File size: 48,501 Bytes
0bb52ff | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 | """
hv-reader
=========
The reading experience, in one call.
Given a text, produce a ReadingProfile: pace, memory, passes, slip, and
wall — the five axes that describe what it is like to read a text.
This unifies seven component models:
hv-tempo → pace
hv-forget → memory
hv-ttu → ttu_s (total time)
hv-fold → passes
hv-slip → slip
hv-hunger → hunger (internal, feeds slip and wall)
hv-wall → wall
The profile is the artifact. The axes are the readings. The signature is
what predicts whether a text gets finished.
Pure stdlib. No dependencies.
Author: zeechimp
License: Apache-2.0
"""
from __future__ import annotations
import argparse
import json
import math
import os
import re
import sys
from collections import Counter, defaultdict
from dataclasses import dataclass, asdict, field
from typing import Any, Dict, List, Optional, Set, Tuple
# ============================================================================
# Lexicons
# ============================================================================
COMMON_WORDS = frozenset("""
the be to of and a in that have i it for not on with he as you do at this
but his by from they we say her she or an will my one all would there their
what so up out if about who get which go me when make can like time no just
him know take people into year your good some could them see other than then
now look only come its over think also back after use two how our work first
well way even new want because any these give day most us is are was were
been being has had having does did doing will would shall should can could
may might must man woman child water fire earth air sun moon star light dark
hand head eye ear mouth nose foot leg arm body face heart mind life death
food bread milk meat fish tree flower grass leaf root seed farm field hill
mountain river sea lake boat ship road street city town house room door
window wall floor roof bed chair table book page word line letter number
name place thing part side end start middle top bottom front back left
right high low long short big small old new young hot cold wet dry clean
dirty light heavy soft hard fast slow easy true false good bad happy sad
love hate fear hope help hurt win lose give take send bring buy sell pay
cost money price work play run walk jump sit stand sleep wake eat drink
cook wash read write speak hear see feel know think learn teach ask
answer tell show hide open close push pull carry hold drop throw catch
break fix build make do try use move turn stop start keep leave stay wait
meet join save spend show thank want wish walk stop continue morning
climb row push pull begin finish start end remain rest return arrive
depart leave enter exit follow lead sit stand lie rise fall drop
still quiet calm slow fast soft loud bright dark warm cool fresh clean
never always often sometimes rarely usually speak spoke spoken
take took taken give gave given see saw seen know knew known
think thought thought come came come go went gone say said said
tell told told find found found hold held held bring brought brought
buy bought bought teach taught taught catch caught caught build built built
send sent sent spend spent lose lost lost lead led led meet met met
read read read write wrote written run ran run swim swam swum
thing things word words time times year years day days
man men woman women child children person people
place places work works way ways life lives hand hands
eye eyes part parts end ends line lines side sides
name names head heads house houses friend friends
family families group groups country countries
world worlds city cities school schools
""".split())
ABSTRACT_SUFFIXES = (
"tion", "sion", "ism", "ity", "ness", "ance", "ence",
"ship", "hood", "ment", "ology", "itude", "acy",
)
SUBORDINATORS = frozenset("""
which that because although though while whereas since when if unless
provided assuming given whenever wherever whoever whichever
""".split())
HEDGES = frozenset("""
may might maybe perhaps possibly probably typically usually often
generally roughly approximately about somewhat rather
""".split())
CONDITIONALS = frozenset("""
if when unless provided assuming suppose supposing
""".split())
NEGATIONS = frozenset("""
not no never none without cannot can't don't doesn't won't
isn't aren't wasn't weren't nor neither
""".split())
BE_FORMS = frozenset("""
is are was were be been being am
""".split())
REENTRY_MARKERS = frozenset("""
above below previous preceding following
aforementioned noted mentioned discussed described
stated referred earlier later
""".split())
ARTICLES = frozenset("""
the this that these those such said aforementioned
""".split())
UNIQUE_REFERENTS = frozenset("""
sun moon earth world sky ground horizon
morning afternoon evening night noon midnight dawn dusk
""".split())
CERTAINTY_MARKERS = frozenset("""
definitely definitively certainly obviously clearly undoubtedly
