Python statistics.FrequencyList类代码示例

 def test_freqlist_tokencount(self):
     """Frequency List (count tokens)"""
     global sentences
     f= FrequencyList()
     for sentence in sentences:
         f.append(sentence)
     self.assertEqual(f.total,13) 

 def test_freqlist_typecount(self):
     """Frequency List (count types)"""
     global sentences
     f= FrequencyList()
     for sentence in sentences:
         f.append(sentence)
     self.assertEqual(len(f),9) 

 def test_freqlist_caseinsens(self):
     """Bigram Frequency List (case insensitive)"""
     global sentences
     f= FrequencyList(None, False)
     for sentence in sentences:
         f.append(Windower(sentence,2))
     self.assertTrue(( f[('is','a')] == 2 and  f[('this','is')] == 1))

def buildclasser(file):
    freqlist = FrequencyList()
    f = open(file,'r')
    for line in f:
        line = line.strip()
        freqlist.append(line.split(' '))
    f.close()
    return Classer(freqlist)

        def buildfromtext(self, files, encoding='utf-8'):
            freqlist = FrequencyList()
            if isinstance(files, str): files = [files]
            for filename in files:
                with open(filename, 'r',encoding=encoding) as f:
                    for line in f:
                        tokens = line.strip().split()
                        freqlist.append(tokens)

            self.buildfromfreqlist(freqlist)

def process(filename):
    print >>sys.stderr, "Processing " + filename
    doc = folia.Document(file=filename)

    freqlist = FrequencyList()
    
    if settings.n == 1:
        for word in doc.words():
            text = word.toktext()
            if settings.casesensitive: text = text.lower()
            freqlist.count(text)
    elif settings.sentencemarkers:
        for sentence in doc.sentences():
            for ngram in Windower(sentence.words(), settings.n):
                text = ' '.join([x for x in ngram.toktext() ])
                if settings.casesensitive: text = text.lower()
                freqlist.count(text)                
    else:
        for word in Windower(sentence.words(), settings.n, None, None):
            text = ' '.join([x for x in ngram.toktext() ])
            if settings.casesensitive: text = text.lower()
            freqlist.count(text)                        
                
    if settings.autooutput:                
        if filename[-len(settings.extension) - 1:].lower() == '.' +settings.extension:
            outfilename = filename[:-len(settings.extension) - 1] + '.freqlist'
        else:
            outfilename += '.freqlist'
        freqlist.save(outfilename,True)    
                                       
    return freqlist

        def buildfromfolia(self, files, encoding='utf-8'):
            freqlist = FrequencyList()
            if isinstance(files, str): files = [files]
            for filename in files:
                f = folia.Document(file=filename)
                for sentence in f.sentences():
                    tokens = sentence.toktext().split(' ')
                    freqlist.append(tokens)


            self.buildfromfreqlist(freqlist)

def main():
    try:
        opts, args = getopt.getopt(sys.argv[1:], "o:OE:htspwrq", ["help"])
    except getopt.GetoptError as err:
        print(str(err),file=sys.stderr)
        usage()
        sys.exit(2)

    outputfile = None


    for o, a in opts:
        if o == '-h' or o == '--help':
            usage()
            sys.exit(0)
        elif o == '-e':
            settings.encoding = a
        elif o == '-E':
            settings.extension = a
        elif o == '-o':
            outputfile = a
        elif o == '-O':
            settings.autooutput = True
        elif o == '-s':
            settings.sentencemarkers = True
        elif o == '-r':
            settings.recurse = True
        elif o == '-q':
            settings.ignoreerrors = True
        else:
            raise Exception("No such option: " + o)


    if outputfile: outputfile = io.open(outputfile,'w',encoding=settings.encoding)

    if len(sys.argv) >= 2:
        freqlist = FrequencyList()
        for x in sys.argv[1:]:
            if os.path.isdir(x):
                processdir(x,freqlist)
            elif os.path.isfile(x):
                freqlist += process(x)
            else:
                print("ERROR: File or directory not found: " + x,file=sys.stderr)
                sys.exit(3)
        if outputfile:
            freqlist.save(outputfile, True)
        else:
            for line in freqlist.output("\t", True):
                print(line)
    else:
        print("ERROR: No files specified",file=sys.stderr)
        sys.exit(2)

