This file is indexed.

/usr/lib/python2.7/dist-packages/pysrt/srtfile.py is in python-pysrt 1.0.1-1.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

  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
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
# -*- coding: utf-8 -*-
import os
import sys
import codecs

try:
    from collections import UserList
except ImportError:
    from UserList import UserList

from itertools import chain
from copy import copy

from pysrt.srtexc import Error
from pysrt.srtitem import SubRipItem
from pysrt.compat import str

BOMS = ((codecs.BOM_UTF32_LE, 'utf_32_le'),
        (codecs.BOM_UTF32_BE, 'utf_32_be'),
        (codecs.BOM_UTF16_LE, 'utf_16_le'),
        (codecs.BOM_UTF16_BE, 'utf_16_be'),
        (codecs.BOM_UTF8, 'utf_8'))
CODECS_BOMS = dict((codec, str(bom, codec)) for bom, codec in BOMS)
BIGGER_BOM = max(len(bom) for bom, encoding in BOMS)


class SubRipFile(UserList, object):
    """
    SubRip file descriptor.

    Provide a pure Python mapping on all metadata.

    SubRipFile(items, eol, path, encoding)

    items -> list of SubRipItem. Default to [].
    eol -> str: end of line character. Default to linesep used in opened file
        if any else to os.linesep.
    path -> str: path where file will be saved. To open an existant file see
        SubRipFile.open.
    encoding -> str: encoding used at file save. Default to utf-8.
    """
    ERROR_PASS = 0
    ERROR_LOG = 1
    ERROR_RAISE = 2

    DEFAULT_ENCODING = 'utf_8'

    def __init__(self, items=None, eol=None, path=None, encoding='utf-8'):
        UserList.__init__(self, items or [])
        self._eol = eol
        self.path = path
        self.encoding = encoding

    def _get_eol(self):
        return self._eol or os.linesep

    def _set_eol(self, eol):
        self._eol = self._eol or eol

    eol = property(_get_eol, _set_eol)

    def slice(self, starts_before=None, starts_after=None, ends_before=None,
              ends_after=None):
        """
        slice([starts_before][, starts_after][, ends_before][, ends_after]) \
-> SubRipFile clone

        All arguments are optional, and should be coercible to SubRipTime
        object.

        It reduce the set of subtitles to those that match match given time
        constraints.

        The returned set is a clone, but still contains references to original
        subtitles. So if you shift this returned set, subs contained in the
        original SubRipFile instance will be altered too.

        Example:
            >>> subs.slice(ends_after={'seconds': 20}).shift(seconds=2)
        """
        clone = copy(self)

        if starts_before:
            clone.data = (i for i in clone.data if i.start < starts_before)
        if starts_after:
            clone.data = (i for i in clone.data if i.start > starts_after)
        if ends_before:
            clone.data = (i for i in clone.data if i.end < ends_before)
        if ends_after:
            clone.data = (i for i in clone.data if i.end > ends_after)

        clone.data = list(clone.data)
        return clone

    def at(self, timestamp=None, **kwargs):
        """
        at(timestamp) -> SubRipFile clone

        timestamp argument should be coercible to SubRipFile object.

        A specialization of slice. Return all subtiles visible at the
        timestamp mark.

        Example:
            >>> subs.at((0, 0, 20, 0)).shift(seconds=2)
            >>> subs.at(seconds=20).shift(seconds=2)
        """
        time = timestamp or kwargs
        return self.slice(starts_before=time, ends_after=time)

    def shift(self, *args, **kwargs):
        """shift(hours, minutes, seconds, milliseconds, ratio)

        Shift `start` and `end` attributes of each items of file either by
        applying a ratio or by adding an offset.

        `ratio` should be either an int or a float.
        Example to convert subtitles from 23.9 fps to 25 fps:
        >>> subs.shift(ratio=25/23.9)

        All "time" arguments are optional and have a default value of 0.
        Example to delay all subs from 2 seconds and half
        >>> subs.shift(seconds=2, milliseconds=500)
        """
        for item in self:
            item.shift(*args, **kwargs)

    def clean_indexes(self):
        """
        clean_indexes()

        Sort subs and reset their index attribute. Should be called after
        destructive operations like split or such.
        """
        self.sort()
        for index, item in enumerate(self):
            item.index = index + 1

    @property
    def text(self):
        return '\n'.join(i.text for i in self)

    @classmethod
    def open(cls, path='', encoding=None, error_handling=ERROR_PASS):
        """
        open([path, [encoding]])

        If you do not provide any encoding, it can be detected if the file
        contain a bit order mark, unless it is set to utf-8 as default.
        """
        new_file = cls(path=path, encoding=encoding)
        source_file = cls._open_unicode_file(path, claimed_encoding=encoding)
        new_file.read(source_file, error_handling=error_handling)
        source_file.close()
        return new_file

    @classmethod
    def from_string(cls, source, **kwargs):
        """
        from_string(source, **kwargs) -> SubRipFile

        `source` -> a unicode instance or at least a str instance encoded with
        `sys.getdefaultencoding()`
        """
        error_handling = kwargs.pop('error_handling', None)
        new_file = cls(**kwargs)
        new_file.read(source.splitlines(True), error_handling=error_handling)
        return new_file

    def read(self, source_file, error_handling=ERROR_PASS):
        """
        read(source_file, [error_handling])

        This method parse subtitles contained in `source_file` and append them
        to the current instance.