unquestionably absolutely surely plainly evidently undeniably
unmistakably decidedly categorically conclusively decisively
resolutely proves proven prove proved proof impossible must always
never guaranteed
""".split())
CONCLUSION_MARKERS = (
"therefore", "thus", "hence", "consequently", "accordingly",
"it follows that", "we conclude", "we can conclude",
"this shows", "this demonstrates", "this proves",
"in conclusion", "as a result",
)
CONTRAST_MARKERS = (
"however", "but", "yet", "although", "though", "nevertheless",
"nonetheless", "conversely", "on the contrary", "in contrast",
"on the other hand", "by contrast", "notwithstanding",
"despite this", "even so",
)
EVIDENCE_MARKERS = (
"according to", "studies show", "studies suggest",
"research shows", "research suggests",
"data show", "data suggest", "evidence indicates",
"we measured", "we observed", "we found",
"for example", "for instance", "specifically", "namely",
"for one", "in fact", "as measured",
)
QUESTION_RAISERS = (
"why", "how", "whether", "what caused", "the reason",
"unclear", "unknown", "remains to be determined",
"remains unclear", "puzzling", "mysterious",
"unexplained", "open question", "puzzle",
)
ANSWER_MARKERS = (
"because", "since", "as a result", "due to", "explained by",
"the reason is", "this explains", "the cause", "attributable to",
"results from", "arises from", "the mechanism is",
)
DIRECTION_GROUPS = {
"up": frozenset("""
increase increases increased increasing rise rises rose risen
grow grows grew grown growth expand expands expanded expansion
raise raises raised raising improve improves improved improving
gain gains gained gaining positive higher highest more most
upward up climb climbs climbed climbing
""".split()),
"down": frozenset("""
decrease decreases decreased decreasing fall falls fell fallen
shrink shrinks shrank shrunk contract contracts contracted
lower lowers lowered lowering reduce reduces reduced reducing
worsen worsens worsened worsening lose loses lost losing
negative lower lowest less least downward down decline declines
declined declining drop drops dropped dropping diminish
""".split()),
"cause": frozenset("""
cause causes caused causing produce produces produced producing
create creates created creating induce induces induced
trigger triggers triggered triggering generate
""".split()),
"prevent": frozenset("""
prevent prevents prevented preventing avoid avoids avoided
block blocks blocked blocking inhibit inhibits inhibited
prohibit prohibits prohibited prohibiting stop stops stopped
""".split()),
"support": frozenset("""
support supports supported supporting confirm confirms confirmed
agree agrees agreed approve approves approved accept accepts accepted
affirm affirms affirmed
""".split()),
"oppose": frozenset("""
oppose opposes opposed opposing deny denies denied contradict
contradicts contradicted disagree disagrees disagreed
reject rejects rejected refuse refuses refused
""".split()),
}
GROUP_OPPOSITES = frozenset([
("up", "down"), ("down", "up"),
("cause", "prevent"), ("prevent", "cause"),
("support", "oppose"), ("oppose", "support"),
])
DISCOURSE_MARKERS = frozenset("""
therefore thus hence consequently accordingly however but yet although
though nevertheless nonetheless conversely meanwhile similarly
moreover furthermore additionally
""".split())
# ============================================================================
# Regexes
# ============================================================================
_WORD_RE = re.compile(r"[A-Za-z][A-Za-z'\-]*")
_SENT_SPLIT_RE = re.compile(r"(?<=[.!?])\s+(?=[A-Z\"'(])")
_NUMBER_RE = re.compile(r"\b\d+(?:[.,]\d+)*\b")
_STANDALONE_DEMON_RE = re.compile(
r'^\s*(this|that|these|those)\s*[.,!?;:]*\s*$', re.IGNORECASE
)
# ============================================================================
# Tokenization helpers
# ============================================================================
def _stem(w: str) -> str:
w = w.lower()
if len(w) <= 4:
return w
for suffix in ("ingly", "edly", "ing", "ed", "ly", "es", "s"):
if w.endswith(suffix) and len(w) - len(suffix) >= 3:
base = w[: -len(suffix)]
if len(base) >= 2 and base[-1] == base[-2] and base[-1] not in "aeiou":
base = base[:-1]
return base
return w
def _words(text: str) -> List[str]:
return _WORD_RE.findall(text)
def _content_words(text: str) -> List[str]:
return [
w.lower() for w in _words(text)
if w.lower() not in COMMON_WORDS
and w.lower() not in DISCOURSE_MARKERS
and len(w) >= 3
]
def _sentences(text: str) -> List[str]:
return [s.strip() for s in _SENT_SPLIT_RE.split(text) if s.strip()]
def _is_common(w: str) -> bool:
lw = w.lower()
if lw in COMMON_WORDS:
return True
return _stem(lw) in COMMON_WORDS
def _is_rare(w: str) -> bool:
return len(w) >= 7 and not _is_common(w)
def _has_conclusion_marker(text: str) -> bool:
low = text.lower().strip()
for m in CONCLUSION_MARKERS:
if low.startswith(m):
return True
if f" {m} " in f" {low} ":
return True
return False
def _has_contrast_marker(text: str) -> Tuple[bool, str]:
