1318 | ] |
||
| 1319 | }, |
||
| 1320 | { |
||
| 1321 | "cell_type": "code", |
||
| 1322 | "execution_count": null, |
||
| 1323 | "id": "e4540c4a", |
||
| 1324 | "metadata": {}, |
||
| 1325 | "outputs": [], |
||
| 1326 | "source": [ |
||
| 1327 | "tokenizer = SimpleMapTokenizer.load(TOKENIZER_PATH)" |
||
| 1328 | ] |
||
| 1329 | }, |
||
| 1330 | { |
||
| 1331 | "cell_type": "code", |
||
| 1332 | "execution_count": null, |
||
| 1333 | "id": "548e8e3b", |
||
| 1334 | "metadata": {}, |
||
| 1335 | "outputs": [], |
||
| 1336 | "source": [ |
||
| 1337 | "encoded_dataset = dataset.map(\n", |
||
| 1338 | " lambda examples: tokenizer.encode(examples),\n", |
||
| 1339 | " batched=True,\n", |
||
| 1340 | " remove_columns=['stream'],\n", |
||
| 1341 | " load_from_cache_file=False,\n", |
||
| 1342 | " num_proc=NUM_PROC)" |
||
| 1343 | ] |
||
| 1344 | }, |
||
| 1345 | { |
||
| 1346 | "cell_type": "code", |
||
| 1347 | "execution_count": null, |
||
| 1348 | "id": "8718ae76", |
||
| 1349 | "metadata": {}, |
||
| 1350 | "outputs": [], |
||
| 1351 | "source": [ |
||
| 1352 | "encoded_dataset.save_to_disk(PREPARED_DATASET_SPLIT_PATH)" |
||
| 1353 | ] |
||
| 1354 | }, |
||
| 1355 | { |
||
| 1356 | "cell_type": "code", |
||
| 1357 | "execution_count": null, |
||
| 1358 | "id": "340dd11d", |
||
| 1359 | "metadata": {}, |
||
| 1360 | "outputs": [], |
||
| 1361 | "source": [ |
||
| 1362 | "PREPARED_DATASET_SPLIT_PATH" |
||
| 1363 | ] |
||
| 1364 | }, |
||
| 1365 | { |
||
| 1366 | "cell_type": "code", |
||
| 1367 | "execution_count": null, |
||
| 1368 | "id": "e512ab8c", |
||
| 1369 | "metadata": {}, |
||
| 1370 | "outputs": [], |
||
| 1371 | "source": [ |
||
| 1372 | "TOKENIZER_PATH" |
||
| 1373 | ] |
||
| 1374 | }, |
||
| 1375 | { |
||
| 1376 | "cell_type": "markdown", |
||
| 1377 | "id": "6715bb47", |
||
| 1378 | "metadata": {}, |
||
| 1379 | "source": [ |
||
| 1380 | "# Test is all OK" |
||
| 1381 | ] |
||
| 1382 | }, |
||
| 1383 | { |
||
| 1384 | "cell_type": "code", |
||
| 1385 | "execution_count": null, |
||
| 1386 | "id": "60881f72", |
||
| 1387 | "metadata": {}, |
||
| 1388 | "outputs": [], |
||
| 1389 | "source": [ |
||
| 1390 | "encoded_dataset = datasets.load_from_disk(PREPARED_DATASET_SPLIT_PATH)" |
||
| 1391 | ] |
||
| 1392 | }, |
||
| 1393 | { |
||
| 1394 | "cell_type": "code", |
||
| 1395 | "execution_count": null, |
||
| 1396 | "id": "2f3c066f", |
||
| 1397 | "metadata": {}, |
||
| 1398 | "outputs": [], |
||
| 1399 | "source": [ |
||
| 1400 | "dataset = datasets.load_from_disk(JUST_BEFORE_ENCODING_DATASET_SPLIT_PATH)" |
||
| 1401 | ] |
||
| 1402 | }, |
||
| 1403 | { |
||
| 1404 | "cell_type": "code", |
||
| 1405 | "execution_count": null, |
||
| 1406 | "id": "6dab3686", |
||
| 1407 | "metadata": {}, |
||
| 1408 | "outputs": [], |
||
| 1409 | "source": [ |
||
| 1410 | "tokenizer = SimpleMapTokenizer.load(TOKENIZER_PATH)" |
||
| 1411 | ] |
||
| 1412 | }, |
||
| 1413 | { |
||
| 1414 | "cell_type": "code", |
||
| 1415 | "execution_count": null, |
||
| 1416 | "id": "035a55fb", |
||
| 1417 | "metadata": {}, |
||
| 1418 | "outputs": [], |
||
| 1419 | "source": [ |
||
| 1420 | "encoded_dataset" |
