blob: f54632ce65e1b1ad424cfc3bfbacdf3cfd697165 [file] [log] [blame]
Guido van Rossum7d5b99d1997-11-21 17:12:59 +00001"""Helper class to quickly write a loop over all standard input files.
2
3Typical use is:
4
5 import fileinput
6 for line in fileinput.input():
7 process(line)
8
9This iterates over the lines of all files listed in sys.argv[1:],
10defaulting to sys.stdin if the list is empty. If a filename is '-' it
11is also replaced by sys.stdin. To specify an alternative list of
12filenames, pass it as the argument to input(). A single file name is
13also allowed.
14
15Functions filename(), lineno() return the filename and cumulative line
16number of the line that has just been read; filelineno() returns its
17line number in the current file; isfirstline() returns true iff the
18line just read is the first line of its file; isstdin() returns true
19iff the line was read from sys.stdin. Function nextfile() closes the
20current file so that the next iteration will read the first line from
21the next file (if any); lines not read from the file will not count
22towards the cumulative line count; the filename is not changed until
23after the first line of the next file has been read. Function close()
24closes the sequence.
25
26Before any lines have been read, filename() returns None and both line
27numbers are zero; nextfile() has no effect. After all lines have been
28read, filename() and the line number functions return the values
29pertaining to the last line read; nextfile() has no effect.
30
Georg Brandlc029f872006-02-19 14:12:34 +000031All files are opened in text mode by default, you can override this by
32setting the mode parameter to input() or FileInput.__init__().
33If an I/O error occurs during opening or reading a file, the IOError
34exception is raised.
Guido van Rossum7d5b99d1997-11-21 17:12:59 +000035
36If sys.stdin is used more than once, the second and further use will
37return no lines, except perhaps for interactive use, or if it has been
38explicitly reset (e.g. using sys.stdin.seek(0)).
39
40Empty files are opened and immediately closed; the only time their
41presence in the list of filenames is noticeable at all is when the
42last file opened is empty.
43
44It is possible that the last line of a file doesn't end in a newline
45character; otherwise lines are returned including the trailing
46newline.
47
48Class FileInput is the implementation; its methods filename(),
49lineno(), fileline(), isfirstline(), isstdin(), nextfile() and close()
50correspond to the functions in the module. In addition it has a
51readline() method which returns the next input line, and a
52__getitem__() method which implements the sequence behavior. The
53sequence must be accessed in strictly sequential order; sequence
54access and readline() cannot be mixed.
55
56Optional in-place filtering: if the keyword argument inplace=1 is
57passed to input() or to the FileInput constructor, the file is moved
58to a backup file and standard output is directed to the input file.
59This makes it possible to write a filter that rewrites its input file
60in place. If the keyword argument backup=".<some extension>" is also
61given, it specifies the extension for the backup file, and the backup
62file remains around; by default, the extension is ".bak" and it is
63deleted when the output file is closed. In-place filtering is
64disabled when standard input is read. XXX The current implementation
65does not work for MS-DOS 8+3 filesystems.
66
67XXX Possible additions:
68
69- optional getopt argument processing
Guido van Rossum7d5b99d1997-11-21 17:12:59 +000070- isatty()
71- read(), read(size), even readlines()
72
73"""
74
Walter Dörwald294bbf32002-06-06 09:48:13 +000075import sys, os
Guido van Rossum7d5b99d1997-11-21 17:12:59 +000076
Skip Montanaroeccd02a2001-01-20 23:34:12 +000077__all__ = ["input","close","nextfile","filename","lineno","filelineno",
78 "isfirstline","isstdin","FileInput"]
79
Guido van Rossum7d5b99d1997-11-21 17:12:59 +000080_state = None
81
Serhiy Storchaka69b7f812016-03-08 18:35:45 +020082# No longer used
Guido van Rossum47955242001-01-05 14:44:39 +000083DEFAULT_BUFSIZE = 8*1024
84
Tim Peters200a5802006-02-19 21:26:07 +000085def input(files=None, inplace=0, backup="", bufsize=0,
Georg Brandlc98eeed2006-02-19 14:57:47 +000086 mode="r", openhook=None):
Terry Jan Reedy35115e62013-06-28 18:59:19 -040087 """Return an instance of the FileInput class, which can be iterated.
