Coverage for nlp_manager/parse_substance

1"""

2crate_anon/nlp_manager/parse_substance_misuse.py

4===============================================================================

7 Created by Rudolf Cardinal (rnc1001@cam.ac.uk).

9 This file is part of CRATE.

11 CRATE is free software: you can redistribute it and/or modify

12 it under the terms of the GNU General Public License as published by

13 the Free Software Foundation, either version 3 of the License, or

14 (at your option) any later version.

16 CRATE is distributed in the hope that it will be useful,

17 but WITHOUT ANY WARRANTY; without even the implied warranty of

18 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the

19 GNU General Public License for more details.

21 You should have received a copy of the GNU General Public License

22 along with CRATE. If not, see <https://www.gnu.org/licenses/>.

24===============================================================================

26**Python regex-based NLP processors for substance misuse.**

28"""

30import logging

31from typing import Any, Dict, Generator, List, Optional, Tuple

33from crate_anon.common.regex_helpers import (

34 at_wb_start_end,

35 noncapture_group,

36 optional_named_capture_group,

37 optional_noncapture_group,

38 regex_or,

39 WORD_BOUNDARY,

40)

41from crate_anon.nlp_manager.nlp_definition import NlpDefinition

42from crate_anon.nlp_manager.number import to_float

43from crate_anon.nlp_manager.regex_func import (

44 compile_regex,

45 compile_regex_dict,

46 get_regex_dict_match,

47 get_regex_dict_search,

48)

49from crate_anon.nlp_manager.regex_parser import (

50 common_tense,

51 EVER,

52 FN_CONTENT,

53 FN_END,

54 FN_RELATION,

55 FN_RELATION_TEXT,

56 FN_START,

57 FN_TENSE,

58 FN_TENSE_TEXT,

59 FN_UNITS,

60 FN_VALUE_TEXT,

61 FN_VARIABLE_NAME,

62 FN_VARIABLE_TEXT,

63 GROUP_NAME_QUANTITY,

64 GROUP_NAME_RELATION,

65 GROUP_NAME_TENSE,

66 GROUP_NAME_UNITS,

67 GROUP_NAME_VALUE,

68 GROUP_NUMBER_WHOLE_EXPRESSION,

69 make_simple_numeric_regex,

70 NumericalResultParser,

71 PAST,

72 PRESENT,

73 ValidatorBase,

74)

75from crate_anon.nlp_manager.regex_units import (

76 ALCOHOL,

77 DAYS_PER_WEEK,

78 UK_ALCOHOL_UNITS_PER_DAY,

79 UK_ALCOHOL_UNITS_PER_MONTH,

80 UK_ALCOHOL_UNITS_PER_WEEK,

81 UK_ALCOHOL_UNITS_PER_YEAR,

82 WEEKS_PER_MONTH_APPROX,

83 WEEKS_PER_YEAR_APPROX,

84)

86log = logging.getLogger(__name__)

89# =============================================================================

90# Alcohol

91# =============================================================================

94class AlcoholUnits(NumericalResultParser):

95 """

96 SUBSTANCE MISUSE.

98 Alcohol consumption, specified explicitly as (UK) units per day or per

99 week, or via non-numeric references to not drinking any.

100

101 - Output is in UK units per week. A UK unit is 10 ml of ethanol [#f1]_ [#f2]_.

102 UK NHS guidelines used to be "per week" and remain broadly week-based [#f1]_.

103 - It doesn't attempt any understanding of other alcohol descriptions (e.g.

104 "pints of beer", "glasses of wine", "bottles of vodka") so is expected to

105 apply where a clinician has converted a (potentially mixed) alcohol

106 description to a units-per-week calculation.

107

108 .. [#f1] https://www.nhs.uk/live-well/alcohol-advice/calculating-alcohol-units/,

109 accessed 2023-01-18.

110 .. [#f2] https://en.wikipedia.org/wiki/Unit_of_alcohol

111 """ # noqa: E501

112

113 # There are no relevant Read codes for alcohol consumption in

114 # v3ReadCode_PBCL.xlsx.