low = text.lower().strip()
for m in CONTRAST_MARKERS:
if low.startswith(m):
return True, m
if f" {m} " in f" {low} ":
return True, m
return False, ""
def _has_evidence_marker(text: str) -> bool:
low = text.lower()
return any(m in low for m in EVIDENCE_MARKERS)
# ============================================================================
# Config
# ============================================================================
@dataclass
class HVReaderConfig:
# Pace (hv-tempo weights)
baseline_wpm: float = 220.0
w_sentence_len_excess: float = 0.40
w_clause_rate: float = 0.08
w_rare_rate: float = 0.70
w_abstract_rate: float = 0.40
w_digit_rate: float = 0.30
w_negation_rate: float = 0.40
w_hedge_rate: float = 0.40
w_conditional_rate: float = 0.60
w_passive_rate: float = 0.30
w_list_bonus: float = -0.50
max_log_slowdown: float = 1.5
# Memory (hv-forget)
memory_target_days: float = 7.0
memory_stability_base: float = 1.0
memory_stability_density: float = 5.0
memory_stability_rare: float = 3.0
memory_stability_salience: float = 2.0
# Fold (hv-fold)
fold_demon_weight: float = 2.0
fold_definite_only: float = 0.5
fold_reentry: float = 1.0
fold_forward: float = 0.5
fold_resolved: float = 0.1
exempt_first_sentence: bool = False
# Slip (hv-slip)
monotone_window: int = 3
slip_monotone_weight: float = 0.4
slip_repetition_weight: float = 0.3
slip_absence_weight: float = 0.3
# Hunger (hv-hunger)
hunger_question_weight: float = 1.0
hunger_answer_weight: float = 1.0
# Wall (hv-wall)
wall_threshold: float = 0.35
certainty_scale: float = 6.0
hedge_scale: float = 5.0
contradiction_scale: float = 0.4
specificity_scale: float = 5.0
gap_confidence_jump_weight: float = 0.5
contrast_bonus: float = 0.15
contradiction_min: float = 0.30
overclaim_min: float = 0.30
gap_min: float = 0.25
# Reporting
min_sentence_words: int = 3
top_reasons: int = 3
version: str = "0.1.1"
# ============================================================================
# Report dataclasses
# ============================================================================
@dataclass
class SpanProfile:
id: int
text: str
span: Tuple[int, int]
n_words: int
wpm: float
slowdown: float
density: float
stability_days: float
retention_week: float
fold_load: float
slip: float
hunger_delta: float
hunger: float
wall: float
wall_type: str
reasons: List[str] = field(default_factory=list)
def to_dict(self) -> dict:
return {
"id": self.id,
"text": self.text,
"span": list(self.span),
"n_words": self.n_words,
"wpm": round(self.wpm, 2),
"slowdown": round(self.slowdown, 3),
"density": round(self.density, 4),
"stability_days": round(self.stability_days, 3),
"retention_week": round(self.retention_week, 4),
"fold_load": round(self.fold_load, 3),
"slip": round(self.slip, 3),
"hunger_delta": round(self.hunger_delta, 3),
"hunger": round(self.hunger, 3),
"wall": round(self.wall, 4),
"wall_type": self.wall_type,
"reasons": list(self.reasons),
}
@dataclass
class ReadingProfile:
text: str
n_sentences: int
n_words: int
pace: float
memory: float
passes: float
slip: float
wall: float
ttu_s: float
mean_wpm: float
hunger_final: float
spans: List[SpanProfile]
slowest_span: Optional[SpanProfile]
wall_span: Optional[SpanProfile]
summary: str
def to_dict(self) -> dict:
return {
"text": self.text,
"n_sentences": self.n_sentences,
"n_words": self.n_words,
"profile": {
"pace": round(self.pace, 4),
"memory": round(self.memory, 4),
"passes": round(self.passes, 4),
"slip": round(self.slip, 4),
"wall": round(self.wall, 4),
},
"ttu_s": round(self.ttu_s, 2),
"mean_wpm": round(self.mean_wpm, 2),
"hunger_final": round(self.hunger_final, 3),
"slowest_span_id": self.slowest_span.id if self.slowest_span else None,
"wall_span_id": self.wall_span.id if self.wall_span else None,
"summary": self.summary,
"spans": [s.to_dict() for s in self.spans],
}
# ============================================================================
# Pace (hv-tempo)
# ============================================================================
def _count_list_markers(text: str) -> int:
bullets = len(re.findall(r"(?:^|\n)\s*[-*•]\s+\S", text))
numbered = re.findall(r"(?:^|[\s;:.])\d+[.)]\s+\S", text)
n_numbered = len(numbered) if len(numbered) >= 2 else 0
return max(bullets, n_numbered)
def _pace_features(sentence: str) -> Dict[str, float]:
words = _words(sentence)
n_words = len(words)
if n_words == 0:
return {k: 0.0 for k in [
"n_words", "mean_sentence_len", "sentence_len_signed",
"clauses_per_sentence", "rare_rate", "abstract_rate",
"digit_rate", "negation_rate", "hedge_rate",
"conditional_rate", "passive_rate", "list_rate",
]}
n_sentences = max(1, len([s for s in _SENT_SPLIT_RE.split(sentence) if s.strip()]))
mean_len = n_words / n_sentences
signed_len = (mean_len - 15.0) / 10.0
n_clauses = (
sentence.count(",") + sentence.count(";") + sentence.count(":")
+ sum(1 for w in words if w.lower() in SUBORDINATORS)
)
clauses_per_sentence = n_clauses / n_sentences
n_rare = sum(1 for w in words if len(w) >= 7 and not _is_common(w))