||
| 1421 | ] |
||
| 1422 | }, |
||
| 1423 | { |
||
| 1424 | "cell_type": "code", |
||
| 1425 | "execution_count": null, |
||
| 1426 | "id": "b89f84af", |
||
| 1427 | "metadata": {}, |
||
| 1428 | "outputs": [], |
||
| 1429 | "source": [ |
||
| 1430 | "dataset" |
||
| 1431 | ] |
||
| 1432 | }, |
||
| 1433 | { |
||
| 1434 | "cell_type": "code", |
||
| 1435 | "execution_count": null, |
||
| 1436 | "id": "21d66a24", |
||
| 1437 | "metadata": {}, |
||
| 1438 | "outputs": [], |
||
| 1439 | "source": [ |
||
| 1440 | "ind = 1096" |
||
| 1441 | ] |
||
| 1442 | }, |
||
| 1443 | { |
||
| 1444 | "cell_type": "code", |
||
| 1445 | "execution_count": null, |
||
| 1446 | "id": "664dd895", |
||
| 1447 | "metadata": {}, |
||
| 1448 | "outputs": [], |
||
| 1449 | "source": [ |
||
| 1450 | "from datetime import datetime" |
||
| 1451 | ] |
||
| 1452 | }, |
||
| 1453 | { |
||
| 1454 | "cell_type": "code", |
||
| 1455 | "execution_count": null, |
||
| 1456 | "id": "49b1cbc1", |
||
| 1457 | "metadata": { |
||
| 1458 | "scrolled": true |
||
| 1459 | }, |
||
| 1460 | "outputs": [], |
||
| 1461 | "source": [ |
||
| 1462 | "[cdb.get_name(x) for x in dataset['train'][ind]['stream']]" |
||
| 1463 | ] |
||
| 1464 | }, |
||
| 1465 | { |
||
| 1466 | "cell_type": "code", |
||
| 1467 | "execution_count": null, |
||
| 1468 | "id": "9ec0594e", |
||
| 1469 | "metadata": { |
||
| 1470 | "scrolled": true |
||
| 1471 | }, |
||
| 1472 | "outputs": [], |
||
| 1473 | "source": [ |
||
| 1474 | "for ty, p, t, c in zip(encoded_dataset['train'][ind]['token_type'], encoded_dataset['train'][ind]['position_ids'], encoded_dataset['train'][ind]['time'], tokenizer.convert_ids2tokens(encoded_dataset['train'][ind]['input_ids'])):\n", |
||
| 1475 | " print(datetime.fromtimestamp(t), p, \"{:20}\".format(ty), c)" |
||
| 1476 | ] |
||
| 1477 | }, |
||
| 1478 | { |
||
| 1479 | "cell_type": "code", |
||
| 1480 | "execution_count": null, |
||
| 1481 | "id": "19003ff8", |
||
| 1482 | "metadata": {}, |
||
| 1483 | "outputs": [], |
||
| 1484 | "source": [ |
||
| 1485 | "encoded_dataset['train'][ind]['patient_id']" |
||
| 1486 | ] |
||
| 1487 | }, |
||
| 1488 | { |
||
| 1489 | "cell_type": "code", |
||
| 1490 | "execution_count": null, |
||
| 1491 | "id": "1fc5460d", |
||
| 1492 | "metadata": {}, |
||
| 1493 | "outputs": [], |
||
| 1494 | "source": [ |
||
| 1495 | "ds_info.close()" |
||
| 1496 | ] |
||
| 1497 | }, |
||
| 1498 | { |
||
| 1499 | "cell_type": "markdown", |
||
| 1500 | "id": "e5a3cc65", |
||
| 1501 | "metadata": {}, |
||
| 1502 | "source": [ |
||
| 1503 | "# Preapre for Foresight" |
||
| 1504 | ] |
||
| 1505 | }, |
||
| 1506 | { |
||
| 1507 | "cell_type": "code", |
||
| 1508 | "execution_count": null, |
||
| 1509 | "id": "2dec8a19", |
||
| 1510 | "metadata": {}, |
||
| 1511 | "outputs": [], |
||
| 1512 | "source": [ |
||
| 1513 | "ind = 32330" |
||
| 1514 | ] |
||
| 1515 | }, |
||
| 1516 | { |
||
| 1517 | "cell_type": "code", |
||
| 1518 | "execution_count": null, |
||
| 1519 | "id": "2821af66", |
||
| 1520 | "metadata": {}, |
||
| 1521 | "outputs": [], |
||
| 1522 | "source": [ |
||
| 1523 | "import json" |
||
| 1524 | ] |
||
