Mercurial > repos > shellac > guppy_basecaller
comparison env/lib/python3.7/site-packages/distlib/manifest.py @ 5:9b1c78e6ba9c draft default tip
"planemo upload commit 6c0a8142489327ece472c84e558c47da711a9142"
author | shellac |
---|---|
date | Mon, 01 Jun 2020 08:59:25 -0400 |
parents | 79f47841a781 |
children |
comparison
equal
deleted
inserted
replaced
4:79f47841a781 | 5:9b1c78e6ba9c |
---|---|
1 # -*- coding: utf-8 -*- | |
2 # | |
3 # Copyright (C) 2012-2013 Python Software Foundation. | |
4 # See LICENSE.txt and CONTRIBUTORS.txt. | |
5 # | |
6 """ | |
7 Class representing the list of files in a distribution. | |
8 | |
9 Equivalent to distutils.filelist, but fixes some problems. | |
10 """ | |
11 import fnmatch | |
12 import logging | |
13 import os | |
14 import re | |
15 import sys | |
16 | |
17 from . import DistlibException | |
18 from .compat import fsdecode | |
19 from .util import convert_path | |
20 | |
21 | |
22 __all__ = ['Manifest'] | |
23 | |
24 logger = logging.getLogger(__name__) | |
25 | |
26 # a \ followed by some spaces + EOL | |
27 _COLLAPSE_PATTERN = re.compile('\\\\w*\n', re.M) | |
28 _COMMENTED_LINE = re.compile('#.*?(?=\n)|\n(?=$)', re.M | re.S) | |
29 | |
30 # | |
31 # Due to the different results returned by fnmatch.translate, we need | |
32 # to do slightly different processing for Python 2.7 and 3.2 ... this needed | |
33 # to be brought in for Python 3.6 onwards. | |
34 # | |
35 _PYTHON_VERSION = sys.version_info[:2] | |
36 | |
37 class Manifest(object): | |
38 """A list of files built by on exploring the filesystem and filtered by | |
39 applying various patterns to what we find there. | |
40 """ | |
41 | |
42 def __init__(self, base=None): | |
43 """ | |
44 Initialise an instance. | |
45 | |
46 :param base: The base directory to explore under. | |
47 """ | |
48 self.base = os.path.abspath(os.path.normpath(base or os.getcwd())) | |
49 self.prefix = self.base + os.sep | |
50 self.allfiles = None | |
51 self.files = set() | |
52 | |
53 # | |
54 # Public API | |
55 # | |
56 | |
57 def findall(self): | |
58 """Find all files under the base and set ``allfiles`` to the absolute | |
59 pathnames of files found. | |
60 """ | |
61 from stat import S_ISREG, S_ISDIR, S_ISLNK | |
62 | |
63 self.allfiles = allfiles = [] | |
64 root = self.base | |
65 stack = [root] | |
66 pop = stack.pop | |
67 push = stack.append | |
68 | |
69 while stack: | |
70 root = pop() | |
71 names = os.listdir(root) | |
72 | |
73 for name in names: | |
74 fullname = os.path.join(root, name) | |
75 | |
76 # Avoid excess stat calls -- just one will do, thank you! | |
77 stat = os.stat(fullname) | |
78 mode = stat.st_mode | |
79 if S_ISREG(mode): | |
80 allfiles.append(fsdecode(fullname)) | |
81 elif S_ISDIR(mode) and not S_ISLNK(mode): | |
82 push(fullname) | |
83 | |
84 def add(self, item): | |
85 """ | |
86 Add a file to the manifest. | |
87 | |
88 :param item: The pathname to add. This can be relative to the base. | |
89 """ | |
90 if not item.startswith(self.prefix): | |
91 item = os.path.join(self.base, item) | |
92 self.files.add(os.path.normpath(item)) | |
93 | |
94 def add_many(self, items): | |
95 """ | |
96 Add a list of files to the manifest. | |
97 | |
98 :param items: The pathnames to add. These can be relative to the base. | |
99 """ | |
100 for item in items: | |
101 self.add(item) | |
102 | |
103 def sorted(self, wantdirs=False): | |
104 """ | |
105 Return sorted files in directory order | |
106 """ | |
107 | |
108 def add_dir(dirs, d): | |
109 dirs.add(d) | |
110 logger.debug('add_dir added %s', d) | |
111 if d != self.base: | |
112 parent, _ = os.path.split(d) | |