rare_rate = n_rare / n_words
n_abstract = sum(
1 for w in words
if len(w) > 5 and w.lower().endswith(ABSTRACT_SUFFIXES)
)
abstract_rate = n_abstract / n_words
digit_rate = len(_NUMBER_RE.findall(sentence)) / n_words
negation_rate = sum(1 for w in words if w.lower() in NEGATIONS) / n_words
hedge_rate = sum(1 for w in words if w.lower() in HEDGES) / n_words
conditional_rate = sum(1 for w in words if w.lower() in CONDITIONALS) / n_words
n_passive = 0
for i, w in enumerate(words):
if w.lower() in BE_FORMS and i + 1 < len(words):
nxt = words[i + 1].lower()
if (nxt.endswith("ed") and len(nxt) > 3) or nxt in (
"gone", "seen", "written", "taken", "made", "known",
"found", "given", "held", "sent", "left", "kept",
):
n_passive += 1
passive_rate = n_passive / n_sentences
n_list = _count_list_markers(sentence)
list_rate = n_list / n_sentences
return {
"n_words": float(n_words),
"mean_sentence_len": mean_len,
"sentence_len_signed": signed_len,
"clauses_per_sentence": clauses_per_sentence,
"rare_rate": rare_rate,
"abstract_rate": abstract_rate,
"digit_rate": digit_rate,
"negation_rate": negation_rate,
"hedge_rate": hedge_rate,
"conditional_rate": conditional_rate,
"passive_rate": passive_rate,
"list_rate": list_rate,
}
def _pace_slowdown(
f: Dict[str, float], cfg: HVReaderConfig
) -> Tuple[float, Dict[str, float]]:
contrib = {
"sentence_length": cfg.w_sentence_len_excess * f["sentence_len_signed"],
"clause_density": cfg.w_clause_rate * f["clauses_per_sentence"],
"rare_words": cfg.w_rare_rate * f["rare_rate"],
"abstract_terms": cfg.w_abstract_rate * f["abstract_rate"],
"numerals": cfg.w_digit_rate * f["digit_rate"],
"negation": cfg.w_negation_rate * f["negation_rate"],
"hedging": cfg.w_hedge_rate * f["hedge_rate"],
"conditionals": cfg.w_conditional_rate * f["conditional_rate"],
"passive_voice": cfg.w_passive_rate * f["passive_rate"],
"list_structure": cfg.w_list_bonus * f["list_rate"],
}
log_slowdown = sum(contrib.values())
log_slowdown = max(
-cfg.max_log_slowdown, min(cfg.max_log_slowdown, log_slowdown)
)
return math.exp(log_slowdown), contrib
# ============================================================================
# Memory (hv-forget)
# ============================================================================
def _memory_stability(
sentence: str, density: float, cfg: HVReaderConfig
) -> Tuple[float, float]:
words = _words(sentence)
n_content = max(1, len(_content_words(sentence)))
n_rare = sum(1 for w in words if _is_rare(w))
rare_ratio = n_rare / n_content
n_digits = len(_NUMBER_RE.findall(sentence))
n_proper = sum(
1 for i, w in enumerate(words)
if i > 0 and w[0].isupper()
and w.lower() not in COMMON_WORDS
and w.lower() not in DISCOURSE_MARKERS
and len(w) >= 3
)
salience = min(1.0, (n_digits + n_proper) / n_content)
stability = (
cfg.memory_stability_base
+ cfg.memory_stability_density * density
+ cfg.memory_stability_rare * rare_ratio
+ cfg.memory_stability_salience * salience
)
return stability, salience
def _memory_retention(stability: float, days: float) -> float:
if stability <= 0:
return 0.0
return math.exp(-days / stability)
# ============================================================================
# Density
# ============================================================================
def _density_per_sentence(sentences: List[str]) -> List[float]:
seen: Set[str] = set()
out: List[float] = []
for i, s in enumerate(sentences):
content = [_stem(w) for w in _content_words(s)]
if not content:
out.append(0.0)
continue
if i == 0:
out.append(1.0)
else:
novel = [w for w in content if w not in seen]
out.append(len(novel) / len(content))
seen.update(content)
return out
# ============================================================================
# Fold (hv-fold)
# ============================================================================
def _definite_nps(sentence: str) -> List[Tuple[str, str]]:
out: List[Tuple[str, str]] = []
words = list(_WORD_RE.finditer(sentence))
for idx, w in enumerate(words):
if w.group(0).lower() not in ARTICLES:
continue
tail = words[idx + 1: idx + 3]
content = [
tw.group(0).lower() for tw in tail
if tw.group(0).lower() not in COMMON_WORDS
and len(tw.group(0)) >= 3
]
if not content:
continue
if any(c in UNIQUE_REFERENTS for c in content):
continue
head = content[-1]
end = tail[-1].end() if tail else w.end()
out.append((sentence[w.start():end], head))
return out
def _fold_loads(sentences: List[str], cfg: HVReaderConfig) -> List[float]:
n = len(sentences)
word_sentences: Dict[str, Set[int]] = defaultdict(set)
for i, s in enumerate(sentences):
for w in _words(s):
lw = w.lower()
if lw not in COMMON_WORDS and len(lw) >= 3:
word_sentences[_stem(lw)].add(i)
loads: List[float] = []
for i, s in enumerate(sentences):
load = 0.0
if _STANDALONE_DEMON_RE.match(s.strip()):
load += cfg.fold_demon_weight
for _, head in _definite_nps(s):
stem = _stem(head)
occ = word_sentences.get(stem, set())
prior = [j for j in occ if j < i]
later = [j for j in occ if j > i]
