blob: 2f895e9dae892e351b1a300e19b46c81d195438e [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
31All files are opened in text mode. If an I/O error occurs during
32opening or reading a file, the IOError exception is raised.
33
34If sys.stdin is used more than once, the second and further use will
35return no lines, except perhaps for interactive use, or if it has been
36explicitly reset (e.g. using sys.stdin.seek(0)).
37
38Empty files are opened and immediately closed; the only time their
39presence in the list of filenames is noticeable at all is when the
40last file opened is empty.
41
42It is possible that the last line of a file doesn't end in a newline
43character; otherwise lines are returned including the trailing
44newline.
45
46Class FileInput is the implementation; its methods filename(),
47lineno(), fileline(), isfirstline(), isstdin(), nextfile() and close()
48correspond to the functions in the module. In addition it has a
49readline() method which returns the next input line, and a
50__getitem__() method which implements the sequence behavior. The
51sequence must be accessed in strictly sequential order; sequence
52access and readline() cannot be mixed.
53
54Optional in-place filtering: if the keyword argument inplace=1 is
55passed to input() or to the FileInput constructor, the file is moved
56to a backup file and standard output is directed to the input file.
57This makes it possible to write a filter that rewrites its input file
58in place. If the keyword argument backup=".<some extension>" is also
59given, it specifies the extension for the backup file, and the backup
60file remains around; by default, the extension is ".bak" and it is
61deleted when the output file is closed. In-place filtering is
62disabled when standard input is read. XXX The current implementation
63does not work for MS-DOS 8+3 filesystems.
64
65XXX Possible additions:
66
67- optional getopt argument processing
68- specify open mode ('r' or 'rb')
69- specify buffer size
70- fileno()
71- isatty()
72- read(), read(size), even readlines()
73
74"""
75
76import sys, os
77
78_state = None
79
80def input(files=(), inplace=0, backup=""):
81 global _state
82 if _state and _state._file:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +000083 raise RuntimeError, "input() already active"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +000084 _state = FileInput(files, inplace, backup)
85 return _state
86
87def close():
88 global _state
89 state = _state
90 _state = None
91 if state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +000092 state.close()
Guido van Rossum7d5b99d1997-11-21 17:12:59 +000093
94def nextfile():
95 if not _state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +000096 raise RuntimeError, "no active input()"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +000097 return _state.nextfile()
98
99def filename():
100 if not _state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000101 raise RuntimeError, "no active input()"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000102 return _state.filename()
103
104def lineno():
105 if not _state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000106 raise RuntimeError, "no active input()"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000107 return _state.lineno()
108
109def filelineno():
110 if not _state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000111 raise RuntimeError, "no active input()"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000112 return _state.filelineno()
113
114def isfirstline():
115 if not _state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000116 raise RuntimeError, "no active input()"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000117 return _state.isfirstline()
118
119def isstdin():
120 if not _state:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000121 raise RuntimeError, "no active input()"
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000122 return _state.isstdin()
123
124class FileInput:
125
126 def __init__(self, files=(), inplace=0, backup=""):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000127 if type(files) == type(''):
128 files = (files,)
129 else:
130 files = tuple(files)
131 if not files:
132 files = tuple(sys.argv[1:])
133 if not files:
134 files = ('-',)
135 self._files = files
136 self._inplace = inplace
137 self._backup = backup
138 self._savestdout = None
139 self._output = None
140 self._filename = None
141 self._lineno = 0
142 self._filelineno = 0
143 self._file = None
144 self._isstdin = 0
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000145
146 def __del__(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000147 self.close()
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000148
149 def close(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000150 self.nextfile()
151 self._files = ()
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000152
153 def __getitem__(self, i):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000154 if i != self._lineno:
155 raise RuntimeError, "accessing lines out of order"
156 line = self.readline()
157 if not line:
158 raise IndexError, "end of input reached"
159 return line
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000160
161 def nextfile(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000162 savestdout = self._savestdout
163 self._savestdout = 0
164 if savestdout:
165 sys.stdout = savestdout
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000166
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000167 output = self._output
168 self._output = 0
169 if output:
170 output.close()
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000171
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000172 file = self._file
173 self._file = 0
174 if file and not self._isstdin:
175 file.close()
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000176
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000177 backupfilename = self._backupfilename
178 self._backupfilename = 0
179 if backupfilename and not self._backup:
180 try: os.unlink(backupfilename)
181 except: pass
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000182
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000183 self._isstdin = 0
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000184
185 def readline(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000186 if not self._file:
187 if not self._files:
188 return ""
189 self._filename = self._files[0]
190 self._files = self._files[1:]
191 self._filelineno = 0
192 self._file = None
193 self._isstdin = 0
194 self._backupfilename = 0
195 if self._filename == '-':
196 self._filename = '<stdin>'
197 self._file = sys.stdin
198 self._isstdin = 1
199 else:
200 if self._inplace:
201 self._backupfilename = (
202 self._filename + (self._backup or ".bak"))
203 try: os.unlink(self._backupfilename)
204 except os.error: pass
205 # The next three lines may raise IOError
206 os.rename(self._filename, self._backupfilename)
207 self._file = open(self._backupfilename, "r")
208 self._output = open(self._filename, "w")
209 self._savestdout = sys.stdout
210 sys.stdout = self._output
211 else:
212 # This may raise IOError
213 self._file = open(self._filename, "r")
214 line = self._file.readline()
215 if line:
216 self._lineno = self._lineno + 1
217 self._filelineno = self._filelineno + 1
218 return line
219 self.nextfile()
220 # Recursive call
221 return self.readline()
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000222
223 def filename(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000224 return self._filename
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000225
226 def lineno(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000227 return self._lineno
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000228
229 def filelineno(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000230 return self._filelineno
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000231
232 def isfirstline(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000233 return self._filelineno == 1
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000234
235 def isstdin(self):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000236 return self._isstdin
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000237
238def _test():
239 import getopt
240 inplace = 0
241 backup = 0
242 opts, args = getopt.getopt(sys.argv[1:], "ib:")
243 for o, a in opts:
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000244 if o == '-i': inplace = 1
245 if o == '-b': backup = a
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000246 for line in input(args, inplace=inplace, backup=backup):
Guido van Rossum45e2fbc1998-03-26 21:13:24 +0000247 if line[-1:] == '\n': line = line[:-1]
248 if line[-1:] == '\r': line = line[:-1]
249 print "%d: %s[%d]%s %s" % (lineno(), filename(), filelineno(),
250 isfirstline() and "*" or "", line)
Guido van Rossum7d5b99d1997-11-21 17:12:59 +0000251 print "%d: %s[%d]" % (lineno(), filename(), filelineno())
252
253if __name__ == '__main__':
254 _test()