113 assert parent not in ('', '/') | |
114 add_dir(dirs, parent) | |
115 | |
116 result = set(self.files) # make a copy! | |
117 if wantdirs: | |
118 dirs = set() | |
119 for f in result: | |
120 add_dir(dirs, os.path.dirname(f)) | |
121 result |= dirs | |
122 return [os.path.join(*path_tuple) for path_tuple in | |
123 sorted(os.path.split(path) for path in result)] | |
124 | |
125 def clear(self): | |
126 """Clear all collected files.""" | |
127 self.files = set() | |
128 self.allfiles = [] | |
129 | |
130 def process_directive(self, directive): | |
131 """ | |
132 Process a directive which either adds some files from ``allfiles`` to | |
133 ``files``, or removes some files from ``files``. | |
134 | |
135 :param directive: The directive to process. This should be in a format | |
136 compatible with distutils ``MANIFEST.in`` files: | |
137 | |
138 http://docs.python.org/distutils/sourcedist.html#commands | |
139 """ | |
140 # Parse the line: split it up, make sure the right number of words | |
141 # is there, and return the relevant words. 'action' is always | |
142 # defined: it's the first word of the line. Which of the other | |
143 # three are defined depends on the action; it'll be either | |
144 # patterns, (dir and patterns), or (dirpattern). | |
145 action, patterns, thedir, dirpattern = self._parse_directive(directive) | |
146 | |
147 # OK, now we know that the action is valid and we have the | |
148 # right number of words on the line for that action -- so we | |
149 # can proceed with minimal error-checking. | |
150 if action == 'include': | |
151 for pattern in patterns: | |
152 if not self._include_pattern(pattern, anchor=True): | |
153 logger.warning('no files found matching %r', pattern) | |
154 | |
155 elif action == 'exclude': | |
156 for pattern in patterns: | |
157 found = self._exclude_pattern(pattern, anchor=True) | |
158 #if not found: | |
159 # logger.warning('no previously-included files ' | |
160 # 'found matching %r', pattern) | |
161 | |
162 elif action == 'global-include': | |
163 for pattern in patterns: | |
164 if not self._include_pattern(pattern, anchor=False): | |
165 logger.warning('no files found matching %r ' | |
166 'anywhere in distribution', pattern) | |
167 | |
168 elif action == 'global-exclude': | |
169 for pattern in patterns: | |
170 found = self._exclude_pattern(pattern, anchor=False) | |
171 #if not found: | |
172 # logger.warning('no previously-included files ' | |
173 # 'matching %r found anywhere in ' | |
174 # 'distribution', pattern) | |
175 | |
176 elif action == 'recursive-include': | |
177 for pattern in patterns: | |
178 if not self._include_pattern(pattern, prefix=thedir): | |
179 logger.warning('no files found matching %r ' | |
180 'under directory %r', pattern, thedir) | |
181 | |
182 elif action == 'recursive-exclude': | |
183 for pattern in patterns: | |
184 found = self._exclude_pattern(pattern, prefix=thedir) | |
185 #if not found: | |
186 # logger.warning('no previously-included files ' | |
187 # 'matching %r found under directory %r', | |
188 # pattern, thedir) | |
189 | |
190 elif action == 'graft': | |
191 if not self._include_pattern(None, prefix=dirpattern): | |
192 logger.warning('no directories found matching %r', | |
193 dirpattern) | |
194 | |
195 elif action == 'prune': | |
196 if not self._exclude_pattern(None, prefix=dirpattern): | |
197 logger.warning('no previously-included directories found ' | |
198 'matching %r', dirpattern) | |
199 else: # pragma: no cover | |
200 # This should never happen, as it should be caught in | |
201 # _parse_template_line | |
202 raise DistlibException( | |