if prior:
load += cfg.fold_resolved
elif later:
delay = min(later) - i
load += cfg.fold_forward * delay
else:
if i == 0 and cfg.exempt_first_sentence:
pass
else:
load += cfg.fold_definite_only
if any(w.lower() in REENTRY_MARKERS for w in _words(s)):
load += cfg.fold_reentry
loads.append(load)
return loads
# ============================================================================
# Slip (hv-slip)
# ============================================================================
def _slip_per_sentence(
sentences: List[str], cfg: HVReaderConfig
) -> List[float]:
n = len(sentences)
if n == 0:
return []
lengths = [len(_words(s)) for s in sentences]
slips: List[float] = []
prev_content: Set[str] = set()
for i, s in enumerate(sentences):
lo = max(0, i - cfg.monotone_window)
hi = min(n, i + cfg.monotone_window + 1)
window = lengths[lo:hi]
if len(window) > 1:
mean = sum(window) / len(window)
var = sum((x - mean) ** 2 for x in window) / len(window)
std = math.sqrt(var)
monotone = max(0.0, 1.0 - std / 8.0)
else:
monotone = 0.0
content = {_stem(w) for w in _content_words(s)}
if prev_content and content:
overlap = len(content & prev_content) / max(1, len(content))
else:
overlap = 0.0
prev_content = content
n_digits = len(_NUMBER_RE.findall(s))
words_s = _words(s)
n_proper = sum(
1 for j, w in enumerate(words_s)
if j > 0 and w[0].isupper()
and w.lower() not in COMMON_WORDS
and len(w) >= 3
)
absence = 1.0 if (n_digits + n_proper) == 0 else 0.0
slip = (
cfg.slip_monotone_weight * monotone
+ cfg.slip_repetition_weight * overlap
+ cfg.slip_absence_weight * absence
)
slips.append(max(0.0, min(1.0, slip)))
return slips
# ============================================================================
# Hunger (hv-hunger)
# ============================================================================
def _hunger_deltas(sentences: List[str], cfg: HVReaderConfig) -> List[float]:
deltas: List[float] = []
for s in sentences:
low = s.lower()
raised = sum(1 for m in QUESTION_RAISERS if m in low)
answered = sum(1 for m in ANSWER_MARKERS if m in low)
delta = (
cfg.hunger_question_weight * raised
- cfg.hunger_answer_weight * answered
)
deltas.append(delta)
return deltas
# ============================================================================
# Wall (hv-wall)
# ============================================================================
def _direction_of(word: str) -> List[str]:
w = word.lower()
sw = _stem(w)
out = []
for g, words in DIRECTION_GROUPS.items():
if w in words or sw in words:
out.append(g)
return out
def _closest_topic(words: List[str], idx: int) -> Optional[int]:
best = None
best_key: Tuple[int, int] = (10, 1)
for j in range(max(0, idx - 3), min(len(words), idx + 4)):
if j == idx:
continue
cand = words[j]
if cand in COMMON_WORDS or len(cand) < 4:
continue
if _direction_of(cand):
continue
if cand in DISCOURSE_MARKERS:
continue
after = 0 if j > idx else 1
key = (abs(j - idx), after)
if key < best_key:
best_key = key
best = j
return best
def _direction_pairs(sentence: str) -> List[Tuple[str, str, str]]:
words = [w.lower() for w in _words(sentence)]
out: List[Tuple[str, str, str]] = []
for i, w in enumerate(words):
groups = _direction_of(w)
if not groups:
continue
j = _closest_topic(words, i)
if j is None:
continue
topic = _stem(words[j])
for g in groups:
out.append((topic, g, w))
return out
def _wall_confidence(
sentence: str, cfg: HVReaderConfig
) -> Tuple[float, int, int]:
words = [w.lower() for w in _words(sentence)]
n = max(1, len(words))
cert = sum(1 for w in words if w in CERTAINTY_MARKERS)
hedg = sum(1 for w in words if w in HEDGES)
conf = min(1.0, cfg.certainty_scale * cert / n)
hedge = min(1.0, cfg.hedge_scale * hedg / n)
return max(0.0, conf - 0.5 * hedge), cert, hedg
def _wall_evidence(
sentence: str, cfg: HVReaderConfig
) -> Tuple[float, float, bool]:
words = _words(sentence)
content = _content_words(sentence)
n_content = max(1, len(content))
n_digits = len(_NUMBER_RE.findall(sentence))
n_proper = sum(
1 for i, w in enumerate(words)
if i > 0 and w[0].isupper()
and w.lower() not in COMMON_WORDS
and w.lower() not in DISCOURSE_MARKERS
and len(w) >= 3
)
n_rare = sum(1 for w in content if len(w) >= 8)
spec = (
0.5 * (n_digits / n_content)
+ 0.3 * (n_proper / n_content)
+ 0.2 * (n_rare / n_content)
)
spec = min(1.0, cfg.specificity_scale * spec)
attribution = _has_evidence_marker(sentence)
evidence = 0.6 * spec + 0.4 * (1.0 if attribution else 0.0)
return evidence, spec, attribution
def _wall_per_sentence(
sentences: List[str], cfg: HVReaderConfig
) -> List[Tuple[float, str, List[str]]]:
n = len(sentences)
out: List[Tuple[float, str, List[str]]] = []
prior_sentences: List[str] = []
prior_conf: List[float] = []
prior_evid: List[float] = []
for i, s in enumerate(sentences):
confidence, cert_count, _ = _wall_confidence(s, cfg)
evidence, _spec, attribution = _wall_evidence(s, cfg)
n_words = len(_words(s))
overclaim = (
max(0.0, confidence - evidence)
if n_words >= cfg.min_sentence_words
else 0.0
)