115

116 # -------------------------------------------------------------------------

117 # Regex building for tense-related statements

118 # -------------------------------------------------------------------------

119

120 # All these are verbose regexes, so don't omit \s+ for whitespace!

121 PAST_ADVERBS = (

122 "formerly",

123 "once",

124 "peak",

125 "previously",

126 "was",

127 )

128 PAST_ADVERBS_RE = noncapture_group(regex_or(*PAST_ADVERBS))

129 DOES_NOT = r"does\s*n[o'’]t" # does not, doesn't

130 PRESENT_ADVERBS = (

131 r"at \s+ present",

132 r"currently",

133 r"has \s+ been",

134 r"now",

135 r"nowadays",

136 r"presently",

137 r"these \s+ days",

138 DOES_NOT,

139 )

140 PRESENT_ADVERBS_RE = noncapture_group(regex_or(*PRESENT_ADVERBS))

141 TEMPORAL_WORDS = tuple(

142 at_wb_start_end(x) for x in PAST_ADVERBS + PRESENT_ADVERBS

143 )

144 TEMPORAL = noncapture_group(regex_or(*TEMPORAL_WORDS))

145 OPT_TEMPORAL = optional_noncapture_group(regex_or(*TEMPORAL_WORDS))

146

147 NEVER = "never"

148 # "Never" is both temporal and negating and thus fiddly. We do *not*

149 # include it in standard temporal words, or a statement about "has never

150 # drunk >100 u/w" would be misinterpreted as positive.

151

152 # -------------------------------------------------------------------------

153 # Regex building for drinking alcohol (and when)

154 # -------------------------------------------------------------------------

155

156 DRINKING_PAST = (

157 # Past infinitive: she used to drink

158 r"\b used \s+ to \s+ drink \b",

159 # Imperfect tense: she [adverb] drank

160 rf"\b (?: {PAST_ADVERBS_RE} \s+ )? drank \b",

161 # Perfect tense: has drunk

162 rf"\b has (?: {PAST_ADVERBS_RE} \s+ )? drunk \b",

163 # Past continuous tense: he was [adverb] drinking

164 # Also abbreviated past continuous tense: previously drinking

165 rf"\b {PAST_ADVERBS_RE} \s+ drinking \b",

166 )

167 # We don't allow the adverbs by themselves, to avoid something that isn't

168 # explicitly about alcohol or drinking, e.g. "[insulin] currently 6

169 # units/day".

170 DRINKING_PRESENT = (

171 # Present tense: he [adverb] drinks

172 rf"\b (?: {PRESENT_ADVERBS_RE} \s+)? drinks \b",

173 # Present continuous tense: he is [adverb] drinking

174 rf"\b (?: is \s+)? (?: {PRESENT_ADVERBS_RE} \s+)? drinking \b",

175 )

176 DRINKING_PAST_PRESENT = DRINKING_PAST + DRINKING_PRESENT

177 DRINKING = noncapture_group(regex_or(*DRINKING_PAST_PRESENT))

178 OPT_DRINKING = optional_noncapture_group(regex_or(*DRINKING_PAST_PRESENT))

179 ALCOHOL_PM_CONSUMPTION = rf"{ALCOHOL} (?: \s+ consumption \b)?"

180 ALC = noncapture_group(ALCOHOL_PM_CONSUMPTION)

181 OPT_ALC = optional_noncapture_group(ALCOHOL_PM_CONSUMPTION)

182

183 # BRK: requires some sort of wordbreak or whitespace, but also disposes of

184 # junk like some punctuation (e.g. "previously: none" versus "previously

185 # none") and words like "at" (e.g. in "drinking at X units/week").

186 BRK = noncapture_group(

187 regex_or(

188 r"\s* : \s*", # colon +/- whitespace

189 r"\s* \b at \b \s*", # "at" +/- whitespace

190 r"\s+", # whitespace

191 WORD_BOUNDARY, # other word break

192 )

193 )

194

195 # Move from more to less specific, or the less specific will capture first.