    def load(self, filename):
        self.freqlistN = FrequencyList(None, self.casesensitive)
        self.freqlistNm1 = FrequencyList(None, self.casesensitive)
        f = io.open(filename,'r',encoding='utf-8')
        mode = False
        for line in f.readlines():
            line = line.strip()
            if line:
                if not mode:
                    if line != "[simplelanguagemodel]":
                        raise Exception("File is not a SimpleLanguageModel")
                    else:
                        mode = 1
                elif mode == 1:
                    if line[:2] == 'n=':
                        self.n = int(line[2:])
                    elif line[:12] == 'beginmarker=':
                        self.beginmarker = line[12:]
                    elif line[:10] == 'endmarker=':
                        self.endmarker = line[10:]
                    elif line[:10] == 'sentences=':
                        self.sentences = int(line[10:])
                    elif line[:14] == 'casesensitive=':
                        self.casesensitive = bool(int(line[14:]))
                        self.freqlistN = FrequencyList(None, self.casesensitive)
                        self.freqlistNm1 = FrequencyList(None, self.casesensitive)
                    elif line == "[freqlistN]":
                        mode = 2
                    else:
                        raise Exception("Syntax error in language model file: ", line)
                elif mode == 2:
                    if line == "[freqlistNm1]":
                        mode = 3
                    else:
                        try:
                            type, count = line.split("\t")
                            self.freqlistN.count(type.split(' '),int(count))
                        except:
                            print("Warning, could not parse line whilst loading frequency list: ", line,file=stderr)
                elif mode == 3:
                        try:
                            type, count = line.split("\t")
                            self.freqlistNm1.count(type.split(' '),int(count))
                        except:
                            print("Warning, could not parse line whilst loading frequency list: ", line,file=stderr)

        if self.beginmarker:
            self._begingram = [self.beginmarker] * (self.n-1)
        if self.endmarker:
            self._endgram = [self.endmarker] * (self.n-1)

    def __init__(self, n=2, casesensitive = True, beginmarker = "<begin>", endmarker = "<end>"):
        self.casesensitive = casesensitive
        self.freqlistN = FrequencyList(None, self.casesensitive)
        self.freqlistNm1 = FrequencyList(None, self.casesensitive)

        assert isinstance(n,int) and n >= 2
        self.n = n
        self.beginmarker = beginmarker
        self.endmarker = endmarker
        self.sentences = 0

        if self.beginmarker:
            self._begingram = tuple([self.beginmarker] * (n-1))
        if self.endmarker:
            self._endgram = tuple([self.endmarker] * (n-1))

def buildclasser():
    global DOTOKENIZE, ENCODING, outputprefix
    log("Counting unigrams (for classer) ...",stream=sys.stderr)
    freqlist = FrequencyList()
    f = open(corpusfile)
    for i, line in enumerate(f):            
        if (i % 10000 == 0): 
            log("\tLine " + str(i+1) + " - (classer construction)", stream=sys.stderr)
        if DOTOKENIZE: 
            line = crude_tokenizer(line.strip())
        line = line.strip().split(' ')
        freqlist.append(['<begin>'] + line + ['<end>'])
    f.close()
    
    log("Building classer ...", stream=sys.stderr)
    classer = Classer(freqlist)
    classer.save(outputprefix + '.cls')
    log("\t" + str(len(classer)) + " classes found", stream=sys.stderr)
    return classer    

def main():
    try:
        opts, files = getopt.getopt(sys.argv[1:], "hn:ie:", ["help"])
    except getopt.GetoptError as err:
        # print help information and exit:
        print(str(err),file=sys.stderr)
        usage()
        sys.exit(2)

    testsetsize = devsetsize = 0
    casesensitive = True
    encoding = 'utf-8'
    n = 1

    for o, a in opts:
        if o == "-n":
            n = int(a)
        elif o == "-i":
            casesensitive =  False
        elif o == "-e":
            encoding = a
        else:
            print("ERROR: Unknown option:",o,file=sys.stderr)
            sys.exit(1)

    if not files:
        print >>sys.stderr, "No files specified"
        sys.exit(1)

    freqlist = FrequencyList(None, casesensitive)
    for filename in files:
        f = codecs.open(filename,'r',encoding)
        for line in f:
            if n > 1:
                freqlist.append(Windower(crude_tokenizer(line),n))
            else:
                freqlist.append(crude_tokenizer(line))

