<Python Text Processing with NLTK 2.0 Cookbook>代码笔记

如下是<Python Text Processing with NLTK 2.0 Cookbook>一书部分章节的代码笔记.

Tokenizing text into sentences

>>> para = "Hello World. It's good to see you. Thanks for buying this book."

>>> from nltk.tokenize import sent_tokenize

>>> sent_tokenize(para) # "sent_tokenize"是一个函数，下文很多中间带下划线的标识符都指的是函数。

['Hello World.', "It's good to see you.", 'Thanks for buying this

book.']

Tokenizing sentences into words

>>> from nltk.tokenize import word_tokenize

>>> word_tokenize('Hello World.')

['Hello', 'World', '.']

# 等同于

>>> from nltk.tokenize import TreebankWordTokenizer

>>> tokenizer = TreebankWordTokenizer()

>>> tokenizer.tokenize('Hello World.')

['Hello', 'World', '.']

# 等同于

>>> import nltk

>>> text = "Hello. Isn't this fun?"

>>> pattern = r"\w+|[^\w\s]+" # r：regular expression；双引号""可以用单引号''代替；\w表示单词字符，等同于字符集合[a-zA-Z0-9_]；+表示一次或者多次，等同于{1,}，即c+ 和 c{1,} 是一个意思；"|"：二选一，正则表达式中的"或"； [...]：字符集（字符类），其对应的位置可以是字符集中任意字符，例如，a[bcd]表abe、ace和ade；^表示只匹配字符串的开头；\s匹配单个空格，等同于[\f\n\r\t\v]。

>>> print nltk.tokenize.regexp_tokenize(text, pattern)

['Hello', '.', 'Isn', "'", 't', 'this', 'fun', '?']

Tokenizing sentences using regular expressions

>>> from nltk.tokenize import RegexpTokenizer

>>> tokenizer = RegexpTokenizer("[\w']+")

>>> tokenizer.tokenize("Can't is a contraction.")

["Can't", 'is', 'a', 'contraction']

# a simple helper function。

the class.

>>> from nltk.tokenize import regexp_tokenize

>>> regexp_tokenize("Can't is a contraction.", "[\w']+")

["Can't", 'is', 'a', 'contraction']

Training a sentence tokenizer

>>> from nltk.tokenize import PunktSentenceTokenizer

>>> from nltk.corpus import webtext

>>> text = webtext.raw('overheard.txt')

>>> sent_tokenizer = PunktSentenceTokenizer(text)

# Let's compare the results to the default sentence tokenizer, as follows:

>>> sents1 = sent_tokenizer.tokenize(text)

>>> sents1[0] #请注意，索引从零开始：第0 个元素写作sent[0]，其实是第1 个词"word1"；而句子的第9 个元素是"word10"。

'White guy: So, do you have any plans for this evening?'

>>> from nltk.tokenize import sent_tokenize

>>> sents2 = sent_tokenize(text)

>>> sents2[0]

'White guy: So, do you have any plans for this evening?'

>>> sents1[678]

'Girl: But you already have a Big Mac...'

>>> sents2[678]

'Girl: But you already have a Big Mac...\\nHobo: Oh, this is all

theatrical.'

Filtering stopwords in a tokenized sentence

>>> from nltk.corpus import stopwords

>>> english_stops = set(stopwords.words('english'))

>>> words = ["Can't", 'is', 'a', 'contraction']

>>> [word for word in words if word not in english_stops]

["Can't", 'contraction']

Looking up synsets for a word in WordNet

# 方法一

>>> from nltk.corpus import wordnet

>>> syn = wordnet.synsets('cookbook')[0]

>>> syn.name()

'cookbook.n.01'

>>> syn.definition()

'a book of recipes and cooking directions'

# 方法二

>>> from nltk.corpus import wordnet

>>> syn = wordnet.synsets('motorcar')[0]

>>> syn.name

<bound method Synset.name of Synset('car.n.01')>

>>> from nltk.corpus import wordnet as wn

>>> wn.synsets("motorcar") # 括号内可以是单引号

[Synset('car.n.01')]

# WordNet层次结构（词汇关系）——Hypernym /ˈhaɪpənɪm//hyponym/ ˈhaɪpənɪm / relation——上级概念与从属概念的关系

>>> from nltk.corpus import wordnet as wn

>>> motorcar = wn.synset('car.n.01')

>>> types_of_motorcar = motorcar.hyponyms() #等号两边的表达式不能换位，否则会出现警示：can't assign to function call.

