Guido van Rossum | aad6761 | 2000-05-08 17:31:04 +0000 | [diff] [blame] | 1 | """Common operations on Posix pathnames. |
| 2 | |
| 3 | Instead of importing this module directly, import os and refer to |
| 4 | this module as os.path. The "os.path" name is an alias for this |
| 5 | module on Posix systems; on other systems (e.g. Mac, Windows), |
| 6 | os.path provides the same operations in a manner specific to that |
| 7 | platform, and is an alias to another module (e.g. macpath, ntpath). |
| 8 | |
| 9 | Some of this can actually be useful on non-Posix systems too, e.g. |
| 10 | for manipulation of the pathname component of URLs. |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 11 | """ |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 12 | |
Guido van Rossum | 3bb710d | 1996-07-30 16:35:26 +0000 | [diff] [blame] | 13 | import os |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 14 | import stat |
| 15 | |
| 16 | |
| 17 | # Normalize the case of a pathname. Trivial in Posix, string.lower on Mac. |
| 18 | # On MS-DOS this may also turn slashes into backslashes; however, other |
| 19 | # normalizations (such as optimizing '../' away) are not allowed |
| 20 | # (another function should be defined to do that). |
| 21 | |
| 22 | def normcase(s): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 23 | """Normalize case of pathname. Has no effect under Posix""" |
| 24 | return s |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 25 | |
| 26 | |
Jeremy Hylton | a05e293 | 2000-06-28 14:48:01 +0000 | [diff] [blame] | 27 | # Return whether a path is absolute. |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 28 | # Trivial in Posix, harder on the Mac or MS-DOS. |
| 29 | |
| 30 | def isabs(s): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 31 | """Test whether a path is absolute""" |
| 32 | return s[:1] == '/' |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 33 | |
| 34 | |
Guido van Rossum | 228b8e8 | 1997-04-02 06:13:34 +0000 | [diff] [blame] | 35 | # Join pathnames. |
| 36 | # Ignore the previous parts if a part is absolute. |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 37 | # Insert a '/' unless the first part is empty or already ends in '/'. |
| 38 | |
Guido van Rossum | 228b8e8 | 1997-04-02 06:13:34 +0000 | [diff] [blame] | 39 | def join(a, *p): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 40 | """Join two or more pathname components, inserting '/' as needed""" |
| 41 | path = a |
| 42 | for b in p: |
| 43 | if b[:1] == '/': |
| 44 | path = b |
| 45 | elif path == '' or path[-1:] == '/': |
| 46 | path = path + b |
| 47 | else: |
| 48 | path = path + '/' + b |
| 49 | return path |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 50 | |
| 51 | |
| 52 | # Split a path in head (everything up to the last '/') and tail (the |
| 53 | # rest). If the path ends in '/', tail will be empty. If there is no |
| 54 | # '/' in the path, head will be empty. |
| 55 | # Trailing '/'es are stripped from head unless it is the root. |
| 56 | |
| 57 | def split(p): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 58 | """Split a pathname. Returns tuple "(head, tail)" where "tail" is |
| 59 | everything after the final slash. Either part may be empty""" |
| 60 | import string |
| 61 | i = string.rfind(p, '/') + 1 |
| 62 | head, tail = p[:i], p[i:] |
| 63 | if head and head <> '/'*len(head): |
| 64 | while head[-1] == '/': |
| 65 | head = head[:-1] |
| 66 | return head, tail |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 67 | |
| 68 | |
| 69 | # Split a path in root and extension. |
Guido van Rossum | a8763e5 | 1996-08-26 18:33:32 +0000 | [diff] [blame] | 70 | # The extension is everything starting at the last dot in the last |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 71 | # pathname component; the root is everything before that. |
| 72 | # It is always true that root + ext == p. |
| 73 | |
| 74 | def splitext(p): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 75 | """Split the extension from a pathname. Extension is everything from the |
| 76 | last dot to the end. Returns "(root, ext)", either part may be empty""" |
| 77 | root, ext = '', '' |
| 78 | for c in p: |
| 79 | if c == '/': |
| 80 | root, ext = root + ext + c, '' |
| 81 | elif c == '.': |
| 82 | if ext: |
| 83 | root, ext = root + ext, c |
| 84 | else: |
| 85 | ext = c |