# Contradiction via direction-group conflicts.
contradiction = 0.0
contra_idx: Optional[int] = None
pairs: List[Tuple[str, str]] = []
if prior_sentences:
curr_pairs = _direction_pairs(s)
prior_dirs: Dict[str, List[Tuple[int, str, str]]] = {}
for j, p in enumerate(prior_sentences):
for topic, g, src in _direction_pairs(p):
prior_dirs.setdefault(topic, []).append((j, g, src))
for topic, curr_g, curr_src in curr_pairs:
for prior_idx, prior_g, prior_src in prior_dirs.get(topic, []):
if (prior_g, curr_g) in GROUP_OPPOSITES:
pairs.append((
f"{topic}:{prior_src}↔{curr_src}",
f"({prior_g} vs {curr_g})",
))
contra_idx = prior_idx
if pairs:
contradiction = min(
1.0, cfg.contradiction_scale * math.sqrt(len(pairs))
)
# Gap.
gap = 0.0
has_conclusion = _has_conclusion_marker(s)
if has_conclusion and prior_evid:
prior_ev_mean = sum(prior_evid) / len(prior_evid)
ev_deficit = max(0.0, confidence - prior_ev_mean)
prior_conf_mean = (
sum(prior_conf) / len(prior_conf) if prior_conf else 0.0
)
conf_jump = max(0.0, confidence - prior_conf_mean)
gap = ev_deficit + cfg.gap_confidence_jump_weight * conf_jump
# Priority ordering.
if contradiction >= cfg.contradiction_min:
wall, wtype = contradiction, "contradiction"
elif overclaim >= cfg.overclaim_min:
wall, wtype = overclaim, "overclaim"
elif gap >= cfg.gap_min:
wall, wtype = gap, "gap"
else:
best = max(overclaim, contradiction, gap)
if best <= 0.0:
wall, wtype = 0.0, "neutral"
else:
wall = best
if best == contradiction:
wtype = "contradiction"
elif best == overclaim:
wtype = "overclaim"
else:
wtype = "gap"
wall = min(1.0, wall)
has_contrast, _contrast_word = _has_contrast_marker(s)
if has_contrast and wall > 0.05:
wall = min(1.0, wall + cfg.contrast_bonus)
reasons: List[str] = []
if wtype == "contradiction" and pairs:
reasons.append(f"conflict with sentence {contra_idx}: {pairs[0][0]}")
elif wtype == "overclaim":
if cert_count:
matched = [
w.lower() for w in _words(s)
if w.lower() in CERTAINTY_MARKERS
]
reasons.append(
f"certainty markers: "
f"{', '.join(repr(m) for m in matched[:3])}"
)
if not attribution:
reasons.append("no attribution marker")
elif wtype == "gap":
if has_conclusion:
reasons.append("conclusion marker with weak prior evidence")
out.append((wall, wtype if wall > 0.0 else "neutral", reasons))
prior_sentences.append(s)
prior_conf.append(confidence)
prior_evid.append(evidence)
return out
# ============================================================================
# The model
# ============================================================================
class HVReader:
"""Unified reading-experience model."""
def __init__(self, config: Optional[HVReaderConfig] = None):
self.config = config or HVReaderConfig()
self._obs = 0
def __repr__(self) -> str:
return (
f"HVReader(baseline_wpm={self.config.baseline_wpm}, "
f"wall_threshold={self.config.wall_threshold}, "
f"version={self.config.version})"
)
def analyze(self, text: str) -> ReadingProfile:
if not text or not text.strip():
return self._empty(text)
sentences = _sentences(text)
n = len(sentences)
if n == 0:
return self._empty(text)
spans: List[Tuple[int, int]] = []
cursor = 0
for s in sentences:
i = text.find(s, cursor)
if i < 0:
i = cursor
spans.append((i, i + len(s)))
cursor = i + len(s)
pace_feats = [_pace_features(s) for s in sentences]
slowdowns: List[float] = []
wpms: List[float] = []
for f in pace_feats:
sd, _ = _pace_slowdown(f, self.config)
slowdowns.append(sd)
wpms.append(
self.config.baseline_wpm / sd
if sd > 0 else self.config.baseline_wpm
)
densities = _density_per_sentence(sentences)
fold_loads = _fold_loads(sentences, self.config)
slips = _slip_per_sentence(sentences, self.config)
hunger_deltas = _hunger_deltas(sentences, self.config)
hunger_cumulative: List[float] = []
h = 0.0
for d in hunger_deltas:
h += d
hunger_cumulative.append(h)
stabilities: List[float] = []
retentions: List[float] = []
for s, dens in zip(sentences, densities):
stab, _ = _memory_stability(s, dens, self.config)
stabilities.append(stab)
retentions.append(
_memory_retention(stab, self.config.memory_target_days)
)
wall_per = _wall_per_sentence(sentences, self.config)
span_profiles: List[SpanProfile] = []
for i in range(n):
wall_score, wall_type, reasons = wall_per[i]
n_words_span = int(pace_feats[i]["n_words"])
span_profiles.append(SpanProfile(
id=i,
text=sentences[i],
span=spans[i],
n_words=n_words_span,
wpm=wpms[i],
slowdown=slowdowns[i],
density=densities[i],
stability_days=stabilities[i],
retention_week=retentions[i],
fold_load=fold_loads[i],
slip=slips[i],
hunger_delta=hunger_deltas[i],
hunger=hunger_cumulative[i],
wall=wall_score,
wall_type=wall_type,
reasons=reasons[: self.config.top_reasons],
))
n_words_total = sum(s.n_words for s in span_profiles)
mean_wpm = (
sum(s.wpm * s.n_words for s in span_profiles)
/ max(1, n_words_total)
)
pace = max(0.0, min(1.0, mean_wpm / 400.0))
memory = sum(retentions) / n if n else 0.0
total_load = sum(fold_loads)
passes = 1.0 + total_load / max(1, n)
slip = sum(slips) / n if n else 0.0
wall_scores = [s.wall for s in span_profiles]