>>> types_of_motorcar

[Synset('ambulance.n.01'), Synset('beach_wagon.n.01'), Synset('bus.n.04'), Synset('cab.n.03'), Synset('compact.n.03'), Synset('convertible.n.01') ... Synset('stanley_steamer.n.01'), Synset('stock_car.n.01'), Synset('subcompact.n.01'), Synset('touring_car.n.01'), Synset('used-car.n.01')]

# 部分整体关系（components (meronyms) holonyms）

>>> from nltk.corpus import wordnet as wn

>>> wn.synset('tree.n.01').part_meronyms()

[Synset('burl.n.02'), Synset('crown.n.07'), Synset('limb.n.02'), Synset('stump.n.01'), Synset('trunk.n.01')]

# 反义词关系

>>> wn.lemma('beautiful.a.01.beautiful').antonyms()

[Lemma('ugly.a.01.ugly')]

# 同义词关系

# 查看词汇关系和同义词集上定义的其它方法

>>> dir(wn.synset('beautiful.a.01'))

['__class__', '__delattr__', '__dict__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__gt__' ... 'substance_holonyms', 'substance_meronyms', 'topic_domains', 'tree', 'unicode_repr', 'usage_domains', 'verb_groups', 'wup_similarity']

# Part-of-Speech (POS)

>>> from nltk.corpus import wordnet

>>> syn = wordnet.synsets('motorcar')[0]

>>> syn.pos()

u'n'

Looking up lemmas and synonyms in WordNet

#方法一

>>> from nltk.corpus import wordnet as wn

>>> wn.synset('car.n.01').lemma_names()

[u'car', u'auto', u'automobile', u'machine', u'motorcar'] #结果字符串有一个u 前缀表示它们是Unicode 字符串

>>> u'motorcar'.encode('utf-8')

'motorcar'

#方法二

>>> a = wn.synset('car.n.01').lemma_names()

>>> print a

[u'car', u'auto', u'automobile', u'machine', u'motorcar']

>>> wn.synset('car.n.01').definition ()

u'a motor vehicle with four wheels; usually propelled by an internal combustion engine'

Calculating WordNet synset similarity

）

>>> from nltk.corpus import wordnet as wn

>>> right = wn.synset('right_whale.n.01')

>>> minke = wn.synset('minke_whale.n.01')

>>> right.path_similarity(minke)

0.25

#量度二：wup_similarity -- wup_similarity is short for Wu-Palmer Similarity, which is a scoring method based on how similar the word senses are and where the synsets occur relative to each other in the

hypernym tree.

>>> from nltk.corpus import wordnet

>>> cb = wordnet.synset('cookbook.n.01')

>>> ib = wordnet.synset('instruction_book.n.01')

>>> cb.wup_similarity(ib)

0.9166666666666666

Discovering word collocations (relative to n-gram)

>>> from nltk import bigrams

>>> a = "Jaganadh is testing this application"

>>> tokens = a.split()

>>> bigrams(tokens)

[('Jaganadh', 'is'), ('is', 'testing'), ('testing', 'this'), ('this', 'application.')]

# 如果已分词，则：

>>> bigrams(['more', 'is', 'said', 'than', 'done'])

[('more', 'is'), ('is', 'said'), ('said', 'than'), ('than', 'done')]

词频统计

>>> from nltk.book import text1

*** Introductory Examples for the NLTK Book ***

Loading text1, ..., text9 and sent1, ..., sent9

Type the name of the text or sentence to view it.