| 1525 | }, |
||
| 1526 | { |
||
| 1527 | "cell_type": "code", |
||
| 1528 | "execution_count": null, |
||
| 1529 | "id": "55457df7", |
||
| 1530 | "metadata": { |
||
| 1531 | "scrolled": true |
||
| 1532 | }, |
||
| 1533 | "outputs": [], |
||
| 1534 | "source": [ |
||
| 1535 | "[cdb.get_name(x) for x in dataset['train'][ind]['stream']]" |
||
| 1536 | ] |
||
| 1537 | }, |
||
| 1538 | { |
||
| 1539 | "cell_type": "code", |
||
| 1540 | "execution_count": null, |
||
| 1541 | "id": "196eaf9d", |
||
| 1542 | "metadata": { |
||
| 1543 | "scrolled": true |
||
| 1544 | }, |
||
| 1545 | "outputs": [], |
||
| 1546 | "source": [ |
||
| 1547 | "for i, c in enumerate(dataset['train'][ind]['stream']):\n", |
||
| 1548 | " print(i)\n", |
||
| 1549 | " if i > 20 and c not in dataset['train'][ind]['stream'][0:i]:\n", |
||
| 1550 | " print(i, c, cdb.get_name(c))" |
||
| 1551 | ] |
||
| 1552 | }, |
||
| 1553 | { |
||
| 1554 | "cell_type": "code", |
||
| 1555 | "execution_count": null, |
||
| 1556 | "id": "4a159ae3", |
||
| 1557 | "metadata": {}, |
||
| 1558 | "outputs": [], |
||
| 1559 | "source": [ |
||
| 1560 | "out = []\n", |
||
| 1561 | "for i, cui in enumerate(dataset['train'][ind]['stream'][:161]):\n", |
||
| 1562 | " d = {\n", |
||
| 1563 | " 'id': cui,\n", |
||
| 1564 | " 'label': cdb.get_name(cui),\n", |
||
| 1565 | " 'count': 1000000,\n", |
||
| 1566 | " 'name': cdb.get_name(cui),\n", |
||
| 1567 | " 'cui': cui,\n", |
||
| 1568 | " 'saliency': 0,\n", |
||
| 1569 | " 'uid': i\n", |
||
| 1570 | " }\n", |
||
| 1571 | " out.append(d)" |
||
| 1572 | ] |
||
| 1573 | }, |
||
| 1574 | { |
||
| 1575 | "cell_type": "code", |
||
| 1576 | "execution_count": null, |
||
| 1577 | "id": "97f40a06", |
||
| 1578 | "metadata": {}, |
||
| 1579 | "outputs": [], |
||
| 1580 | "source": [ |
||
| 1581 | "json.dump(out, open(\"./data/tmp/timeline_example_1.json\", 'w'))" |
||
| 1582 | ] |
||
| 1583 | }, |
||
| 1584 | { |
||
| 1585 | "cell_type": "code", |
||
| 1586 | "execution_count": null, |
||
| 1587 | "id": "da6e079d", |
||
| 1588 | "metadata": {}, |
||
| 1589 | "outputs": [], |
||
| 1590 | "source": [ |
||
| 1591 | "len(out)" |
||
| 1592 | ] |
||
| 1593 | }, |
||
| 1594 | { |
||
| 1595 | "cell_type": "code", |
||
| 1596 | "execution_count": null, |
||
| 1597 | "id": "e2730480", |
||
| 1598 | "metadata": {}, |
||
| 1599 | "outputs": [], |
||
| 1600 | "source": [ |
||
| 1601 | "out" |
||
| 1602 | ] |
||
| 1603 | } |
||
| 1604 | ], |
||
| 1605 | "metadata": { |
||
| 1606 | "kernelspec": { |
||
| 1607 | "display_name": "Python 3 (ipykernel)", |
||
| 1608 | "language": "python", |
||
| 1609 | "name": "python3" |
||
| 1610 | }, |
||
| 1611 | "language_info": { |
||
| 1612 | "codemirror_mode": { |
||
| 1613 | "name": "ipython", |
||
| 1614 | "version": 3 |
||
| 1615 | }, |
||
| 1616 | "file_extension": ".py", |
||
| 1617 | "mimetype": "text/x-python", |
||
| 1618 | "name": "python", |
||
| 1619 | "nbconvert_exporter": "python", |
||
| 1620 | "pygments_lexer": "ipython3", |
||
| 1621 | "version": "3.8.0" |
||
| 1622 | } |
||
| 1623 | }, |
||
| 1624 | "nbformat": 4, |
||
| 1625 | "nbformat_minor": 5 |
||
| 1626 | } |