nltk /tools/find_deprecated.py

Language Python Lines 227
MD5 Hash d36d548282ec51d8714a5443b4920d08
Repository https://github.com/BrucePHill/nltk.git View Raw File
  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
#!/usr/bin/env python
#
## Natural Language Toolkit: Deprecated Function & Class Finder
#
# Copyright (C) 2001-2013 NLTK Project
# Author: Edward Loper <edloper@gradient.cis.upenn.edu>
# URL: <http://www.nltk.org/>
# For license information, see LICENSE.TXT

from __future__ import print_statement

"""
This command-line tool takes a list of python files or directories,
and searches them for calls to deprecated NLTK functions, or uses of
deprecated NLTK classes.  For each use of a deprecated object it
finds, it will print out a warning containing the offending line, as
well as its line number and containing file name.  If the terminal has
color support (and if epydoc is installed), then the offending
identifier will be highlighted in red.
"""

######################################################################
# Imports
######################################################################

import os, re, sys, tokenize, textwrap
import nltk, nltk.corpus
from doctest import DocTestParser, register_optionflag
from cStringIO import StringIO
from nltk import defaultdict

######################################################################
# Regexps
######################################################################

#: A little over-simplified, but it'll do.
STRING_PAT = (r'\s*[ur]{0,2}(?:'
              '"""[\s\S]*?"""|'  '"[^"\n]+?"|'
              "'''[\s\S]*?'''|"  "'[^'\n]+?'" ")\s*")
STRING_RE = re.compile(STRING_PAT)

STRINGS_PAT = '%s(?:[+]?%s)*' % (STRING_PAT, STRING_PAT)
STRINGS_RE = re.compile(STRINGS_PAT)

# Define a regexp to search for deprecated definitions.
DEPRECATED_DEF_PAT = (
    r'^\s*@deprecated\s*\(\s*(%s)\s*\)\s*\n+' % STRINGS_PAT +
    r'\s*def\s*(\w+).*' +
    r'|' +
    r'^\s*class\s+(\w+)\s*\(.*Deprecated.*\):\s*')
DEPRECATED_DEF_RE = re.compile(DEPRECATED_DEF_PAT, re.MULTILINE)

CORPUS_READ_METHOD_RE = re.compile(
    '(%s)\.read\(' % ('|'.join(re.escape(n) for n in dir(nltk.corpus))))

CLASS_DEF_RE = re.compile('^\s*class\s+(\w+)\s*[:\(]')

######################################################################
# Globals
######################################################################
# Yes, it's bad programming practice, but this is a little hack
# script. :)  These get initialized by find_deprecated_defs.

deprecated_funcs = defaultdict(set)
deprecated_classes = defaultdict(set)
deprecated_methods = defaultdict(set)

try:
    from epydoc.cli import TerminalController
except ImportError:
    class TerminalController:
        def __getattr__(self, attr): return ''

term = TerminalController()

######################################################################
# Code
######################################################################

# If we're using py24, then ignore the +SKIP directive.
if sys.version_info[:2] < (2,5): register_optionflag('SKIP')

def strip_quotes(s):
    s = s.strip()
    while (s and (s[0] in "ur") and (s[-1] in "'\"")):
        s = s[1:]
    while (s and (s[0] in "'\"" and (s[0] == s[-1]))):
        s = s[1:-1]
    s = s.strip()
    return s

def find_class(s, index):
    lines = s[:index].split('\n')
    while lines:
        m = CLASS_DEF_RE.match(lines[-1])
        if m: return m.group(1)+'.'
        lines.pop()
    return '?.'

def find_deprecated_defs(pkg_dir):
    """
    Return a list of all functions marked with the @deprecated
    decorator, and classes with an immediate Deprecated base class, in
    all Python files in the given directory.
    """
    # Walk through the directory, finding python files.
    for root, dirs, files in os.walk(pkg_dir):
        for filename in files:
            if filename.endswith('.py'):
                # Search the file for any deprecated definitions.
                s = open(os.path.join(root, filename)).read()
                for m in DEPRECATED_DEF_RE.finditer(s):
                    if m.group(2):
                        name = m.group(2)
                        msg = ' '.join(strip_quotes(s) for s in
                                       STRING_RE.findall(m.group(1)))
                        msg = ' '.join(msg.split())
                        if m.group()[0] in ' \t':
                            cls = find_class(s, m.start())
                            deprecated_methods[name].add( (msg, cls, '()') )
                        else:
                            deprecated_funcs[name].add( (msg, '', '()') )
                    else:
                        name = m.group(3)
                        m2 = STRING_RE.match(s, m.end())
                        if m2: msg = strip_quotes(m2.group())
                        else: msg = ''
                        msg = ' '.join(msg.split())
                        deprecated_classes[name].add( (msg, '', ''))

