nlp - 使用 NLTK 访问 Switchboard 语料库中的对话行为注释

Question

有什么方法可以使用 NLTK 访问 Switchboard 语料库中的对话行为注释吗？

我试过使用 nltk.corpus.reader.switchboard :

import nltk
print('nltk.corpus.switchboard.words(): {0}'.format(nltk.corpus.switchboard.words()))
print('nltk.corpus.switchboard.tagged_words(): {0}'.format(nltk.corpus.switchboard.tagged_words()))
print('nltk.corpus.switchboard.turns(): {0}'.format(nltk.corpus.switchboard.turns()))
print('nltk.corpus.switchboard.tagged_turns(): {0}'.format(nltk.corpus.switchboard.tagged_turns()))
print('nltk.corpus.switchboard.discourses(): {0}'.format(nltk.corpus.switchboard.discourses()))
print('nltk.corpus.switchboard.tagged_discourses(): {0}'.format(nltk.corpus.switchboard.tagged_discourses()))

但这些似乎都不包含对话行为注释。

Answer 1

您可能错过了语料库的实际加载。尝试这样做:

from nltk.corpus import switchboard
from nltk import download

download('switchboard')

switchboard.ensure_loaded()
switchboard.tagged_discourses()

你应该得到这样的输出:

from nltk.corpus import switchboard
from nltk import download

download('switchboard') [nltk_data] 
Downloading package switchboard to [nltk_data]     /usr/local/lib/nltk_data... [nltk_data] Unzipping corpora/switchboard.zip. Out[22]: True

switchboard.ensure_loaded()