| 86 | elif ext: |
| 87 | ext = ext + c |
| 88 | else: |
| 89 | root = root + c |
| 90 | return root, ext |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 91 | |
| 92 | |
| 93 | # Split a pathname into a drive specification and the rest of the |
| 94 | # path. Useful on DOS/Windows/NT; on Unix, the drive is always empty. |
| 95 | |
| 96 | def splitdrive(p): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 97 | """Split a pathname into drive and path. On Posix, drive is always |
| 98 | empty""" |
| 99 | return '', p |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 100 | |
| 101 | |
| 102 | # Return the tail (basename) part of a path. |
| 103 | |
| 104 | def basename(p): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 105 | """Returns the final component of a pathname""" |
| 106 | return split(p)[1] |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 107 | |
| 108 | |
| 109 | # Return the head (dirname) part of a path. |
| 110 | |
| 111 | def dirname(p): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 112 | """Returns the directory component of a pathname""" |
| 113 | return split(p)[0] |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 114 | |
| 115 | |
| 116 | # Return the longest prefix of all list elements. |
| 117 | |
| 118 | def commonprefix(m): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 119 | "Given a list of pathnames, returns the longest common leading component" |
| 120 | if not m: return '' |
| 121 | prefix = m[0] |
| 122 | for item in m: |
| 123 | for i in range(len(prefix)): |
| 124 | if prefix[:i+1] <> item[:i+1]: |
| 125 | prefix = prefix[:i] |
| 126 | if i == 0: return '' |
| 127 | break |
| 128 | return prefix |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 129 | |
| 130 | |
Guido van Rossum | e03c050 | 1998-08-12 02:38:11 +0000 | [diff] [blame] | 131 | # Get size, mtime, atime of files. |
| 132 | |
| 133 | def getsize(filename): |
| 134 | """Return the size of a file, reported by os.stat().""" |
| 135 | st = os.stat(filename) |
| 136 | return st[stat.ST_SIZE] |
| 137 | |
| 138 | def getmtime(filename): |
| 139 | """Return the last modification time of a file, reported by os.stat().""" |
| 140 | st = os.stat(filename) |
| 141 | return st[stat.ST_MTIME] |
| 142 | |
| 143 | def getatime(filename): |
| 144 | """Return the last access time of a file, reported by os.stat().""" |
| 145 | st = os.stat(filename) |
Guido van Rossum | aad6761 | 2000-05-08 17:31:04 +0000 | [diff] [blame] | 146 | return st[stat.ST_ATIME] |
Guido van Rossum | e03c050 | 1998-08-12 02:38:11 +0000 | [diff] [blame] | 147 | |
| 148 | |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 149 | # Is a path a symbolic link? |
Guido van Rossum | 3bb710d | 1996-07-30 16:35:26 +0000 | [diff] [blame] | 150 | # This will always return false on systems where os.lstat doesn't exist. |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 151 | |
| 152 | def islink(path): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 153 | """Test whether a path is a symbolic link""" |
| 154 | try: |
| 155 | st = os.lstat(path) |
| 156 | except (os.error, AttributeError): |
| 157 | return 0 |
| 158 | return stat.S_ISLNK(st[stat.ST_MODE]) |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 159 | |
| 160 | |
| 161 | # Does a path exist? |
| 162 | # This is false for dangling symbolic links. |
| 163 | |
| 164 | def exists(path): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 165 | """Test whether a path exists. Returns false for broken symbolic links""" |
| 166 | try: |
| 167 | st = os.stat(path) |
| 168 | except os.error: |
| 169 | return 0 |
| 170 | return 1 |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 171 | |
| 172 | |
Guido van Rossum | 3bb710d | 1996-07-30 16:35:26 +0000 | [diff] [blame] | 173 | # Is a path a directory? |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 174 | # This follows symbolic links, so both islink() and isdir() can be true |
| 175 | # for the same path. |
| 176 | |
| 177 | def isdir(path): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 178 | """Test whether a path is a directory""" |
| 179 | try: |
| 180 | st = os.stat(path) |
| 181 | except os.error: |
| 182 | return 0 |
| 183 | return stat.S_ISDIR(st[stat.ST_MODE]) |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 184 | |
| 185 | |
| 186 | # Is a path a regular file? |
