1 | #! /usr/bin/env python
|
---|
2 |
|
---|
3 | class Markov:
|
---|
4 | def __init__(self, histsize, choice):
|
---|
5 | self.histsize = histsize
|
---|
6 | self.choice = choice
|
---|
7 | self.trans = {}
|
---|
8 |
|
---|
9 | def add(self, state, next):
|
---|
10 | self.trans.setdefault(state, []).append(next)
|
---|
11 |
|
---|
12 | def put(self, seq):
|
---|
13 | n = self.histsize
|
---|
14 | add = self.add
|
---|
15 | add(None, seq[:0])
|
---|
16 | for i in range(len(seq)):
|
---|
17 | add(seq[max(0, i-n):i], seq[i:i+1])
|
---|
18 | add(seq[len(seq)-n:], None)
|
---|
19 |
|
---|
20 | def get(self):
|
---|
21 | choice = self.choice
|
---|
22 | trans = self.trans
|
---|
23 | n = self.histsize
|
---|
24 | seq = choice(trans[None])
|
---|
25 | while True:
|
---|
26 | subseq = seq[max(0, len(seq)-n):]
|
---|
27 | options = trans[subseq]
|
---|
28 | next = choice(options)
|
---|
29 | if not next:
|
---|
30 | break
|
---|
31 | seq += next
|
---|
32 | return seq
|
---|
33 |
|
---|
34 |
|
---|
35 | def test():
|
---|
36 | import sys, random, getopt
|
---|
37 | args = sys.argv[1:]
|
---|
38 | try:
|
---|
39 | opts, args = getopt.getopt(args, '0123456789cdwq')
|
---|
40 | except getopt.error:
|
---|
41 | print 'Usage: %s [-#] [-cddqw] [file] ...' % sys.argv[0]
|
---|
42 | print 'Options:'
|
---|
43 | print '-#: 1-digit history size (default 2)'
|
---|
44 | print '-c: characters (default)'
|
---|
45 | print '-w: words'
|
---|
46 | print '-d: more debugging output'
|
---|
47 | print '-q: no debugging output'
|
---|
48 | print 'Input files (default stdin) are split in paragraphs'
|
---|
49 | print 'separated blank lines and each paragraph is split'
|
---|
50 | print 'in words by whitespace, then reconcatenated with'
|
---|
51 | print 'exactly one space separating words.'
|
---|
52 | print 'Output consists of paragraphs separated by blank'
|
---|
53 | print 'lines, where lines are no longer than 72 characters.'
|
---|
54 | sys.exit(2)
|
---|
55 | histsize = 2
|
---|
56 | do_words = False
|
---|
57 | debug = 1
|
---|
58 | for o, a in opts:
|
---|
59 | if '-0' <= o <= '-9': histsize = int(o[1:])
|
---|
60 | if o == '-c': do_words = False
|
---|
61 | if o == '-d': debug += 1
|
---|
62 | if o == '-q': debug = 0
|
---|
63 | if o == '-w': do_words = True
|
---|
64 | if not args:
|
---|
65 | args = ['-']
|
---|
66 |
|
---|
67 | m = Markov(histsize, random.choice)
|
---|
68 | try:
|
---|
69 | for filename in args:
|
---|
70 | if filename == '-':
|
---|
71 | f = sys.stdin
|
---|
72 | if f.isatty():
|
---|
73 | print 'Sorry, need stdin from file'
|
---|
74 | continue
|
---|
75 | else:
|
---|
76 | f = open(filename, 'r')
|
---|
77 | if debug: print 'processing', filename, '...'
|
---|
78 | text = f.read()
|
---|
79 | f.close()
|
---|
80 | paralist = text.split('\n\n')
|
---|
81 | for para in paralist:
|
---|
82 | if debug > 1: print 'feeding ...'
|
---|
83 | words = para.split()
|
---|
84 | if words:
|
---|
85 | if do_words:
|
---|
86 | data = tuple(words)
|
---|
87 | else:
|
---|
88 | data = ' '.join(words)
|
---|
89 | m.put(data)
|
---|
90 | except KeyboardInterrupt:
|
---|
91 | print 'Interrupted -- continue with data read so far'
|
---|
92 | if not m.trans:
|
---|
93 | print 'No valid input files'
|
---|
94 | return
|
---|
95 | if debug: print 'done.'
|
---|
96 |
|
---|
97 | if debug > 1:
|
---|
98 | for key in m.trans.keys():
|
---|
99 | if key is None or len(key) < histsize:
|
---|
100 | print repr(key), m.trans[key]
|
---|
101 | if histsize == 0: print repr(''), m.trans['']
|
---|
102 | print
|
---|
103 | while True:
|
---|
104 | data = m.get()
|
---|
105 | if do_words:
|
---|
106 | words = data
|
---|
107 | else:
|
---|
108 | words = data.split()
|
---|
109 | n = 0
|
---|
110 | limit = 72
|
---|
111 | for w in words:
|
---|
112 | if n + len(w) > limit:
|
---|
113 | print
|
---|
114 | n = 0
|
---|
115 | print w,
|
---|
116 | n += len(w) + 1
|
---|
117 | print
|
---|
118 | print
|
---|
119 |
|
---|
120 | if __name__ == "__main__":
|
---|
121 | test()
|
---|