switchboard.tagged_discourses()
[[<A.1: u'Uh/UH ,/, do/VBP you/PRP have/VB a/DT pet/NN Randy/NNP ?/.'>, <B.2: u'Uh/UH ,/, yeah/UH ,/, currently/RB we/PRP have/VBP a/DT poodle/NN ./.'>, <A.3: u'A/DT poodle/NN ,/, miniature/JJ or/CC ,/, uh/UH ,/, full/JJ size/NN ?/.'>, <B.4: u"Yeah/UH ,/, uh/UH ,/, it/PRP 's/BES ,/, uh/UH miniature/JJ ./.">, <A.5: u'Uh-huh/UH ./.'>, <B.6: u'Yeah/UH ./.'>, <A.7: u'I/PRP read/VBD somewhere/RB that/IN ,/, the/DT poodles/NNS is/VBZ one/CD of/IN the/DT ,/, the/DT most/RBS intelligent/JJ dogs/NNS ,/, uh/UH ,/, around/RB ./.'>, <B.8: u"Well/UH ,/, um/UH ,/, I/PRP would/MD n't/RB ,/, uh/UH ,/, I/PRP definitely/RB would/MD n't/RB dispute/VB that/IN ,/, it/PRP ,/, it/PRP 's/BES actually/RB my/PRP$ wife/NN 's/POS dog/NN ,/, uh/UH ,/, I/PRP ,/, I/PRP became/VBD part/NN owner/NN six/CD months/NNS ago/RB when/WRB we/PRP got/VBD married/VBN ,/, but/CC ,/, uh/UH ,/, it/PRP ,/, uh/UH ,/, definitely/RB responds/VBZ to/IN ,/, uh/UH ,/, to/IN authority/NN and/CC ,/, I/PRP 've/VBP had/VBN dogs/NNS in/IN the/DT past/JJ and/CC ,/, uh/UH ,/, it/PRP seems/VBZ ,/, it/PRP seems/VBZ to/TO ,/, uh/UH ,/, respond/VB real/RB well/RB ,/, it/PRP ,/, it/PRP she/PRP 's/HVS ,/, she/PRP 's/HVS picked/VBN up/RP a/DT lot/NN of/IN things/NNS ,/, uh/UH ,/, just/RB ,/, just/RB by/IN ,/, uh/UH ,/, teaching/VBG by/IN force/NN ,/, I/PRP guess/VBP is/VBZ what/WP I/PRP 'd/MD like/VB to/TO say/VB ./.">, <A.9: u"Oh/UH ,/, uh-huh/UH ./. So/RB ,/, you/PRP ,/, you/PRP 've/VBP only/RB known/VBN the/DT dog/NN ,/, wh-/XX ,/, how/WRB long/JJ did/VBD you/PRP say/VB ./.">, <B.10: u'Well/UH ,/, about/RB a/DT year/NN I/PRP guess/VBP ./.'>, <A.11: u'Oh/UH ,/, well/UH ,/, uh/UH ,/, is/VBZ it/PRP ,/, uh/UH ,/, uh/UH ,/, how/WRB old/JJ is/VBZ the/DT dog/NN ?/.'>, <B.12: u'It/PRP just/RB turned/VBD two/CD ,/, I/PRP believe/VBP ./.'>, <A.13: u"Oh/UH ,/, it/PRP 's/BES still/RB just/RB a/DT pup/NN ./.">, <B.14: u'Pretty/RB much/JJ ,/, yeah/UH ,/, yeah/UH ./.'>, <A.15: u'Yeah/UH ,/, I/PRP have/VBP a/DT ,/, uh/UH ,/, well/UH a/DT mutt/NN ,/, myself/PRP ./. I/PRP call/VBP it/PRP a/DT ,/, uh/UH ,/, uh/UH ,/, Chowperd/NNP ./.'>, <B.16: u'Okay/UH ./.'>, <A.17: u"It/PRP 's/BES ,/, uh/UH ,/, part/NN Chow/NNP and/CC part/NN Shepherd/NNP and/CC it/PRP ,/, as/IN I/PRP understand/VBP it/PRP ,/, uh/UH ,/, both/DT sides/NNS of/IN the/DT ,/, were/VBD thoroughbreds/NNS ./. So/RB ,/, she/PRP 's/BES a/DT genuine/JJ Chowperd/NNP ./.">, <B.18: u'Oh/UH ,/, that/DT sounds/VBZ interesting/JJ ./.'>, <A.19: u'She/PRP has/VBZ the/DT ,/, the/DT color/NN and/CC the/DT black/JJ to-/NN ,/, tongue/NN of/IN a/DT Chow/NNP ,/, but/CC ,/, uh/UH ,/, she/PRP has/VBZ the/DT shap-/NN ,/, the/DT shape/NN of/IN the/DT ,/, uh/UH ,/, uh/UH ,/, Shepherd/NNP ./.'>, <B.20: u"Oh/UH ,/, that/DT 's/BES ,/, that/DT 's/BES neat/JJ ./. How/WRB ,/, about/RB how/WRB big/JJ then/RB ?