root/galaxy-central/eggs/bx_python-0.5.0_dev_f74aec067563-py2.6-macosx-10.6-universal-ucs2.egg/bx/misc/filecache.py

リビジョン 3, 3.6 KB (コミッタ: kohda, 14 年 前)

Install Unix tools  http://hannonlab.cshl.edu/galaxy_unix_tools/galaxy.html

行番号 
1from __future__ import division
2
3import sys
4from bx_extras.lrucache import LRUCache
5from cStringIO import StringIO
6
7DEFAULT_CACHE_SIZE=10
8DEFAULT_BLOCK_SIZE=1024*1024*2
9
10class FileCache( object ):
11    """
12    Wrapper for a file that cache blocks of data in memory.
13   
14    **NOTE:** this is currently an incomplete file-like object, it only
15    supports seek, tell, and readline (plus iteration). Reading bytes is
16    currently not implemented.
17    """
18    def __init__( self, file, size, cache_size=DEFAULT_CACHE_SIZE,
19                                    block_size=DEFAULT_BLOCK_SIZE ):
20        """
21        Create a new `FileCache` wrapping the file-like object `file` that
22        has total size `size` and caching blocks of size `block_size`.
23        """
24        self.file = file
25        self.size = size
26        self.cache_size = cache_size
27        self.block_size = block_size
28        # Setup the cache
29        self.nblocks = ( self.size // self.block_size ) + 1
30        self.cache = LRUCache( self.cache_size )
31        # Position in file
32        self.dirty = True
33        self.at_eof = False
34        self.file_pos = 0
35        self.current_block_index = -1
36        self.current_block = None
37    def fix_dirty( self ):
38        chunk, offset = self.get_block_and_offset( self.file_pos )
39        if self.current_block_index != chunk:
40            self.current_block = StringIO( self.load_block( chunk ) )
41            self.current_block.read( offset )
42            self.current_block_index = chunk
43        else:
44            self.current_block.seek( offset )
45        self.dirty = False
46    def get_block_and_offset( self, index ):
47        return int( index // self.block_size ), int( index % self.block_size )
48    def load_block( self, index ):
49        if index in self.cache:
50            return self.cache[index]
51        else:
52            real_offset = index * self.block_size
53            self.file.seek( real_offset )
54            block = self.file.read( self.block_size )
55            self.cache[index] = block
56            return block
57    def seek( self, offset, whence=0 ):
58        """
59        Move the file pointer to a particular offset.
60        """
61        # Determine absolute target position
62        if whence == 0:
63            target_pos = offset
64        elif whence == 1:
65            target_pos = self.file_pos + offset
66        elif whence == 2:
67            target_pos = self.size - offset
68        else:
69            raise Exception( "Invalid `whence` argument: %r", whence )
70        # Check if this is a noop
71        if target_pos == self.file_pos:
72            return   
73        # Verify it is valid
74        assert 0 <= target_pos < self.size, "Attempt to seek outside file"
75        # Move the position
76        self.file_pos = target_pos
77        # Mark as dirty, the next time a read is done we need to actually
78        # move the position in the bzip2 file
79        self.dirty = True
80    def readline( self ):
81        if self.dirty:
82            self.fix_dirty()
83        if self.at_eof:
84            return ""
85        rval = []
86        while 1:
87            line = self.current_block.readline()
88            rval.append( line )
89            if len( line ) > 0 and line[-1] == '\n':
90                break
91            elif self.current_block_index == self.nblocks - 1:
92                self.at_eof = True
93                break
94            else:
95                self.current_block_index += 1
96                self.current_block = StringIO( self.load_block( self.current_block_index ) )     
97        return "".join( rval )     
98    def next( self ):
99        line = self.readline()
100        if line == "":
101            raise StopIteration
102    def __iter__( self ):
103        return self
104    def close( self ):
105        self.file.close()
Note: リポジトリブラウザについてのヘルプは TracBrowser を参照してください。