196 ALCOHOL_DRINKING = rf"""

197 {WORD_BOUNDARY}

198 # Alcohol drinking:

199 (?:

200 # 1. ... DRINKING ... [ALC] ...

201 {OPT_TEMPORAL} {BRK}

202 {DRINKING} {BRK}

203 {OPT_TEMPORAL} {BRK}

204 {OPT_ALC} {BRK}

205 {OPT_TEMPORAL}

206 |

207 # 2. ... ALC ... [DRINKING] ...

208 {OPT_TEMPORAL} {BRK}

209 {ALC} {BRK}

210 {OPT_TEMPORAL} {BRK}

211 {OPT_DRINKING} {BRK}

212 {OPT_TEMPORAL}

213 )

214 {WORD_BOUNDARY}

215 """

216

217 _drinking_tense_dict = {} # type: Dict[str, str]

218 for _past in DRINKING_PAST + PAST_ADVERBS:

219 _drinking_tense_dict[_past] = PAST

220 for _present in DRINKING_PRESENT + PRESENT_ADVERBS:

221 _drinking_tense_dict[_present] = PRESENT

222 TENSE_PAST_PRESENT_LOOKUP = compile_regex_dict(_drinking_tense_dict)

223 TENSE_NEVER_LOOKUP = compile_regex_dict({NEVER: EVER})

224

225 # -------------------------------------------------------------------------

226 # Regex building for "drinking alcohol at X units per week"

227 # -------------------------------------------------------------------------

228

229 # A temporal suffix allows e.g. "drinking X units/week previously".

230 GROUP_NAME_SUFFIX = "suffix"

231 group_suffix = r"\b \s*" + optional_named_capture_group(

232 TEMPORAL, GROUP_NAME_SUFFIX

233 )

234 REGEX_ALCOHOL_UNITS = (

235 make_simple_numeric_regex(

236 quantity=ALCOHOL_DRINKING,

237 units=regex_or(

238 UK_ALCOHOL_UNITS_PER_DAY,

239 UK_ALCOHOL_UNITS_PER_WEEK,

240 UK_ALCOHOL_UNITS_PER_MONTH, # perhaps unusual!

241 UK_ALCOHOL_UNITS_PER_YEAR, # perhaps unusual!

242 ),

243 units_optional=False,

244 )

245 + group_suffix

246 )

247

248 # -------------------------------------------------------------------------

249 # Regex building for "no alcohol" statements

250 # -------------------------------------------------------------------------

251

252 ABSTINENT = r"\b abstin[ae]nt \b" # "abstinent", or typo "abstinant"

253 NONE = noncapture_group(

254 WORD_BOUNDARY

255 + noncapture_group(

256 regex_or(

257 "0",

258 rf"{ABSTINENT} (?: \s+ from \b )?",

259 NEVER,

260 "no",

261 "none",

262 "zero",

263 )

264 )

265 + WORD_BOUNDARY

266 )

267 TEETOTAL = noncapture_group(

268 r"\b te[ea][-]?total(?:l?er)? \b",

269 )

270 DOES_NOT_DRINK = noncapture_group(

271 regex_or(

272 rf"\b {DOES_NOT} \s+ drink \b",

273 rf"\b has \s+ {NEVER} \s+ drunk \b",

274 )

275 )

276 OPT_TEMPORAL_AND_OR_DRINKING_BRK = (

277 f"{OPT_TEMPORAL} {BRK} {OPT_DRINKING} {BRK} {OPT_TEMPORAL} {BRK}"

278 )

279 NO_ALCOHOL = rf"""

280 {WORD_BOUNDARY}

281 # "No alcohol" statements.

282 # Temporal modifiers might be found in all sorts of places.

283 (?:

284 # 1. [DRINKING] ... ALC ... [DRINKING] ... NONE ...