Type: 'texts()' or 'sents()' to list the materials.

text3: The Book of Genesis

text4: Inaugural Address Corpus

text5: Chat Corpus

text6: Monty Python and the Holy Grail

text7: Wall Street Journal

text8: Personals Corpus

>>> from nltk import FreqDist

>>> fdist1 = FreqDist(text1)

>>> print fdist1

<FreqDist with 19317 samples and 260819 outcomes>

>>> fdist1

FreqDist({u',': 18713, u'the': 13721, u'.': 6862, u'of': 6536, u'and': 6024, u'a': 4569, u'to': 4542, u';': 4072, u'in': 3916, u'that': 2982, ...})

>>> fdist1.most_common(50)

[(u',', 18713), (u'the', 13721), (u'.', 6862), (u'of', 6536), (u'and', 6024), (u'a', 4569), (u'to', 4542), (u';', 4072), (u'in', 3916), (u'that', 2982), (u"'", 2684), (u'-', 2552), (u'his', 2459), (u'it', 2209), (u'I', 2124), (u's', 1739), (u'is', 1695), (u'he', 1661), (u'with', 1659), (u'was', 1632), (u'as', 1620), (u'"', 1478), (u'all', 1462), (u'for', 1414), (u'this', 1280), (u'!', 1269), (u'at', 1231), (u'by', 1137), (u'but', 1113), (u'not', 1103), (u'--', 1070), (u'him', 1058), (u'from', 1052), (u'be', 1030), (u'on', 1005), (u'so', 918), (u'whale', 906), (u'one', 889), (u'you', 841), (u'had', 767), (u'have', 760), (u'there', 715), (u'But', 705), (u'or', 697), (u'were', 680), (u'now', 646), (u'which', 640), (u'?', 637), (u'me', 627), (u'like', 624)]

# 绘制累积频率图

>>> import matplotlib # 不能直接运行from matplotlib import plot。

>>> fdist1.plot(50, cumulative=True) # 可能处理的时间比较长。

Stemming words

# 单个

>>> from nltk.stem import PorterStemmer # "Poter"是一种词干提取的算法。

>>> stemmer = PorterStemmer()

>>> stemmer.stem('cooking')

u'cook'

>>> stemmer.stem('cookery')

u'cookeri' # The resulting stem is not always a valid word. For example, the

stem of "cookery" is "cookeri". This is a feature, not a bug.

# 多个

>>> import nltk

>>> stemmer = nltk.PorterStemmer()

>>> verbs = ['appears', 'appear', 'appeared', 'calling', 'called']

>>> stems = []

>>> for verb in verbs:

stemmed_verb = stemmer.stem(verb)

stems.append(stemmed_verb) # 一定要按两次回车键，然后再输入下面的语句。

>>> sorted(set(stems))

[u'appear', u'call']

Lemmatizing words with WordNet

>>> from nltk.stem import WordNetLemmatizer

>>> lemmatizer = WordNetLemmatizer()

>>> lemmatizer.lemmatize('cooking')

'cooking'

>>> lemmatizer.lemmatize('cooking', pos='v')

u'cook'

>>> lemmatizer.lemmatize('cookbooks')

u'cookbook'

Replacing words matching regular expressions

# 第一步：新建一个名为"replacers.py"的模块(不是在控制台上操作)，放置于安装Python的目录下的"Lib"文件夹，也可置于安装Python的根目录下，但最好放在"Lib"下，以表明这是一个库（即Python中的模块）。

import re

replacement_patterns = [

(r'won\'t', 'will not'),

(r'can\'t', 'cannot'),

(r'i\'m', 'i am'),

(r'ain\'t', 'is not'),

(r'(\w+)\'ll', '\g<1> will'),

(r'(\w+)n\'t', '\g<1> not'),

(r'(\w+)\'ve', '\g<1> have'),

(r'(\w+)\'s', '\g<1> is'),

(r'(\w+)\'re', '\g<1> are'),

(r'(\w+)\'d', '\g<1> would')

]

class RegexpReplacer(object):

def __init__(self, patterns=replacement_patterns):

self.patterns = [(re.compile(regex), repl) for (regex, repl) in patterns]

def replace(self, text):

s = text

for (pattern, repl) in self.patterns:

s = re.sub(pattern, repl, s)

return s

# 第二步

>>> from replacers import RegexpReplacer

>>> replacer = RegexpReplacer()