wall_max = max(wall_scores) if wall_scores else 0.0
wall = wall_max
wps = mean_wpm / 60.0
ttu_s = n_words_total / wps if wps > 0 else 0.0
hunger_final = hunger_cumulative[-1] if hunger_cumulative else 0.0
slowest = (
min(span_profiles, key=lambda s: s.wpm)
if span_profiles else None
)
wall_span: Optional[SpanProfile] = None
if wall > 0.15:
wall_span = max(span_profiles, key=lambda s: s.wall)
summary = self._summary(
span_profiles, pace, memory, passes, slip, wall
)
self._obs += 1
return ReadingProfile(
text=text,
n_sentences=n,
n_words=n_words_total,
pace=pace,
memory=memory,
passes=passes,
slip=slip,
wall=wall,
ttu_s=ttu_s,
mean_wpm=mean_wpm,
hunger_final=hunger_final,
spans=span_profiles,
slowest_span=slowest,
wall_span=wall_span,
summary=summary,
)
def _empty(self, text: str) -> ReadingProfile:
return ReadingProfile(
text=text,
n_sentences=0,
n_words=0,
pace=0.0,
memory=0.0,
passes=1.0,
slip=0.0,
wall=0.0,
ttu_s=0.0,
mean_wpm=0.0,
hunger_final=0.0,
spans=[],
slowest_span=None,
wall_span=None,
summary="Empty text.",
)
@staticmethod
def _summary(
spans: List[SpanProfile],
pace: float,
memory: float,
passes: float,
slip: float,
wall: float,
) -> str:
if not spans:
return "Empty text."
parts = []
parts.append(f"reads at {pace * 400:.0f} WPM (pace {pace:.2f})")
parts.append(f"memory after a week: {memory * 100:.0f}%")
parts.append(f"requires {passes:.2f} passes")
parts.append(f"slip probability: {slip:.2f}")
if wall > 0.15:
wall_span = max(spans, key=lambda s: s.wall)
parts.append(
f"wall at sentence {wall_span.id} "
f"({wall_span.wall_type}, {wall:.2f})"
)
else:
parts.append("no wall")
return ". ".join(parts).capitalize() + "."
def render(self, profile: ReadingProfile) -> str:
lines: List[str] = []
bar = "=" * 72
lines.append(bar)
lines.append("hv-reader — the reading experience")
lines.append(bar)
lines.append("")
lines.append(f" text : {profile.n_sentences} sentences, "
f"{profile.n_words} words")
lines.append("")
lines.append(" READING PROFILE")
lines.append(" " + "-" * 68)
lines.append(f" pace {profile.pace:>6.3f} "
f"({profile.mean_wpm:.0f} WPM)")
lines.append(f" memory {profile.memory:>6.3f} "
f"(fraction surviving 1 week)")
lines.append(f" passes {profile.passes:>6.3f} "
f"(reads needed)")
lines.append(f" slip {profile.slip:>6.3f} "
f"(attention-lapse probability)")
lines.append(f" wall {profile.wall:>6.3f} "
f"(reader-refusal probability)")
lines.append("")
lines.append(f" ttu : {profile.ttu_s:.1f} s "
f"(total reading time)")
lines.append(f" hunger : {profile.hunger_final:+.2f} "
f"(unresolved questions)")
lines.append("")
if not profile.spans:
lines.append(" (no content)")
return "\n".join(lines)
lines.append(" PER-SENTENCE")
lines.append(" " + "-" * 68)
lines.append(
f" {'id':>3} {'wpm':>5} {'dens':>5} {'ret':>5} "
f"{'fold':>5} {'slip':>5} {'wall':>5} type"
)
for s in profile.spans:
marker = (
"*" if (profile.wall_span and s.id == profile.wall_span.id)
else " "
)
lines.append(
f" {marker}{s.id:>2} {s.wpm:>5.0f} {s.density:>5.2f} "
f"{s.retention_week:>5.2f} {s.fold_load:>5.2f} "
f"{s.slip:>5.2f} {s.wall:>5.2f} {s.wall_type}"
)
lines.append("")
if profile.slowest_span:
s = profile.slowest_span
lines.append(" SLOWEST SPAN")
lines.append(" " + "-" * 68)
lines.append(f" [{s.id}] {s.wpm:.0f} WPM "
f"(slowdown {s.slowdown:.2f}x)")
lines.append(f" \"{self._shorten(s.text, 60)}\"")
lines.append("")
if profile.wall_span:
s = profile.wall_span
lines.append(" WALL SPAN")
lines.append(" " + "-" * 68)
lines.append(f" [{s.id}] {s.wall_type} "
f"(score {s.wall:.3f})")
lines.append(f" \"{self._shorten(s.text, 60)}\"")
for r in s.reasons:
lines.append(f" - {r}")
lines.append("")
lines.append(" SUMMARY")
lines.append(" " + "-" * 68)
lines.append(f" {profile.summary}")
lines.append("")
return "\n".join(lines)
@staticmethod
def _shorten(s: str, n: int) -> str:
s = s.strip().replace("\n", " ")
if len(s) <= n:
return s
return s[: n - 1].rsplit(" ", 1)[0] + "…"
def save_pretrained(self, save_dir: str) -> None:
os.makedirs(save_dir, exist_ok=True)
payload = {
"config": asdict(self.config),
"observations": self._obs,
}
with open(os.path.join(save_dir, "config.json"), "w") as f:
json.dump(payload, f, indent=2)
@classmethod
def from_pretrained(cls, save_dir: str) -> "HVReader":
with open(os.path.join(save_dir, "config.json"), "r") as f:
payload = json.load(f)
cfg_dict = payload.get("config", {})
known = {f.name for f in HVReaderConfig.__dataclass_fields__.values()}
cfg_dict = {k: v for k, v in cfg_dict.items() if k in known}
cfg = HVReaderConfig(**cfg_dict)
obj = cls(config=cfg)
obj._obs = int(payload.get("observations", 0))
return obj
# ============================================================================
# Demo
# ============================================================================
SAMPLE_FICTION = (
"The old man walked slowly to the boat. He stopped, looked at the "
"water, and then continued. The sea was quiet that morning. He "
"pushed the boat into the water and climbed in. The oars were cold "
"in his hands. He rowed out past the harbor and into the open sea."