| 187 | # This follows symbolic links, so both islink() and isfile() can be true |
| 188 | # for the same path. |
| 189 | |
| 190 | def isfile(path): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 191 | """Test whether a path is a regular file""" |
| 192 | try: |
| 193 | st = os.stat(path) |
| 194 | except os.error: |
| 195 | return 0 |
| 196 | return stat.S_ISREG(st[stat.ST_MODE]) |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 197 | |
| 198 | |
| 199 | # Are two filenames really pointing to the same file? |
| 200 | |
| 201 | def samefile(f1, f2): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 202 | """Test whether two pathnames reference the same actual file""" |
| 203 | s1 = os.stat(f1) |
| 204 | s2 = os.stat(f2) |
| 205 | return samestat(s1, s2) |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 206 | |
| 207 | |
| 208 | # Are two open files really referencing the same file? |
| 209 | # (Not necessarily the same file descriptor!) |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 210 | |
| 211 | def sameopenfile(fp1, fp2): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 212 | """Test whether two open file objects reference the same file""" |
| 213 | s1 = os.fstat(fp1) |
| 214 | s2 = os.fstat(fp2) |
| 215 | return samestat(s1, s2) |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 216 | |
| 217 | |
| 218 | # Are two stat buffers (obtained from stat, fstat or lstat) |
| 219 | # describing the same file? |
| 220 | |
| 221 | def samestat(s1, s2): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 222 | """Test whether two stat buffers reference the same file""" |
| 223 | return s1[stat.ST_INO] == s2[stat.ST_INO] and \ |
| 224 | s1[stat.ST_DEV] == s2[stat.ST_DEV] |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 225 | |
| 226 | |
| 227 | # Is a path a mount point? |
Guido van Rossum | 3bb710d | 1996-07-30 16:35:26 +0000 | [diff] [blame] | 228 | # (Does this work for all UNIXes? Is it even guaranteed to work by Posix?) |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 229 | |
| 230 | def ismount(path): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 231 | """Test whether a path is a mount point""" |
| 232 | try: |
| 233 | s1 = os.stat(path) |
| 234 | s2 = os.stat(join(path, '..')) |
| 235 | except os.error: |
| 236 | return 0 # It doesn't exist -- so not a mount point :-) |
| 237 | dev1 = s1[stat.ST_DEV] |
| 238 | dev2 = s2[stat.ST_DEV] |
| 239 | if dev1 != dev2: |
| 240 | return 1 # path/.. on a different device as path |
| 241 | ino1 = s1[stat.ST_INO] |
| 242 | ino2 = s2[stat.ST_INO] |
| 243 | if ino1 == ino2: |
| 244 | return 1 # path/.. is the same i-node as path |
| 245 | return 0 |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 246 | |
| 247 | |
| 248 | # Directory tree walk. |
| 249 | # For each directory under top (including top itself, but excluding |
| 250 | # '.' and '..'), func(arg, dirname, filenames) is called, where |
| 251 | # dirname is the name of the directory and filenames is the list |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 252 | # of files (and subdirectories etc.) in the directory. |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 253 | # The func may modify the filenames list, to implement a filter, |
| 254 | # or to impose a different order of visiting. |
| 255 | |
| 256 | def walk(top, func, arg): |
Guido van Rossum | aad6761 | 2000-05-08 17:31:04 +0000 | [diff] [blame] | 257 | """walk(top,func,arg) calls func(arg, d, files) for each directory "d" |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 258 | in the tree rooted at "top" (including "top" itself). "files" is a list |
| 259 | of all the files and subdirs in directory "d". |
| 260 | """ |
| 261 | try: |
| 262 | names = os.listdir(top) |
| 263 | except os.error: |
| 264 | return |
| 265 | func(arg, top, names) |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 266 | for name in names: |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 267 | name = join(top, name) |
Guido van Rossum | aad6761 | 2000-05-08 17:31:04 +0000 | [diff] [blame] | 268 | st = os.lstat(name) |
| 269 | if stat.S_ISDIR(st[stat.ST_MODE]): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 270 | walk(name, func, arg) |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 271 | |
| 272 | |
| 273 | # Expand paths beginning with '~' or '~user'. |