def print_deprecated_uses(paths):
    dep_names = set()
    dep_files = set()
    for path in sorted(paths):
        if os.path.isdir(path):
            dep_names.update(print_deprecated_uses(
                [os.path.join(path,f) for f in os.listdir(path)]))
        elif path.endswith('.py'):
            print_deprecated_uses_in(open(path).readline, path,
                                     dep_files, dep_names, 0)
        elif path.endswith('.doctest') or path.endswith('.txt'):
            for example in DocTestParser().get_examples(open(path).read()):
                ex = StringIO(example.source)
                try:
                    print_deprecated_uses_in(ex.readline, path, dep_files,
                                             dep_names, example.lineno)
                except tokenize.TokenError:
                    print(term.RED + 'Caught TokenError -- '
                           'malformatted doctest?' + term.NORMAL)
    return dep_names

def print_deprecated_uses_in(readline, path, dep_files, dep_names,
                             lineno_offset):
    tokiter = tokenize.generate_tokens(readline)
    context = ['']
    for (typ, tok, start, end, line) in tokiter:
        # Remember the previous line -- it might contain
        # the @deprecated decorator.
        if line is not context[-1]:
            context.append(line)
            if len(context) > 10: del context[0]
        esctok = re.escape(tok)
        # Ignore all tokens except deprecated names.
        if not (tok in deprecated_classes or
                (tok in deprecated_funcs and
                 re.search(r'\b%s\s*\(' % esctok, line)) or
                (tok in deprecated_methods and
                 re.search(r'(?!<\bself)[.]\s*%s\s*\(' % esctok, line))):
            continue
        # Hack: only complain about read if it's used after a corpus.
        if tok == 'read' and not CORPUS_READ_METHOD_RE.search(line):
            continue
        # Ignore deprecated definitions:
        if DEPRECATED_DEF_RE.search(''.join(context)):
            continue
        # Print a header for the first use in a file:
        if path not in dep_files:
            print('\n'+term.BOLD + path + term.NORMAL)
            print('  %slinenum%s' % (term.YELLOW, term.NORMAL))
            dep_files.add(path)
        # Mark the offending token.
        dep_names.add(tok)
        if term.RED: sub = term.RED+tok+term.NORMAL
        elif term.BOLD: sub = term.BOLD+tok+term.NORMAL
        else: sub = '<<'+tok+'>>'
        line = re.sub(r'\b%s\b' % esctok, sub, line)
        # Print the offending line.
        print('  %s[%5d]%s %s' % (term.YELLOW, start[0]+lineno_offset,
                                  term.NORMAL, line.rstrip()))


def main():
    paths = sys.argv[1:] or ['.']

    print('Importing nltk...')
    try:
        import nltk
    except ImportError:
        print('Unable to import nltk -- check your PYTHONPATH.')
        sys.exit(-1)

    print('Finding definitions of deprecated funtions & classes in nltk...')
    find_deprecated_defs(nltk.__path__[0])

    print('Looking for possible uses of deprecated funcs & classes...')
    dep_names = print_deprecated_uses(paths)

    if not dep_names:
        print('No deprecated funcs or classes found!')
    else:
        print("\n"+term.BOLD+"What you should use instead:"+term.NORMAL)
        for name in sorted(dep_names):
            msgs = deprecated_funcs[name].union(
                deprecated_classes[name]).union(
                deprecated_methods[name])
            for msg, prefix, suffix in msgs:
                print(textwrap.fill(term.RED+prefix+name+suffix+)
                                    term.NORMAL+': '+msg,
                                    width=75, initial_indent=' '*2,
                                    subsequent_indent=' '*6)



if __name__ == '__main__':
    main()
Back to Top