285 {OPT_TEMPORAL_AND_OR_DRINKING_BRK}

286 {ALC} {BRK}

287 {OPT_TEMPORAL_AND_OR_DRINKING_BRK}

288 {NONE} {BRK}

289 {OPT_TEMPORAL_AND_OR_DRINKING_BRK}

290 |

291 # 2. NONE ... ALC (e.g. "never alcohol")

292 {OPT_TEMPORAL_AND_OR_DRINKING_BRK}

293 {NONE} {BRK}

294 {OPT_TEMPORAL_AND_OR_DRINKING_BRK}

295 {ALC} {BRK}

296 {OPT_TEMPORAL_AND_OR_DRINKING_BRK}

297 |

298 # 3. "has never drunk... alcohol", etc.

299 {DOES_NOT_DRINK} {BRK} {ALC} {BRK}

300 |

301 # 4. "teetotal" with typos

302 {TEETOTAL}

303 # ... but not just "drinking... none" (could be water etc.)

304 )

305 {WORD_BOUNDARY}

306 """

307

308 # -------------------------------------------------------------------------

309 # Other class variables

310 # -------------------------------------------------------------------------

311

312 NAME = "AlcoholUnits"

313 PREFERRED_UNIT_COLUMN = "value_uk_units_per_week"

314 UNIT_MAPPING = {

315 UK_ALCOHOL_UNITS_PER_WEEK: 1, # preferred unit

316 UK_ALCOHOL_UNITS_PER_DAY: DAYS_PER_WEEK, # 1 unit/day -> 7 units/week

317 UK_ALCOHOL_UNITS_PER_MONTH: 1 / WEEKS_PER_MONTH_APPROX,

318 UK_ALCOHOL_UNITS_PER_YEAR: 1 / WEEKS_PER_YEAR_APPROX,

319 }

320

321 # -------------------------------------------------------------------------

322 # Init

323 # -------------------------------------------------------------------------

324

325 def __init__(

326 self,

327 nlpdef: Optional[NlpDefinition],

328 cfg_processor_name: Optional[str],

329 commit: bool = False,

330 ) -> None:

331 # see documentation above

332 super().__init__(

333 nlpdef=nlpdef,

334 cfg_processor_name=cfg_processor_name,

335 variable=self.NAME,

336 target_unit=self.PREFERRED_UNIT_COLUMN,

337 regex_str_for_debugging=self.REGEX_ALCOHOL_UNITS,

338 commit=commit,

339 )

340 self.compiled_regex_alcohol = compile_regex(self.REGEX_ALCOHOL_UNITS)

341 self.units_to_factor = compile_regex_dict(self.UNIT_MAPPING)

342 self.compiled_regex_no_alcohol = compile_regex(self.NO_ALCOHOL)

343

344 # -------------------------------------------------------------------------

345 # Parse

346 # -------------------------------------------------------------------------

347

348 def parse(

349 self, text: str, debug: bool = False

350 ) -> Generator[Tuple[str, Dict[str, Any]], None, None]:

351 """

352 Parse for two regexes which operate slightly differently.

353 """

354 if not text:

355 return

356 yield from self.parse_alcohol_units(text, debug)

357 yield from self.parse_alcohol_none(text, debug)

358

359 def parse_alcohol_units(

360 self, text: str, debug: bool = False

361 ) -> Generator[Tuple[str, Dict[str, Any]], None, None]:

362 """

363 We amend SimpleNumericalResultParser.parse() to deal with tense a bit

364 better (e.g. "used to drink"). Comments from that version not repeated.

365 That version also shortened a bit since we guarantee some aspects of

366 the flags.

367 """

368 for m in self.compiled_regex_alcohol.finditer(text):

369 startpos = m.start()

370 endpos = m.end()

371 matching_text = m.group(GROUP_NUMBER_WHOLE_EXPRESSION)

372 variable_text = m.group(GROUP_NAME_QUANTITY)

373 tense_text = m.group(GROUP_NAME_TENSE)

374 relation_text = m.group(GROUP_NAME_RELATION)

375 value_text = m.group(GROUP_NAME_VALUE)

376 units = m.group(GROUP_NAME_UNITS)

377 suffix_text = m.group(self.GROUP_NAME_SUFFIX)

378

379 value_in_target_units = None

380 if units:

381 matched_unit, multiple_or_fn = get_regex_dict_match(

382 units, self.units_to_factor

383 )

384 if not matched_unit:

385 continue

386 # MODIFIED: no need to check callable(multiple_or_fn); always

387 # no

388 value_in_target_units = to_float(value_text) * multiple_or_fn

389 # MODIFIED: no need to check self.assume_preferred_unit (we never

390 # assume that here)

391

392 # MODIFIED: no need to check self.take_absolute (always yes)

393 if value_in_target_units is not None:

394 value_in_target_units = abs(value_in_target_units)

395

396 tense, relation = common_tense(tense_text, relation_text)

397

398 # MODIFIED: Extra bit here to detect tense information in a

399 # different place:

400 for temporal_info in (variable_text, suffix_text):

401 if tense:

402 break

403 tense = self._get_tense(temporal_info)

404 if tense:

405 tense_text = temporal_info

406

407 # Back to the previous code:

408 result = {

409 FN_VARIABLE_NAME: self.variable,

410 FN_CONTENT: matching_text,

411 FN_START: startpos,

412 FN_END: endpos,

413 FN_VARIABLE_TEXT: variable_text,

414 FN_RELATION_TEXT: relation_text,

415 FN_RELATION: relation,

416 FN_VALUE_TEXT: value_text,

417 FN_UNITS: units,

418 self.target_unit: value_in_target_units,

419 FN_TENSE_TEXT: tense_text,

420 FN_TENSE: tense,

421 }

422 if debug:

423 log.debug(f"Match {m} for {text!r} -> {result}")

424 yield self.tablename, result

425

426 def parse_alcohol_none(

427 self, text: str, debug: bool = False

428 ) -> Generator[Tuple[str, Dict[str, Any]], None, None]:

429 """

430 Deal with references to not drinking any alcohol (except those referred

431 to as e.g. "0 units per week", which will be picked up by the

432 units-per-week function -- that will be rare!).

433 """

434 for m in self.compiled_regex_no_alcohol.finditer(text):

435 startpos = m.start()

436 endpos = m.end()

437 matching_text = m.group(GROUP_NUMBER_WHOLE_EXPRESSION)

438 tense = self._get_tense(matching_text)

439 tense_text = matching_text if tense else None

440

441 result = {

442 FN_VARIABLE_NAME: self.variable,

443 FN_CONTENT: matching_text,

444 FN_START: startpos,

445 FN_END: endpos,

446 FN_VARIABLE_TEXT: matching_text,

447 FN_RELATION_TEXT: None,

448 FN_RELATION: None,

449 FN_VALUE_TEXT: matching_text,

450 FN_UNITS: None,

451 self.target_unit: 0, # zero units

452 FN_TENSE_TEXT: tense_text,

453 FN_TENSE: tense,

454 }

455 if debug:

456 log.debug(f"Match {m} for {text!r} -> {result}")

457 yield self.tablename, result

458

459 def _get_tense(self, text: str) -> Optional[str]:

460 """

461 Find a tense indicator and return the corresponding text, or None.

462 """

463 # We deal with "never" first because otherwise "never drank" may hit

464 # "[optional_stuff] drank" and be classified as the past tense.

465 _, tense = get_regex_dict_search(text, self.TENSE_NEVER_LOOKUP)

466 if not tense:

467 _, tense = get_regex_dict_search(

468 text, self.TENSE_PAST_PRESENT_LOOKUP

469 )

470 return tense

471

472 # -------------------------------------------------------------------------

473 # Test

474 # -------------------------------------------------------------------------

475

476 def test(self, verbose: bool = False) -> None:

477 # docstring in parent class

478 # Test via e.g.:

479 # pytest -k SubstanceMisuseTests # self-tests

480 # crate_run_crate_nlp_demo - --processors AlcoholUnits # interactive

481 no_results = []

482 six_no_tense = [{self.target_unit: 6, FN_TENSE: None}]

483 six_past = [{self.target_unit: 6, FN_TENSE: PAST}]