203 'invalid action %r' % action) | |
204 | |
205 # | |
206 # Private API | |
207 # | |
208 | |
209 def _parse_directive(self, directive): | |
210 """ | |
211 Validate a directive. | |
212 :param directive: The directive to validate. | |
213 :return: A tuple of action, patterns, thedir, dir_patterns | |
214 """ | |
215 words = directive.split() | |
216 if len(words) == 1 and words[0] not in ('include', 'exclude', | |
217 'global-include', | |
218 'global-exclude', | |
219 'recursive-include', | |
220 'recursive-exclude', | |
221 'graft', 'prune'): | |
222 # no action given, let's use the default 'include' | |
223 words.insert(0, 'include') | |
224 | |
225 action = words[0] | |
226 patterns = thedir = dir_pattern = None | |
227 | |
228 if action in ('include', 'exclude', | |
229 'global-include', 'global-exclude'): | |
230 if len(words) < 2: | |
231 raise DistlibException( | |
232 '%r expects <pattern1> <pattern2> ...' % action) | |
233 | |
234 patterns = [convert_path(word) for word in words[1:]] | |
235 | |
236 elif action in ('recursive-include', 'recursive-exclude'): | |
237 if len(words) < 3: | |
238 raise DistlibException( | |
239 '%r expects <dir> <pattern1> <pattern2> ...' % action) | |
240 | |
241 thedir = convert_path(words[1]) | |
242 patterns = [convert_path(word) for word in words[2:]] | |
243 | |
244 elif action in ('graft', 'prune'): | |
245 if len(words) != 2: | |
246 raise DistlibException( | |
247 '%r expects a single <dir_pattern>' % action) | |
248 | |
249 dir_pattern = convert_path(words[1]) | |
250 | |
251 else: | |
252 raise DistlibException('unknown action %r' % action) | |
253 | |
254 return action, patterns, thedir, dir_pattern | |
255 | |
256 def _include_pattern(self, pattern, anchor=True, prefix=None, | |
257 is_regex=False): | |
258 """Select strings (presumably filenames) from 'self.files' that | |
259 match 'pattern', a Unix-style wildcard (glob) pattern. | |
260 | |
261 Patterns are not quite the same as implemented by the 'fnmatch' | |
262 module: '*' and '?' match non-special characters, where "special" | |
263 is platform-dependent: slash on Unix; colon, slash, and backslash on | |
264 DOS/Windows; and colon on Mac OS. | |
265 | |
266 If 'anchor' is true (the default), then the pattern match is more | |
267 stringent: "*.py" will match "foo.py" but not "foo/bar.py". If | |
268 'anchor' is false, both of these will match. | |
269 | |
270 If 'prefix' is supplied, then only filenames starting with 'prefix' | |
271 (itself a pattern) and ending with 'pattern', with anything in between | |
272 them, will match. 'anchor' is ignored in this case. | |
273 | |
274 If 'is_regex' is true, 'anchor' and 'prefix' are ignored, and | |
275 'pattern' is assumed to be either a string containing a regex or a | |
276 regex object -- no translation is done, the regex is just compiled | |
277 and used as-is. | |
278 | |
279 Selected strings will be added to self.files. | |
280 | |
281 Return True if files are found. | |
282 """ | |
283 # XXX docstring lying about what the special chars are? | |
284 found = False | |
285 pattern_re = self._translate_pattern(pattern, anchor, prefix, is_regex) | |
286 | |
287 # delayed loading of allfiles list | |
288 if self.allfiles is None: | |
289 self.findall() | |
290 | |
291 for name in self.allfiles: | |
292 if pattern_re.search(name): | |
293 self.files.add(name) | |
294 found = True | |
295 return found | |
296 | |
297 def _exclude_pattern(self, pattern, anchor=True, prefix=None, | |
298 is_regex=False): | |
299 """Remove strings (presumably filenames) from 'files' that match | |
300 'pattern'. | |
301 | |