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +000088
Terry Jan Reedy35115e62013-06-28 18:59:19 -040089 The parameters are passed to the constructor of the FileInput class.
90 The returned instance, in addition to being an iterator,
91 keeps global state for the functions of this module,.
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +000092 """
Guido van Rossum7d5b99d1997-11-21 17:12:59 +000093 global _state
94 if _state and _state._file:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +000095 raise RuntimeError, "input() already active"
Georg Brandlc98eeed2006-02-19 14:57:47 +000096 _state = FileInput(files, inplace, backup, bufsize, mode, openhook)
Guido van Rossum7d5b99d1997-11-21 17:12:59 +000097 return _state
98
99def close():
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000100 """Close the sequence."""
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000101 global _state
102 state = _state
103 _state = None
104 if state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000105 state.close()
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000106
107def nextfile():
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000108 """
109 Close the current file so that the next iteration will read the first
110 line from the next file (if any); lines not read from the file will
111 not count towards the cumulative line count. The filename is not
112 changed until after the first line of the next file has been read.
113 Before the first line has been read, this function has no effect;
114 it cannot be used to skip the first file. After the last line of the
Tim Peters8ac14952002-05-23 15:15:30 +0000115 last file has been read, this function has no effect.
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000116 """
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000117 if not _state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000118 raise RuntimeError, "no active input()"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000119 return _state.nextfile()
120
121def filename():
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000122 """
123 Return the name of the file currently being read.
Tim Peters8ac14952002-05-23 15:15:30 +0000124 Before the first line has been read, returns None.
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000125 """
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000126 if not _state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000127 raise RuntimeError, "no active input()"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000128 return _state.filename()
129
130def lineno():
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000131 """
132 Return the cumulative line number of the line that has just been read.
133 Before the first line has been read, returns 0. After the last line
Tim Peters8ac14952002-05-23 15:15:30 +0000134 of the last file has been read, returns the line number of that line.
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000135 """
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000136 if not _state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000137 raise RuntimeError, "no active input()"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000138 return _state.lineno()
139
140def filelineno():
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000141 """
142 Return the line number in the current file. Before the first line
143 has been read, returns 0. After the last line of the last file has
Tim Peters8ac14952002-05-23 15:15:30 +0000144 been read, returns the line number of that line within the file.
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000145 """
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000146 if not _state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000147 raise RuntimeError, "no active input()"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000148 return _state.filelineno()
149
Georg Brandl67e9fb92006-02-19 13:56:17 +0000150def fileno():
151 """
152 Return the file number of the current file. When no file is currently
153 opened, returns -1.
154 """
155 if not _state:
156 raise RuntimeError, "no active input()"
157 return _state.fileno()
158
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000159def isfirstline():
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000160 """
161 Returns true the line just read is the first line of its file,
Tim Peters8ac14952002-05-23 15:15:30 +0000162 otherwise returns false.
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000163 """
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000164 if not _state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000165 raise RuntimeError, "no active input()"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000166 return _state.isfirstline()
167
168def isstdin():
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000169 """
170 Returns true if the last line was read from sys.stdin,
Tim Peters8ac14952002-05-23 15:15:30 +0000171 otherwise returns false.
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000172 """
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000173 if not _state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000174 raise RuntimeError, "no active input()"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000175 return _state.isstdin()
176
177class FileInput:
Terry Jan Reedy35115e62013-06-28 18:59:19 -0400178 """FileInput([files[, inplace[, backup[, bufsize[, mode[, openhook]]]]]])
Tim Peters8ac14952002-05-23 15:15:30 +0000179
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000180 Class FileInput is the implementation of the module; its methods
Georg Brandl67e9fb92006-02-19 13:56:17 +0000181 filename(), lineno(), fileline(), isfirstline(), isstdin(), fileno(),
182 nextfile() and close() correspond to the functions of the same name
183 in the module.
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000184 In addition it has a readline() method which returns the next
185 input line, and a __getitem__() method which implements the
186 sequence behavior. The sequence must be accessed in strictly
Tim Peters8ac14952002-05-23 15:15:30 +0000187 sequential order; random access and readline() cannot be mixed.