        `source_file` -> Any iterable that yield unicode strings, like a file
            opened with `codecs.open()` or an array of unicode.
        """
        self.eol = self._guess_eol(source_file)
        self.extend(self.stream(source_file, error_handling=error_handling))
        return self

    @classmethod
    def stream(cls, source_file, error_handling=ERROR_PASS):
        """
        stream(source_file, [error_handling])

        This method yield SubRipItem instances a soon as they have been parsed
        without storing them. It is a kind of SAX parser for .srt files.

        `source_file` -> Any iterable that yield unicode strings, like a file
            opened with `codecs.open()` or an array of unicode.

        Example:
            >>> import pysrt
            >>> import codecs
            >>> file = codecs.open('movie.srt', encoding='utf-8')
            >>> for sub in pysrt.stream(file):
            ...     sub.text += "\nHello !"
            ...     print unicode(sub)
        """
        string_buffer = []
        for index, line in enumerate(chain(source_file, '\n')):
            if line.strip():
                string_buffer.append(line)
            else:
                source = string_buffer
                string_buffer = []
                if source and all(source):
                    try:
                        yield SubRipItem.from_lines(source)
                    except Error as error:
                        error.args += (''.join(source), )
                        cls._handle_error(error, error_handling, index)

    def save(self, path=None, encoding=None, eol=None):
        """
        save([path][, encoding][, eol])

        Use initial path if no other provided.
        Use initial encoding if no other provided.
        Use initial eol if no other provided.
        """
        path = path or self.path
        encoding = encoding or self.encoding

        save_file = codecs.open(path, 'w+', encoding=encoding)
        self.write_into(save_file, eol=eol)
        save_file.close()

    def write_into(self, output_file, eol=None):
        """
        write_into(output_file [, eol])

        Serialize current state into `output_file`.

        `output_file` -> Any instance that respond to `write()`, typically a
        file object
        """
        output_eol = eol or self.eol

        for item in self:
            string_repr = str(item)
            if output_eol != '\n':
                string_repr = string_repr.replace('\n', output_eol)
            output_file.write(string_repr)
            # Only add trailing eol if it's not already present.
            # It was kept in the SubRipItem's text before but it really
            # belongs here. Existing applications might give us subtitles
            # which already contain a trailing eol though.
            if not string_repr.endswith(2 * output_eol):
                output_file.write(output_eol)

    @classmethod
    def _guess_eol(cls, string_iterable):
        first_line = cls._get_first_line(string_iterable)
        for eol in ('\r\n', '\r', '\n'):
            if first_line.endswith(eol):
                return eol
        return os.linesep

    @classmethod
    def _get_first_line(cls, string_iterable):
        if hasattr(string_iterable, 'tell'):
            previous_position = string_iterable.tell()

        try:
            first_line = next(iter(string_iterable))
        except StopIteration:
            return ''
        if hasattr(string_iterable, 'seek'):
            string_iterable.seek(previous_position)

        return first_line

    @classmethod
    def _detect_encoding(cls, path):
        file_descriptor = open(path, 'rb')
        first_chars = file_descriptor.read(BIGGER_BOM)
        file_descriptor.close()

        for bom, encoding in BOMS:
            if first_chars.startswith(bom):
                return encoding

        # TODO: maybe a chardet integration
        return cls.DEFAULT_ENCODING

    @classmethod
    def _open_unicode_file(cls, path, claimed_encoding=None):
        encoding = claimed_encoding or cls._detect_encoding(path)
        source_file = codecs.open(path, 'rU', encoding=encoding)

        # get rid of BOM if any
        possible_bom = CODECS_BOMS.get(encoding, None)
        if possible_bom:
            file_bom = source_file.read(len(possible_bom))
            if not file_bom == possible_bom:
                source_file.seek(0)  # if not rewind
        return source_file

    @classmethod
    def _handle_error(cls, error, error_handling, index):
        if error_handling == cls.ERROR_RAISE:
            error.args = (index, ) + error.args
            raise error
        if error_handling == cls.ERROR_LOG:
            name = type(error).__name__
            sys.stderr.write('PySRT-%s(line %s): \n' % (name, index))
            sys.stderr.write(error.args[0].encode('ascii', 'replace'))
            sys.stderr.write('\n')