/.">, <A.21: u"Oh/UH ,/, she/PRP weighs/VBZ in/RP at/IN about/RB fifty/CD pounds/NNS ,/, so/RB she/PRP 's/BES a/DT medium/JJ size/NN ./.">, <B.22: u'Yeah/UH ,/, yeah/UH ./.'>, <A.23: u"But/CC she/PRP 's/BES big/JJ enough/RB to/TO be/VB intimidating/JJ ,/,">, <B.24: u'Most/JJS definitely/RB ./.'>, <A.25: u'it/PRP is/VBZ a/DT fi/VBN ,/, fixed/VBN female/NN ,/, by/IN the/DT way/NN ,/,'>, <B.26: u'Yeah/UH ./.'>, <A.27: u'and/CC right/RB from/IN day/NN one/CD ,/, she/PRP was/VBD teaching/VBG me/PRP ./.'>, <B.28: u"Oh/UH ,/, I/PRP would/MD n't/RB doubt/VB it/PRP ,/, yeah/UH ./.">, <A.29: u"She/PRP 's/BES the/DT most/RBS intelligent/JJ dog/NN I/PRP 've/VBP ever/RB seen/VBN ./. Course/NN ,/, I/PRP 'm/VBP a/DT little/RB prejudiced/JJ ,/, of/IN course/NN ./.">, <B.30: u"Well/UH that/DT 's/BES understandable/JJ ,/, yeah/UH ,/, it/PRP 's/BES ,/, uh/UH ,/,">, <A.31: u'You/PRP know/VBP ,/, the/DT first/JJ time/NN I/PRP brought/VBD her/PRP$ home/NN ,/, she/PRP was/VBD only/RB ,/, uh/UH ,/, was/VBD it/PRP six/CD weeks/NNS old/JJ ./. And/CC I/PRP spread/VBD the/DT newspapers/NNS out/RP in/IN the/DT kitchen/NN area/NN ./.'>, <B.32: u'Uh-huh/UH ./.'>, <A.33: u'But/CC ,/, uh/UH ,/, next/JJ morning/NN ,/, she/PRP let/VBD me/PRP know/VB in/IN no/DT uncertain/JJ terms/NNS that/IN she/PRP wanted/VBD to/TO use/VB the/DT bathroom/NN ./.'>, <B.34: u'Okay/UH ./.'>, <A.35: u'So/RB ,/, on/IN next/JJ night/NN ,/, I/PRP spread/VBD the/DT newspaper/NN in/IN the/DT bathroom/NN and/CC she/PRP used/VBD them/PRP there/RB ./.'>, <B.36: u'Oh/UH ./.'>, <A.37: u"But/CC it/PRP was/VBD n't/RB too/RB long/JJ until/IN she/PRP ,/, uh/UH ,/, found/VBD out/RP she/PRP could/MD wait/VB until/IN I/PRP let/VBP her/PRP out/RB in/IN the/DT morning/NN ./.">, <B.38: u'Yeah/UH ./.'>, <A.39: u'And/CC since/IN then/RB ,/, I/PRP ,/, I/PRP live/VBP alone/RB ,/,'>, <B.40: u'Okay/UH ./.'>, <A.41: u"and/CC ,/, uh/UH ,/, I/PRP live/VBP in/IN motor/NN home/NN ,/, by/IN the/DT way/NN ,/, I/PRP 'm/VBP ,/, uh/UH ,/, an/DT R/NN V/NN ,/, full/GW time/^RB R/NN V/NN -er/NN ,/, and/CC it/PRP 's/BES ,/, it/PRP 's/BES such/PDT a/DT pleasure/NN to/TO come/VB home/NN at/IN night/NN and/CC you/PRP can/MD see/VB her/PRP$ smiling/VBG from/IN ear/NN to/IN ear/NN ,/, she/PRP 's/BES so/RB happy/JJ to/TO see/VB me/PRP ./.">, <B.42: u'Yeah/UH ,/, definitely/UH ./.'>, <A.43: u"And/CC ,/, uh/UH ,/, I/PRP do/VBP n't/RB know/VB if/IN you/PRP get/VBP that/DT kind/NN of/IN greeting/NN or/CC not/RB ./.">, <B.44: u"Yeah/UH ,/, I/PRP can/MD honestly/RB say/VB we/PRP do/VBP ,/, uh/UH ,/, we/PRP ,/, uh/UH ,/, just/RB recently/RB put/VBD a/DT security/NN system/NN in/IN our/PRP$ house/NN and/CC so/RB now/RB ,/, uh/UH ,/, in/IN order/NN to/TO ,/, uh/UH ,/, to/TO accommodate/VB the/DT motion/NN detectors/NNS we/PRP have/VBP to/TO keep/VB her/PRP$ ,/, uh/UH ,/, uh/UH ,/, locked/VBN up/RP in/IN the/DT ,/, the/DT master/NN bedroom/NN during/IN the/DT day/NN and/CC then/RB she/PRP 's/HVS got/VBN the/DT ,/, the/DT bedroom/NN and/CC the/DT bathroom/NN to/IN ,/, for/IN free/JJ run/NN during/IN the/DT day/NN but/CC ,/,">, <A.45: u'Uh-huh/UH ./.'