Raymond Hettingerd1fa3db2002-05-15 02:56:03 +0000188 """
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000189
Tim Peters200a5802006-02-19 21:26:07 +0000190 def __init__(self, files=None, inplace=0, backup="", bufsize=0,
Georg Brandlc98eeed2006-02-19 14:57:47 +0000191 mode="r", openhook=None):
Georg Brandle4662172006-02-19 09:51:27 +0000192 if isinstance(files, basestring):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000193 files = (files,)
194 else:
Guido van Rossum2516b392000-04-10 17:16:12 +0000195 if files is None:
196 files = sys.argv[1:]
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000197 if not files:
Guido van Rossum2516b392000-04-10 17:16:12 +0000198 files = ('-',)
199 else:
200 files = tuple(files)
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000201 self._files = files
202 self._inplace = inplace
203 self._backup = backup
204 self._savestdout = None
205 self._output = None
206 self._filename = None
Serhiy Storchaka69b7f812016-03-08 18:35:45 +0200207 self._startlineno = 0
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000208 self._filelineno = 0
209 self._file = None
Serhiy Storchaka69b7f812016-03-08 18:35:45 +0200210 self._readline = self._start_readline
Guido van Rossum8ca162f2002-04-07 06:36:23 +0000211 self._isstdin = False
Guido van Rossum0aec9fb1998-07-20 15:49:28 +0000212 self._backupfilename = None
Georg Brandlc029f872006-02-19 14:12:34 +0000213 # restrict mode argument to reading modes
214 if mode not in ('r', 'rU', 'U', 'rb'):
215 raise ValueError("FileInput opening mode must be one of "
216 "'r', 'rU', 'U' and 'rb'")
217 self._mode = mode
Georg Brandlc98eeed2006-02-19 14:57:47 +0000218 if inplace and openhook:
219 raise ValueError("FileInput cannot use an opening hook in inplace mode")
Brett Cannon36bed8a2008-08-03 23:52:32 +0000220 elif openhook and not hasattr(openhook, '__call__'):
Georg Brandlc98eeed2006-02-19 14:57:47 +0000221 raise ValueError("FileInput openhook must be callable")
222 self._openhook = openhook
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000223
224 def __del__(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000225 self.close()
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000226
227 def close(self):
Serhiy Storchaka1aa2c0f2015-04-10 13:24:10 +0300228 try:
229 self.nextfile()
230 finally:
231 self._files = ()
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000232
Neil Schemenauer908632a2002-03-26 20:28:40 +0000233 def __iter__(self):
234 return self
235
236 def next(self):
Serhiy Storchaka69b7f812016-03-08 18:35:45 +0200237 line = self._readline()
238 if line:
Guido van Rossum47955242001-01-05 14:44:39 +0000239 self._filelineno += 1
240 return line
Serhiy Storchaka69b7f812016-03-08 18:35:45 +0200241 if not self._file:
Neil Schemenauer908632a2002-03-26 20:28:40 +0000242 raise StopIteration
Serhiy Storchaka69b7f812016-03-08 18:35:45 +0200243 self.nextfile()
244 # Recursive call
245 return self.next()
Tim Peters863ac442002-04-16 01:38:40 +0000246
Neil Schemenauer908632a2002-03-26 20:28:40 +0000247 def __getitem__(self, i):
Serhiy Storchaka69b7f812016-03-08 18:35:45 +0200248 if i != self.lineno():
Neil Schemenauer908632a2002-03-26 20:28:40 +0000249 raise RuntimeError, "accessing lines out of order"
250 try:
251 return self.next()
252 except StopIteration:
253 raise IndexError, "end of input reached"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000254
255 def nextfile(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000256 savestdout = self._savestdout
257 self._savestdout = 0
258 if savestdout:
259 sys.stdout = savestdout
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000260
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000261 output = self._output
262 self._output = 0
Serhiy Storchaka1aa2c0f2015-04-10 13:24:10 +0300263 try:
264 if output:
265 output.close()
266 finally:
267 file = self._file
Serhiy Storchaka69b7f812016-03-08 18:35:45 +0200268 self._file = None
269 self._readline = self._start_readline
Serhiy Storchaka1aa2c0f2015-04-10 13:24:10 +0300270 try:
271 if file and not self._isstdin:
272 file.close()
273 finally:
274 backupfilename = self._backupfilename
275 self._backupfilename = 0
276 if backupfilename and not self._backup:
277 try: os.unlink(backupfilename)
278 except OSError: pass
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000279
Serhiy Storchaka1aa2c0f2015-04-10 13:24:10 +0300280 self._isstdin = False
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000281