302 Other parameters are the same as for 'include_pattern()', above. | |
303 The list 'self.files' is modified in place. Return True if files are | |
304 found. | |
305 | |
306 This API is public to allow e.g. exclusion of SCM subdirs, e.g. when | |
307 packaging source distributions | |
308 """ | |
309 found = False | |
310 pattern_re = self._translate_pattern(pattern, anchor, prefix, is_regex) | |
311 for f in list(self.files): | |
312 if pattern_re.search(f): | |
313 self.files.remove(f) | |
314 found = True | |
315 return found | |
316 | |
317 def _translate_pattern(self, pattern, anchor=True, prefix=None, | |
318 is_regex=False): | |
319 """Translate a shell-like wildcard pattern to a compiled regular | |
320 expression. | |
321 | |
322 Return the compiled regex. If 'is_regex' true, | |
323 then 'pattern' is directly compiled to a regex (if it's a string) | |
324 or just returned as-is (assumes it's a regex object). | |
325 """ | |
326 if is_regex: | |
327 if isinstance(pattern, str): | |
328 return re.compile(pattern) | |
329 else: | |
330 return pattern | |
331 | |
332 if _PYTHON_VERSION > (3, 2): | |
333 # ditch start and end characters | |
334 start, _, end = self._glob_to_re('_').partition('_') | |
335 | |
336 if pattern: | |
337 pattern_re = self._glob_to_re(pattern) | |
338 if _PYTHON_VERSION > (3, 2): | |
339 assert pattern_re.startswith(start) and pattern_re.endswith(end) | |
340 else: | |
341 pattern_re = '' | |
342 | |
343 base = re.escape(os.path.join(self.base, '')) | |
344 if prefix is not None: | |
345 # ditch end of pattern character | |
346 if _PYTHON_VERSION <= (3, 2): | |
347 empty_pattern = self._glob_to_re('') | |
348 prefix_re = self._glob_to_re(prefix)[:-len(empty_pattern)] | |
349 else: | |
350 prefix_re = self._glob_to_re(prefix) | |
351 assert prefix_re.startswith(start) and prefix_re.endswith(end) | |
352 prefix_re = prefix_re[len(start): len(prefix_re) - len(end)] | |
353 sep = os.sep | |
354 if os.sep == '\\': | |
355 sep = r'\\' | |
356 if _PYTHON_VERSION <= (3, 2): | |
357 pattern_re = '^' + base + sep.join((prefix_re, | |
358 '.*' + pattern_re)) | |
359 else: | |
360 pattern_re = pattern_re[len(start): len(pattern_re) - len(end)] | |
361 pattern_re = r'%s%s%s%s.*%s%s' % (start, base, prefix_re, sep, | |
362 pattern_re, end) | |
363 else: # no prefix -- respect anchor flag | |
364 if anchor: | |
365 if _PYTHON_VERSION <= (3, 2): | |
366 pattern_re = '^' + base + pattern_re | |
367 else: | |
368 pattern_re = r'%s%s%s' % (start, base, pattern_re[len(start):]) | |
369 | |
370 return re.compile(pattern_re) | |
371 | |
372 def _glob_to_re(self, pattern): | |
373 """Translate a shell-like glob pattern to a regular expression. | |
374 | |
375 Return a string containing the regex. Differs from | |
376 'fnmatch.translate()' in that '*' does not match "special characters" | |
377 (which are platform-specific). | |
378 """ | |
379 pattern_re = fnmatch.translate(pattern) | |
380 | |
381 # '?' and '*' in the glob pattern become '.' and '.*' in the RE, which | |
382 # IMHO is wrong -- '?' and '*' aren't supposed to match slash in Unix, | |
383 # and by extension they shouldn't match such "special characters" under | |
384 # any OS. So change all non-escaped dots in the RE to match any | |
385 # character except the special characters (currently: just os.sep). | |
386 sep = os.sep | |
387 if os.sep == '\\': | |
388 # we're using a regex to manipulate a regex, so we need | |
389 # to escape the backslash twice | |
390 sep = r'\\\\' | |
391 escaped = r'\1[^%s]' % sep | |
392 pattern_re = re.sub(r'((?<!\\)(\\\\)*)\.', escaped, pattern_re) | |
393 return pattern_re |