>, <B.46: u"we/PRP 've/VBP always/RB got/VBN ,/, uh/UH ,/, got/VBN a/DT nose/NN and/CC tongue/NN pressed/VBN up/IN against/IN the/DT window/NN when/WRB we/PRP come/VBP walking/VBG up/IN to/IN the/DT front/JJ door/NN ./.">, <A.47: u'./.'>, <B.48: u"She/PRP 's/BES definitely/RB ready/JJ to/TO get/VB out/RB and/CC run/VB around/RB ./.">, <A.49: u"Well/UH my/PRP$ dog/NN 's/BES an/DT outdoor/JJ type/NN ,/, she/PRP does/VBZ not/RB like/VB to/TO be/VB indoors/RB ./.">, <B.50: u'Really/UH ./.'>, <A.51: u"Uh/UH ,/, she/PRP 'd/MD rather/RB sle/VB ,/, sleep/VB outside/RB on/IN the/DT ,/, the/DT cold/JJ ground/NN at/IN night/NN ./.">, <B.52: u'Oh/UH wow/UH ./.'>, <A.53: u'But/CC ,/, uh/UH ,/, I/PRP do/VBP make/VB her/PRP$ come/VB in/RB ./.'>, <B.54: u'Yeah/UH ./.'>, <A.55: u"And/CC I/PRP feed/VBP her/PRP$ indoors/RB ,/, that/DT 's/BES to/TO lure/VB her/PRP in/RB ,/, but/CC during/IN the/DT day/NN I/PRP have/VBP her/PRP on/IN a/DT ,/, uh/UH ,/, on/IN a/DT leash/NN ,/,">, <B.56: u'Okay/UH ./.'>, <A.57: u'which/WDT is/VBZ ,/, uh/UH ,/, on/IN sort/RB of/RB a/DT run/NN ./. I/PRP have/VBP a/DT ,/, a/DT thirty/CD foot/NN cable/NN ,/,'>, <B.58: u'Okay/UH ./.'>, <A.59: u'running/VBG from/IN one/CD stake/NN to/IN another/DT ,/, and/CC then/RB attached/VBN to/IN that/DT is/VBZ a/DT ,/, uh/UH ,/, twelve/CD foot/NN leash/NN ,/,'>, <B.60: u'Okay/UH ./.'>, <A.61: u'so/RB she/PRP can/MD cover/VB quite/PDT an/DT area/NN ./.'>, <B.62: u'Most/JJS definitely/RB ./.'>, <A.63: u"And/CC ,/, uh/UH ,/, she/PRP 's/BES the/DT best/JJS ,/, uh/UH ,/, burglar/NN alarm/NN going/VBG ./.">, <B.64: u"Yeah/UH ,/, yeah/UH ,/, yeah/UH that/DT 's/BES ,/, uh/UH ,/, definite/JJ security/NN involved/VBN in/IN ,/, uh/UH ,/, in/IN a/DT dog/NN like/IN that/DT ./.">, <A.65: u"Oh/UH ,/, yeah/UH ,/, she/PRP ,/, uh/UH ,/, it/PRP 's/BES the/DT strangest/JJS thing/NN ,/, though/RB ,/, uh/UH ,/, children/NNS ,/, no/DT matter/NN how/WRB strange/JJ they/PRP are/VBP ,/, or/CC how/WRB new/JJ they/PRP might/MD be/VB can/MD walk/VB ,/, uh/UH ,/, right/RB up/IN to/IN her/PRP$ ,/,">, <B.66: u'Uh-huh/UH ./.'>, <A.67: u"but/CC adults/NNS ,/, if/IN they/PRP 're/VBP strange/JJ to/IN her/PRP$ ,/, or/CC ,/, or/CC they/PRP look/VBP suspicious/JJ or/CC something/NN ,/, boy/UH she/PRP acts/VBZ like/IN she/PRP wants/VBZ to/TO chew/VB their/PRP$ leg/NN off/RP ./.">, <B.68: u'Wow/UH ./.'>, <A.69: u'And/CC I/PRP have/VBP not/RB discovered/VBN yet/RB where/WRB the/DT ,/, the/DT line/NN is/VBZ between/IN children/NNS and/CC adults/NNS ./.'