)
SAMPLE_ACADEMIC = (
"A black hole is a region of spacetime where gravity is so strong "
"that nothing — no particles or even electromagnetic radiation such "
"as light — can escape from it. The theory of general relativity "
"predicts that a sufficiently compact mass can deform spacetime to "
"form a black hole. The boundary of the region from which no escape "
"is possible is called the event horizon. Although the event horizon "
"has profound effects on the fate of an object that crosses it, it "
"has no locally detectable features. A black hole acts as a perfect "
"black body, and moreover, it emits Hawking radiation."
)
SAMPLE_OVERCLAIM = (
"The data suggests some correlation between the policy and the outcome. "
"Results appear to indicate a modest effect in some subpopulations. "
"The mechanism remains unclear, and further work is needed to establish "
"causality. "
"Therefore, the policy definitively causes the outcome in all cases, "
"and this is unquestionably proven by the evidence."
)
SAMPLE_MONOTONE = (
"The system processes the input. The system processes the data. "
"The system processes the output. The system processes the result. "
"The system processes the value. The system processes the record."
)
def _demo(output_dir: str = "./hv_reader_output") -> None:
os.makedirs(output_dir, exist_ok=True)
m = HVReader()
samples = [
("Fiction", SAMPLE_FICTION),
("Academic", SAMPLE_ACADEMIC),
("Overclaim", SAMPLE_OVERCLAIM),
("Monotone", SAMPLE_MONOTONE),
]
for name, text in samples:
print()
print("#" * 72)
print(f"# {name}")
print("#" * 72)
print(m.render(m.analyze(text)))
print()
print("=" * 72)
print("Summary across samples")
print("=" * 72)
print(
f" {'sample':<12} {'words':>6} {'pace':>6} {'mem':>6} "
f"{'pass':>5} {'slip':>6} {'wall':>5} {'ttu':>6}"
)
print(" " + "-" * 68)
for name, text in samples:
r = m.analyze(text)
print(
f" {name:<12} {r.n_words:>6} {r.pace:>6.3f} "
f"{r.memory:>6.3f} {r.passes:>5.2f} "
f"{r.slip:>6.3f} {r.wall:>5.2f} {r.ttu_s:>5.1f}s"
)
print()
print("=" * 72)
print("Save / load round trip")
print("=" * 72)
save_path = os.path.join(output_dir, "hv_reader_model")
m.save_pretrained(save_path)
m2 = HVReader.from_pretrained(save_path)
print(f" saved to : {save_path}")
print(f" reloaded : {m2!r}")
a = m.analyze(SAMPLE_OVERCLAIM)
b = m2.analyze(SAMPLE_OVERCLAIM)
print(f" pace : {a.pace:.4f}")
print(f" wall : {a.wall:.4f}")
print(f" identical : "
f"{abs(a.pace - b.pace) < 1e-9 and abs(a.wall - b.wall) < 1e-9}")
print()
# ============================================================================
# CLI
# ============================================================================
def _cli() -> None:
p = argparse.ArgumentParser(
description="hv-reader: the reading experience in one call."
)
p.add_argument("--text", type=str, default="",
help="text to analyze (or '-' to read stdin)")
p.add_argument("--json", action="store_true",
help="output JSON instead of a rendered report")
p.add_argument("--profile", action="store_true",
help="print only the five-axis profile")
p.add_argument("--save-to", type=str, default="",
help="save the model to this directory")
p.add_argument("--outdir", type=str, default="./hv_reader_output")
args = p.parse_args()
text = sys.stdin.read() if args.text == "-" else args.text
m = HVReader()
if args.save_to:
m.save_pretrained(args.save_to)
print(f"saved to {args.save_to}", file=sys.stderr)
if not text:
_demo(args.outdir)
return
profile = m.analyze(text)
if args.profile:
print(f"pace {profile.pace:.3f}")
print(f"memory {profile.memory:.3f}")
print(f"passes {profile.passes:.3f}")
print(f"slip {profile.slip:.3f}")
print(f"wall {profile.wall:.3f}")
return
if args.json:
print(json.dumps(profile.to_dict(), indent=2, ensure_ascii=False))
else:
print(m.render(profile))
if __name__ == "__main__":
_cli() |