>>> replacer.replace("can't is a contraction")

'cannot is a contraction'

>>> replacer.replace("I should've done that thing I didn't do")

'I should have done that thing I did not do'

Accessing Corpora

>>> from nltk.corpus import gutenberg

>>> for filename in gutenberg.fileids():

r = gutenberg.raw(filename)

w = gutenberg.words(filename)

s = gutenberg.sents(filename)

v = set(w)

print filename, len(r)/len(w), len(w)/len(s), len(w)/len(v) # 要按两次回车键才能显示结果。

austen-emma.txt 4 24 24 #语料库的文件名，平均字长，平均句长，每个词平均出现的次数，下同。

...

...

>>> from nltk.book import *

*** Introductory Examples for the NLTK Book ***

Loading text1, ..., text9 and sent1, ..., sent9

Type the name of the text or sentence to view it.

Type: 'texts()' or 'sents()' to list the materials.

...

...

text9: The Man Who Was Thursday by G . K . Chesterton 1908

>>> text1.concordance("monstrous")

Displaying 11 of 11 matches:

ong the former , one was of a most monstrous size . ... This came towards us ,

ON OF THE PSALMS . " Touching that monstrous bulk of the whale or ork we have r

...

...

ght have been rummaged out of this monstrous cabinet there is no telling . But

of Whale - Bones ; for Whales of a monstrous size are oftentimes cast up dead u

外部文档操作

# 外部文档操作

# 读取一个txt文件（已经新建一个"good.txt"文件在D:/Python27下）

>>> f = open('document.txt')

>>> raw = f.read()

#打印"good.txt"内容。# 即使打印Walden这样比较大的文件也比较快

>>> f = open ('good.txt')

>>> print f.read()

happy

Lucy

Lilei

>>> # -*- coding:utf-8 -*-

>>> f = open ('语言学.txt')

>>> print f.read()

计算语言学

自然语言处理

# 建立自己的语料库，并对语料库里的文件进行检索

# 第一步

>>> corpus_root = 'D:/Python27/my own data'

>>> from nltk.corpus import PlaintextCorpusReader

>>> corpus_root = 'D:/Python27/my own data'

>>> wordlists = PlaintextCorpusReader(corpus_root, 'Walden.txt')

>>> wordlists.fileids()

['Walden.txt']

#注意：

>>> from nltk.corpus import PlaintextCorpusReader

>>> corpus_root = 'D:/Python27/my own data'

>>> wordlists = PlaintextCorpusReader(corpus_root, '.*') # .*具有贪婪的性质，首先匹配到不能匹配为止，根据后面的正则表达式，会进行回溯。

>>> wordlists.fileids()

['Gone with the Wind.txt', 'Walden.txt']

# 第二步

>>> n = nltk.word_tokenize(wordlists.raw(fileids="Walden.txt"))

>>> complete_Walden = nltk.Text(n)

>>> complete_Walden.concordance("love")

Displaying 25 of 40 matches:

r even to found a school , but so to love wisdom as to live according to its d

, perhaps we are led oftener by the love of novelty and a regard for the opin

...

...

eed have you to employ punishments ? Love virtue , and the people will be virt

abardine dressed . '' `` Come ye who love , And ye who hate , Children of the

# 获取网络文本

>>> from urllib import urlopen

>>> url = "http://news.bbc.co.uk/2/hi/health/2284783.stm"

>>> html = urlopen(url).read()

>>> html[:60]

'<!doctype html public "-//W3C//DTD HTML 4.0 Transitional//EN'

# 接下来如果输入print html可以看到HTML 的全部内容，包括meta 元标签、图像标签、map 标

签、JavaScript、表单和表格。

# NLTK本来提供了一个辅助函数nltk.clean_html()将HTML 字符串作为参数，返回原始文本；但现在这个函数已经不被支持了，而是用BeautifulSoup的函数get_text()。

>>> import urllib

>>> from bs4 import BeautifulSoup

>>> url = "http://news.bbc.co.uk/2/hi/health/2284783.stm"

>>> html = urllib.urlopen(url).read()

>>> soup = BeautifulSoup(html)