| 274 | # '~' means $HOME; '~user' means that user's home directory. |
| 275 | # If the path doesn't begin with '~', or if the user or $HOME is unknown, |
| 276 | # the path is returned unchanged (leaving error reporting to whatever |
| 277 | # function is called with the expanded path as argument). |
| 278 | # See also module 'glob' for expansion of *, ? and [...] in pathnames. |
| 279 | # (A function should also be defined to do full *sh-style environment |
| 280 | # variable expansion.) |
| 281 | |
| 282 | def expanduser(path): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 283 | """Expand ~ and ~user constructions. If user or $HOME is unknown, |
| 284 | do nothing""" |
| 285 | if path[:1] <> '~': |
| 286 | return path |
| 287 | i, n = 1, len(path) |
| 288 | while i < n and path[i] <> '/': |
| 289 | i = i+1 |
| 290 | if i == 1: |
| 291 | if not os.environ.has_key('HOME'): |
| 292 | return path |
| 293 | userhome = os.environ['HOME'] |
| 294 | else: |
| 295 | import pwd |
| 296 | try: |
| 297 | pwent = pwd.getpwnam(path[1:i]) |
| 298 | except KeyError: |
| 299 | return path |
| 300 | userhome = pwent[5] |
| 301 | if userhome[-1:] == '/': i = i+1 |
| 302 | return userhome + path[i:] |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 303 | |
| 304 | |
| 305 | # Expand paths containing shell variable substitutions. |
| 306 | # This expands the forms $variable and ${variable} only. |
Jeremy Hylton | a05e293 | 2000-06-28 14:48:01 +0000 | [diff] [blame] | 307 | # Non-existent variables are left unchanged. |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 308 | |
| 309 | _varprog = None |
| 310 | |
| 311 | def expandvars(path): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 312 | """Expand shell variables of form $var and ${var}. Unknown variables |
| 313 | are left unchanged""" |
| 314 | global _varprog |
| 315 | if '$' not in path: |
| 316 | return path |
| 317 | if not _varprog: |
| 318 | import re |
| 319 | _varprog = re.compile(r'\$(\w+|\{[^}]*\})') |
| 320 | i = 0 |
| 321 | while 1: |
| 322 | m = _varprog.search(path, i) |
| 323 | if not m: |
| 324 | break |
| 325 | i, j = m.span(0) |
| 326 | name = m.group(1) |
| 327 | if name[:1] == '{' and name[-1:] == '}': |
| 328 | name = name[1:-1] |
| 329 | if os.environ.has_key(name): |
| 330 | tail = path[j:] |
| 331 | path = path[:i] + os.environ[name] |
| 332 | i = len(path) |
| 333 | path = path + tail |
| 334 | else: |
| 335 | i = j |
| 336 | return path |
Guido van Rossum | 5c97167 | 1996-07-22 15:23:25 +0000 | [diff] [blame] | 337 | |
| 338 | |
| 339 | # Normalize a path, e.g. A//B, A/./B and A/foo/../B all become A/B. |
| 340 | # It should be understood that this may change the meaning of the path |
| 341 | # if it contains symbolic links! |
| 342 | |
| 343 | def normpath(path): |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 344 | """Normalize path, eliminating double slashes, etc.""" |
| 345 | import string |
| 346 | # Treat initial slashes specially |
| 347 | slashes = '' |
| 348 | while path[:1] == '/': |
| 349 | slashes = slashes + '/' |
| 350 | path = path[1:] |
| 351 | comps = string.splitfields(path, '/') |
| 352 | i = 0 |
| 353 | while i < len(comps): |
| 354 | if comps[i] == '.': |
| 355 | del comps[i] |
Guido van Rossum | 64e736b | 1998-10-02 01:23:47 +0000 | [diff] [blame] | 356 | while i < len(comps) and comps[i] == '': |
| 357 | del comps[i] |
Guido van Rossum | b978d18 | 1997-12-09 16:56:41 +0000 | [diff] [blame] | 358 | elif comps[i] == '..' and i > 0 and comps[i-1] not in ('', '..'): |
| 359 | del comps[i-1:i+1] |
| 360 | i = i-1 |
| 361 | elif comps[i] == '' and i > 0 and comps[i-1] <> '': |
| 362 | del comps[i] |
| 363 | else: |
| 364 | i = i+1 |
| 365 | # If the path is now empty, substitute '.' |
| 366 | if not comps and not slashes: |
| 367 | comps.append('.') |
| 368 | return slashes + string.joinfields(comps, '/') |
Guido van Rossum | 2e7840f | 1999-02-09 18:40:13 +0000 | [diff] [blame] | 369 | |
| 370 | |
Guido van Rossum | 2e7840f | 1999-02-09 18:40:13 +0000 | [diff] [blame] | 371 | def abspath(path): |
Guido van Rossum | aad6761 | 2000-05-08 17:31:04 +0000 | [diff] [blame] | 372 | """Return an absolute path.""" |
Guido van Rossum | 2e7840f | 1999-02-09 18:40:13 +0000 | [diff] [blame] | 373 | if not isabs(path): |
| 374 | path = join(os.getcwd(), path) |
| 375 | return normpath(path) |