        f.close()

    dist = Distribution(freqlist)
    for type, count in freqlist:
        if isinstance(type,tuple) or isinstance(type,list):
            type = " ".join(type)
        s =  type + "\t" + str(count) + "\t" + str(dist[type]) + "\t" + str(dist.information(type))
        print(s)

    print("Tokens:           ", freqlist.tokens(),file=sys.stderr)
    print("Types:            ", len(freqlist),file=sys.stderr)
    print("Type-token ratio: ", freqlist.typetokenratio(),file=sys.stderr)
    print("Entropy:          ", dist.entropy(),file=sys.stderr)

def process(filename):
    try:
        print("Processing " + filename,file=sys.stderr)
        doc = folia.Document(file=filename)

        freqlist = FrequencyList()

        if settings.n == 1:
            for word in doc.words():
                text = word.toktext()
                if settings.casesensitive: text = text.lower()
                freqlist.count(text)
        elif settings.sentencemarkers:
            for sentence in doc.sentences():
                for ngram in Windower(sentence.words(), settings.n):
                    text = ' '.join([x for x in ngram.toktext() ])
                    if settings.casesensitive: text = text.lower()
                    freqlist.count(text)
        else:
            for word in Windower(sentence.words(), settings.n, None, None):
                text = ' '.join([x for x in ngram.toktext() ])
                if settings.casesensitive: text = text.lower()
                freqlist.count(text)

        if settings.autooutput:
            if filename[-len(settings.extension) - 1:].lower() == '.' +settings.extension:
                outfilename = filename[:-len(settings.extension) - 1] + '.freqlist'
            else:
                outfilename += '.freqlist'
            freqlist.save(outfilename,True)
    except Exception as e:
        if settings.ignoreerrors:
            print("ERROR: An exception was raised whilst processing " + filename, e,file=sys.stderr)
        else:
            raise

    return freqlist

def main():
    parser = argparse.ArgumentParser(description="Generate an n-gram frequency list", formatter_class=argparse.ArgumentDefaultsHelpFormatter)
    parser.add_argument('-n','--ngramsize', help="N-gram size", type=int, action='store',default=1)
    parser.add_argument('-i','--caseinsensitive', help="Case insensitive", action="store_true")
    parser.add_argument('-e','--encoding', help="Character encoding", type=str, action='store',default='utf-8')
    parser.add_argument('files', type=str, nargs='+', help="The data sets to sample from, must be of equal size (i.e., same number of lines)")


    args = parser.parse_args()

    if not args.files:
        print("No files specified", file=sys.stderr)
        sys.exit(1)

    freqlist = FrequencyList(None, args.caseinsensitive)
    for filename in args.files:
        f = io.open(filename,'r',encoding=args.encoding)
        for line in f:
            if args.ngramsize > 1:
                freqlist.append(Windower(crude_tokenizer(line),args.ngramsize))
            else:
                freqlist.append(crude_tokenizer(line))

        f.close()

    dist = Distribution(freqlist)
    for type, count in freqlist:
        if isinstance(type,tuple) or isinstance(type,list):
            type = " ".join(type)
        s =  type + "\t" + str(count) + "\t" + str(dist[type]) + "\t" + str(dist.information(type))
        print(s)

    print("Tokens:           ", freqlist.tokens(),file=sys.stderr)
    print("Types:            ", len(freqlist),file=sys.stderr)
    print("Type-token ratio: ", freqlist.typetokenratio(),file=sys.stderr)
    print("Entropy:          ", dist.entropy(),file=sys.stderr)

#!/usr/bin/env python
#-*- coding:utf-8 -*-

from pynlpl.textprocessors import Classer
from pynlpl.statistics import FrequencyList
import sys

filename = sys.argv[1]

print >>sys.stderr, "Counting tokens"
f = open(filename)
freqlist = FrequencyList()
for i, line in enumerate(f):            
    if (i % 10000 == 0): 
        print >>sys.stderr, "\tLine " + str(i+1)
    line = ['<s>'] + line.strip().split(' ') + ['</s>']
    
    freqlist.append(line)
f.close()

print >>sys.stderr, "Building classer"
classer = Classer(freqlist, filesupport=True )
classer.save(filename + '.cls')

print >>sys.stderr, "Encoding data"
classer.encodefile(filename, filename + '.clsenc')