>>> print soup.get_text()

# 对网络文本分词

>>> import urllib

>>> from bs4 import BeautifulSoup

>>> url = "http://news.bbc.co.uk/2/hi/health/2284783.stm"

>>> html = urllib.urlopen(url).read()

>>> soup = BeautifulSoup(html)

>>> raw = BeautifulSoup.get_text(soup)

>>> from nltk.tokenize import word_tokenize

>>> token = nltk.word_tokenize(raw)

>>> token # 可以在token前加上"print"。

[u'BBC', u'NEWS', u'|', u'Health', u'|', u'Blondes', u"'to", u'die', u'out', u'in', u'200', u"years'", u'NEWS', u'SPORT', u'WEATHER', u'WORLD', u'SERVICE', u'A-Z', u'INDEX', ...]

# 重点

>>> s = [u'r118', u'BB']

>>> [str(item) for item in s]

['r118', 'BB']

Default tagging

# 版本一

>>> import nltk

>>> text=nltk.word_tokenize("We are going out.Just you and me.")

>>> print nltk.pos_tag(text)

[('We', 'PRP'), ('are', 'VBP'), ('going', 'VBG'), ('out.Just', 'JJ'), ('you', 'PRP'), ('and', 'CC'), ('me', 'PRP'), ('.', '.')]

# 版本二

>>> sentence = """At eight o'clock on Thursday morning Arthur didn't feel very good."""

>>> tokens = nltk.word_tokenize(sentence)

>>> tokens

['At', 'eight', "o'clock", 'on', 'Thursday', 'morning', 'Arthur', 'did', "n't", 'feel', 'very', 'good', '.']

>>> tagged = nltk.pos_tag(tokens)

>>> tagged

[('At', 'IN'), ('eight', 'CD'), ("o'clock", 'JJ'), ('on', 'IN'), ('Thursday', 'NNP'), ('morning', 'NN'), ('Arthur', 'NNP'), ('did', 'VBD'), ("n't", 'RB'), ('feel', 'VB'), ('very', 'RB'), ('good', 'JJ'), ('.', '.')]

>>> tagged[0:6]

[('At', 'IN'), ('eight', 'CD'), ("o'clock", 'JJ'), ('on', 'IN'), ('Thursday', 'NNP'), ('morning', 'NN')]

Chunking and chinking with regular expressions

# Chunking & Parsing

# Chart Parsing 是描述CFG(Context Free Grammar)语法的一种方法，两者不是平行关系。

import nltk

grammar = r"""

NP: {<DT|PP\$>?<JJ>*<NN>} # chunk determiner/possessive, adjectives and nouns

{<NNP>+} # chunk sequences of proper nouns

"""

cp = nltk.RegexpParser(grammar)

tagged_tokens = [("Rapunzel", "NNP"), ("let", "VBD"), ("down", "RP"), ("her", "PP$"), ("golden", "JJ"), ("hair", "NN")] # 实际运用中，需先对"Rapunzel let down her golden hair."这句进行tokenization.

print cp.parse(tagged_tokens)

# CFG Parsing

>>> import nltk # 这两行代码也可以写成：from nltk import CFG; groucho_grammar = CFG.fromstring("""

>>> groucho_grammar = nltk.CFG.fromstring("""

S -> NP VP

PP -> P NP

NP -> D N | D N PP | 'I'

VP -> V NP | V PP

D -> 'an' | 'a' | 'my' | 'the'

N -> 'elephant' | 'pajamas'

V -> 'shot'

P -> 'in'

""")

>>> sent = "I shot an elephant in my pajamas".split() # 有可能实现就分好词了，即：sent = ['I','shot','an','elephant','in','my','pajamas']

>>> parser = nltk.ChartParser(groucho_grammar) # Chart Parsing 是描述CFG语法的一种方法。

>>> all_the_parses = parser.parse(sent)

>>> all_the_parses

<generator object parses at 0x030E8AD0>

>>> for parse in all_the_parses:

print(parse)

(S

(NP I)

(VP

(V shot)

(NP (D an) (N elephant) (PP (P in) (NP (D my) (N pajamas))))))