>, <B.70: u"Yeah/UH ,/, that/DT 's/BES interesting/JJ ./.">, <A.71: u'But/CC ,/, uh/UH ,/, she/PRP is/VBZ a/DT great/JJ comfort/NN to/IN me/PRP ./.'>, <B.72: u"Yeah/UH ,/, I/PRP know/VBP our/PRP$ dog/NN has/VBZ had/VBN ,/, uh/UH ,/, some/DT different/JJ reactions/NNS ,/, she/PRP 's/HVS never/RB really/RB been/VBN around/IN children/NNS and/CC ,/, uh/UH ,/, if/IN ,/, if/IN the/DT child/NN is/VBZ ,/, is/VBZ pretty/RB straight/GW forward/^JJ ,/, um/UH ,/, she/PRP 's/BES fine/JJ ./. If/IN ,/, if/IN a/DT child/NN is/VBZ a/DT little/RB intimidated/JJ ,/, she/PRP 'll/MD jump/VB around/RB and/CC ,/, and/CC yip/VB and/CC bark/VB quite/PDT a/DT bit/NN ,/, and/CC if/IN the/DT child/NN gets/VBZ scared/JJ ,/, uh/UH ,/, she/PRP 's/BES still/RB trying/VBG to/TO play/VB ,/, but/CC she/PRP does/VBZ n't/RB completely/RB understand/VB what/WP 's/BES going/VBG on/RP and/CC we/PRP 've/VBP had/VBN a/DT little/JJ confusion/NN with/IN ,/, with/IN ,/, uh/UH ,/, with/IN younger/JJR kids/NNS ./.">, <A.73: u'Uh-huh/UH ./.'>, <B.74: u"But/CC ,/, uh/UH ,/, you/PRP know/VBP ,/, that/DT 's/BES ,/, it/PRP 's/BES a/DT matter/NN of/IN exposure/NN really/RB ./. Um/UH ,/, we/PRP ,/, uh/UH ,/, took/VBD her/PRP$ home/NN to/IN ,/, uh/UH ,/, my/PRP$ family/NN 's/POS place/NN in/IN South/NNP Dakota/NNP ,/, and/CC she/PRP was/VBD the/DT one/NN that/WDT was/VBD intimidated/JJ then/RB ./. There/RB was/VBD about/RB seven/CD kids/NNS ranging/VBG from/IN about/RB ,/, uh/UH ,/, three/CD years/NNS to/IN ten/CD years/NNS running/VBG around/IN the/DT house/NN all/RB at/IN one/CD time/NN ,/,">, <A.75: u'Oh/UH ,/, uh-huh/UH ./.'>, <B.76: u"you/PRP know/VBP come/VB to/TO visit/VB Grandpa/NNP and/CC Grandma/NNP and/CC ,/, the/DT dog/NN kind/RB of/RB ,/, kind/RB of/RB felt/VBD out/IN of/IN place/NN then/RB because/IN she/PRP was/VBD ,/, she/PRP was/VBD being/VBG fed/VBN ,/, and/CC everything/NN else/RB from/IN all/DT directions/NNS ./. She/PRP really/RB did/VBD n't/RB know/VB how/WRB to/TO handle/VB herself/PRP ./.">, <A.77: u"You/PRP mean/VBP she/PRP did/VBD n't/RB appreciate/VB all/PDT that/DT attention/NN ./.">, <B.78: u"She/PRP really/RB did/VBD ,/, she/PRP just/RB ,/, uh/UH ,/, she/PRP ,/, she/PRP was/VBD alm-/XX ,/, she/PRP was/VBD just/RB inundated/VBN with/IN ,/, with/IN all/PDT the/DT attention/NN ./. Uh/UH ,/, she/PRP ,/, she/PRP kind/RB of/RB ,/, she/PRP kind/RB of/RB sat/VBD and/CC MUMBLEx/XX it/PRP all/DT in/RB for/IN a/DT little/JJ while/IN and/CC then/RB she/PRP 'd/MD go/VB get/VB back/RB in/RB and/CC try/VB to/TO play/VB and/CC ,/, and/CC what/WP not/RB ,/, but/CC ,/, uh/UH ,/, it/PRP was/VBD ,/, it/PRP was/VBD just/RB such/PDT a/DT ,/, such/JJ a/DT new/JJ experience/NN for/IN her/PRP$ ./. She/PRP 's/HVS only/RB been/VBN around/RB one/CD and/CC ,/, and/CC sometimes/RB two/CD people/NNS at/IN the/DT most/RBS so/RB ,/, uh/UH ./.">, <A.79: u"Uh-huh/UH ./. What/WP 's/BES her/PRP$ name/NN by/IN the/DT way/NN ?