282 def readline(self):
Serhiy Storchaka69b7f812016-03-08 18:35:45 +0200283 while 1:
284 line = self._readline()
285 if line:
286 self._filelineno += 1
287 return line
288 if not self._file:
289 return line
Guido van Rossum47955242001-01-05 14:44:39 +0000290 self.nextfile()
Serhiy Storchaka69b7f812016-03-08 18:35:45 +0200291 # repeat with next file
292
293 def _start_readline(self):
294 if not self._files:
295 return ""
296 self._filename = self._files[0]
297 self._files = self._files[1:]
298 self._startlineno = self.lineno()
299 self._filelineno = 0
300 self._file = None
301 self._isstdin = False
302 self._backupfilename = 0
303 if self._filename == '-':
304 self._filename = '<stdin>'
305 self._file = sys.stdin
306 self._isstdin = True
307 else:
308 if self._inplace:
309 self._backupfilename = (
310 self._filename + (self._backup or os.extsep+"bak"))
311 try: os.unlink(self._backupfilename)
312 except os.error: pass
313 # The next few lines may raise IOError
314 os.rename(self._filename, self._backupfilename)
315 self._file = open(self._backupfilename, self._mode)
316 try:
317 perm = os.fstat(self._file.fileno()).st_mode
318 except OSError:
319 self._output = open(self._filename, "w")
320 else:
321 fd = os.open(self._filename,
322 os.O_CREAT | os.O_WRONLY | os.O_TRUNC,
323 perm)
324 self._output = os.fdopen(fd, "w")
325 try:
326 if hasattr(os, 'chmod'):
327 os.chmod(self._filename, perm)
328 except OSError:
329 pass
330 self._savestdout = sys.stdout
331 sys.stdout = self._output
332 else:
333 # This may raise IOError
334 if self._openhook:
335 self._file = self._openhook(self._filename, self._mode)
336 else:
337 self._file = open(self._filename, self._mode)
338
339 self._readline = self._file.readline
340 return self._readline()
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000341
342 def filename(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000343 return self._filename
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000344
345 def lineno(self):
Serhiy Storchaka69b7f812016-03-08 18:35:45 +0200346 return self._startlineno + self._filelineno
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000347
348 def filelineno(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000349 return self._filelineno
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000350
Georg Brandl67e9fb92006-02-19 13:56:17 +0000351 def fileno(self):
352 if self._file:
353 try:
354 return self._file.fileno()
355 except ValueError:
356 return -1
357 else:
358 return -1
359
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000360 def isfirstline(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000361 return self._filelineno == 1
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000362
363 def isstdin(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000364 return self._isstdin
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000365
Georg Brandlc98eeed2006-02-19 14:57:47 +0000366
367def hook_compressed(filename, mode):
368 ext = os.path.splitext(filename)[1]
369 if ext == '.gz':
370 import gzip
371 return gzip.open(filename, mode)
372 elif ext == '.bz2':
373 import bz2
374 return bz2.BZ2File(filename, mode)
375 else:
376 return open(filename, mode)
377
378
379def hook_encoded(encoding):
Serhiy Storchaka68b8a942014-02-26 20:59:08 +0200380 import io
Georg Brandlc98eeed2006-02-19 14:57:47 +0000381 def openhook(filename, mode):
Serhiy Storchaka68b8a942014-02-26 20:59:08 +0200382 mode = mode.replace('U', '').replace('b', '') or 'r'
383 return io.open(filename, mode, encoding=encoding, newline='')
Georg Brandlc98eeed2006-02-19 14:57:47 +0000384 return openhook
385
386
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000387def _test():
388 import getopt
389 inplace = 0
390 backup = 0
391 opts, args = getopt.getopt(sys.argv[1:], "ib:")
392 for o, a in opts:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000393 if o == '-i': inplace = 1
394 if o == '-b': backup = a
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000395 for line in input(args, inplace=inplace, backup=backup):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000396 if line[-1:] == '\n': line = line[:-1]
397 if line[-1:] == '\r': line = line[:-1]
398 print "%d: %s[%d]%s %s" % (lineno(), filename(), filelineno(),
399 isfirstline() and "*" or "", line)
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000400 print "%d: %s[%d]" % (lineno(), filename(), filelineno())
401
402if __name__ == '__main__':
403 _test()