484 six_present = [{self.target_unit: 6, FN_TENSE: PRESENT}]

485 six_per_day_present = [

486 {self.target_unit: 6 * DAYS_PER_WEEK, FN_TENSE: PRESENT}

487 ]

488 six_per_month_present = [

489 {self.target_unit: 6 / WEEKS_PER_MONTH_APPROX, FN_TENSE: PRESENT}

490 ]

491 six_per_year_present = [

492 {self.target_unit: 6 / WEEKS_PER_YEAR_APPROX, FN_TENSE: PRESENT}

493 ]

494 under_6_present = [

495 {self.target_unit: 6, FN_RELATION: "<", FN_TENSE: PRESENT}

496 ]

497 over_200_present = [

498 {self.target_unit: 200, FN_RELATION: ">", FN_TENSE: PRESENT}

499 ]

500 no_alcohol_no_tense = [{self.target_unit: 0, FN_TENSE: None}]

501 no_alcohol_past = [{self.target_unit: 0, FN_TENSE: PAST}]

502 no_alcohol_present = [{self.target_unit: 0, FN_TENSE: PRESENT}]

503 no_alcohol_ever = [{self.target_unit: 0, FN_TENSE: EVER}]

504 self.detailed_test_multiple(

505 [

506 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

507 # No results expected:

508 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

509 ("Alcohol", no_results),

510 ("He used to drink like a fish", no_results),

511 ("[e.g. insulin] currently 6 units per week", no_results),

512 ("[e.g. insulin] previously 6 units per week", no_results),

513 ("[could be insulin] peak 6 u/w", no_results),

514 ("[!] methylalcohol 6 u/w", no_results),

515 ("[not starts with no] Alcohol: not explored", no_results),

516 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

517 # Value with no tense:

518 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

519 ("Alcohol 6 u/w", six_no_tense),

520 ("Alcohol - 6 u/w", six_no_tense),

521 ("EtOH = 6 u/w", six_no_tense),

522 ("EtOH = 6 u/wk", six_no_tense),

523 ("Alcohol (units/week): 6", six_no_tense),

524 ("Ethanol 6 units/week", six_no_tense),

525 ("[not international but] alcohol 6 IU/week", six_no_tense),

526 ("alcohol 6 I.U./week", six_no_tense),

527 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

528 # Past tense:

529 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

530 ("Alcohol: was 6 u/w", six_past), # other tenses fail (= good)

531 ("Alcohol: formerly 6 u/w", six_past),

532 ("Alcohol: previously 6 u/w", six_past),

533 ("Alcohol: once 6 u/w", six_past),

534 ("Alcohol: peak 6 u/w", six_past),

535 ("Used to drink 6 u/w", six_past),

536 ("Peak drinking 6 u/w", six_past),

537 ("Peak alcohol consumption: 6 u/w", six_past),

538 ("Drank 6 u/w", six_past),

539 ("Formerly drank 6 u/w", six_past),

540 ("Previously drank 6 u/w", six_past),

541 ("Was drinking 6 u/w", six_past),

542 ("Was previously drinking 6 u/w", six_past),

543 ("Was formerly drinking 6 u/w", six_past),

544 ("Alcohol: formerly 6 u/w", six_past),

545 ("Alcohol: previously 6 u/w", six_past),

546 ("Alcohol: 6 u/w previously", six_past),

547 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

548 # Present tense:

549 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

550 ("Drinks 6 units per week", six_present),

551 ("Drinks 6 alcohol units per week", six_present),

552 ("Drinks 6 UK units per week", six_present),

553 ("Drinks 6 UK alcohol units per week", six_present),

554 ("[silly] Drinks 6 UK alcohol IU per week", six_present),

555 ("Drinks 6 units/d", six_per_day_present),

556 ("Drinks 6 units/dy", six_per_day_present),

557 ("Drinks 6 units/day", six_per_day_present),

558 ("Currently drinks 6 units per week", six_present),

559 ("These days drinks 6 units per week", six_present),

560 ("Now drinks 6 units per week", six_present),

561 ("Nowadays drinks 6 units per week", six_present),

562 ("Drinking 6 units per week", six_present),

563 ("Currently drinking 6 units per week", six_present),

564 ("Presently drinking 6 units per week", six_present),

565 ("Alcohol: currently 6 u/w", six_present),

566 ("Alcohol: presently 6 u/w", six_present),

567 ("In terms of alcohol she drinks 6 units/week", six_present),

568 ("Has been drinking 6 units per week", six_present),

569 ("Drinks 6 units per month", six_per_month_present),

570 ("Drinks 6 units per year", six_per_year_present),

571 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

572 # Inequalities:

573 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

574 ("Alcohol: presently less than 6 u/w", under_6_present),

575 ("Alcohol: presently under 6 u/w", under_6_present),

576 ("Alcohol: presently >200 u/w", over_200_present),

577 ("Alcohol: currently more than 200 u/w", over_200_present),

578 ("Alcohol: currently over 200 u/w", over_200_present),

579 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

580 # References to not drinking -- no tense:

581 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

582 ("Alcohol: none", no_alcohol_no_tense),

583 ("Teetotal", no_alcohol_no_tense),

584 ("Tee-total", no_alcohol_no_tense), # typo

585 ("Teetotaller", no_alcohol_no_tense),

586 ("Teetotaler", no_alcohol_no_tense), # typo

587 ("Abstinent from alcohol", no_alcohol_no_tense),

588 ("Alcohol: abstinent", no_alcohol_no_tense),

589 ("Alcohol: abstinant", no_alcohol_no_tense), # typo

590 ("Alcohol: zero", no_alcohol_no_tense),

591 ("Alcohol: 0", no_alcohol_no_tense),

592 ("Alcohol: no", no_alcohol_no_tense),

593 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

594 # References to not drinking -- past tense:

595 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

596 ("Alcohol: was abstinent", no_alcohol_past),

597 ("Alcohol: previously abstinent", no_alcohol_past),

598 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

599 # References to not drinking -- present tense:

600 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

601 ("Alcohol: has been abstinent", no_alcohol_present),

602 ("Alcohol: currently abstinent", no_alcohol_present),

603 ("Alcohol: currently none", no_alcohol_present),

604 ("Drinks no alcohol", no_alcohol_present),

605 ("Drinks zero alcohol", no_alcohol_present),

606 ("Does not drink alcohol", no_alcohol_present),

607 ("Doesn't drink alcohol", no_alcohol_present),

608 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

609 # References to not drinking -- ever:

610 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

611 ("Has never drunk alcohol", no_alcohol_ever),

612 ("Never drank alcohol", no_alcohol_ever),

613 ("Alcohol: never", no_alcohol_ever),

614 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

615 # Vague references to not drinking, not interpreted:

616 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

617 ("Has not drunk alcohol", no_results),

618 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

619 # Potential teetotal statements, but very tricky to be sure:

620 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

621 ("Doesn't drink [coffee]", no_results),

622 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

623 # Distractors:

624 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

625 ("Lemonade, which he does not drink.", no_results),

626 ],

627 verbose=verbose,

628 )

629

630

631class AlcoholUnitsValidator(ValidatorBase):

632 """

633 Validator for AlcoholUnits (see help for explanation).

634 """

635

636 @classmethod

637 def get_variablename_regexstrlist(cls) -> Tuple[str, List[str]]:

638 # We're very broad here:

639 return AlcoholUnits.NAME, [

640 regex_or(

641 ALCOHOL,

642 r"\b dr[iau]nk ", # drink/drank/drunk plus any ending

643 AlcoholUnits.ABSTINENT,

644 AlcoholUnits.TEETOTAL,

645 )

646 ]

647

648

649# =============================================================================

650# All classes in this module

651# =============================================================================

652

653ALL_SUBSTANCE_MISUSE_NLP_AND_VALIDATORS = [

654 (AlcoholUnits, AlcoholUnitsValidator),

655]

Coverage for nlp_manager/parse_substance_misuse.py: 97%

124 statements