/.">, <B.80: u'Uh/UH ,/, pardon/UH ?/.'>, <A.81: u'What/WP ,/, what/WP do/VBP you/PRP call/VB the/DT dog/NN ?/.'>, <B.82: u"Oh/UH ,/, it/PRP 's/BES ,/, uh/UH ,/, Mitzi/NNP ./.">, <A.83: u'Mitzi/NNP ./.'>, <B.84: u'Yeah/UH ./.'>, <A.85: u'Mine/PRP is/VBZ Gin/NNP ./.'>, <B.86: u'Oh/UH ,/, okay/UH ./.'>, <A.87: u'As/IN in/IN ,/, uh/UH ,/, martini/NN ./.'>, <B.88: u'Yeah/UH ./.'>, <A.89: u"Actually/UH ,/, it/PRP 's/BES Gin/NNP two/CD ./.">, <B.90: u'I/PRP ,/, I/PRP see/VBP ./.'>, <A.91: u"Because/IN ,/, uh/UH ,/, when/WRB I/PRP was/VBD a/DT teenager/NN ,/, in/IN high/JJ school/NN ,/, I/PRP had/VBD Gin/NNP one/CD ,/, but/CC then/RB when/WRB I/PRP went/VBD out/IN in/IN the/DT world/NN ,/, I/PRP could/MD n't/RB take/VB her/PRP with/IN me/PRP ./.">, <B.92: u'Yeah/UH ,/, yeah/UH ,/, yeah/UH ,/, I/PRP had/VBD ,/, uh/UH ,/, a/DT similar/JJ ,/, uh/UH ,/, experience/NN ./. I/PRP ,/, I/PRP grew/VBD up/RP on/IN a/DT farm/NN so/RB I/PRP always/RB had/VBD ,/, uh/UH ,/, outdoor/JJ pets/NNS and/CC ,/,'>, <A.93: u'Uh-huh/UH ./.'>, <B.94: u'uh/UH ,/, the/DT dog/NN I/PRP had/VBD when/WRB I/PRP moved/VBD to/IN Dallas/NNP about/RB five/CD years/NNS ago/RB was/VBD a/DT ,/, uh/UH ,/, Springer/NNP Spaniel/NNP ,/, black/JJ Lab/NNP cross/NN ./. And/CC he/PRP was/VBD a/DT real/RB ,/, a/DT real/RB lovable/JJ type/NN ,/,'>, <A.95: u'Oh/UH yeah/UH ./.'>, <B.96: u'but/CC ,/, uh/UH ,/, definitely/RB not/RB an/DT apartment/NN type/NN animal/NN so/RB he/PRP ,/, uh/UH ,/, he/PRP had/VBD to/TO stay/VB home/NN ./.'>, <A.97: u'Oh/UH ,/, what/WP a/DT shame/NN ./.'>, <B.98: u'Yeah/UH ,/, yeah/UH ,/, it/PRP really/RB was/VBD ./. He/PRP ,/, uh/UH ,/, apparently/RB had/VBD a/DT tough/JJ time/NN with/IN it/PRP for/IN a/DT little/JJ while/IN and/CC then/RB he/PRP ,/, he/PRP got/VBD ,/, he/PRP came/VBD to/TO accept/VB the/DT fact/NN that/DT Mom/NNP and/CC Dad/NNP were/VBD his/PRP$ company/NN from/IN then/RB on/RB ,/, but/CC ,/, uh/UH ./.'>, <A.99: u'Uh-huh/UH ,/, oh/UH well/UH ./.'>, <B.100: u'Yeah/UH ./.'>, <A.101: u"Well/UH Randy/UH ,/, we/PRP 've/VBP just/RB about/IN used/VBN up/RP our/PRP$ time/NN here/RB ,/,">, <B.102: u'Okay/UH ./.'>, <A.103: u'and/CC I/PRP must/MD say/VB it/PRP was/VBD interesting/JJ ./.'>, <B.104: u'Most/JJS definitely/RB ./.'>, <A.105: u'I/PRP enjoyed/VBD talking/VBG about/IN pets/NNS with/IN you/PRP ./.'>, <B.106: u"Well/UH that/DT 's/BES great/JJ ./.">, <A.107: u"Maybe/RB we/PRP 'll/MD get/VB together/RB again/RB in/IN the/DT future/NN ./.">, <B.108: u'That/DT sounds/VBZ real/RB good/JJ ./.'>, <A.109: u'Take/VB care/NN now/RB ./.'>, <B.110: u'You/PRP too/RB ./.'>, ...]

然后您可能想查看您的 nltk_data 目录并查看下载的示例数据:

ls -r *
switchboard.zip

switchboard:
transcript      tagged          discourse
timed-transcript    disfluency      README

我仍在研究如何将完整的配电盘数据集导入 NLTK，如果我弄清楚了，我会通知您。但这些示例应该足以让您入门。