linesource 0.0.1 → 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- data/Rakefile +1 -0
- data/VERSION +1 -1
- data/lib/linesource.rb +28 -5
- data/test/test_linesource.rb +55 -2
- metadata +1 -1
data/Rakefile
CHANGED
@@ -10,6 +10,7 @@ begin
|
|
10
10
|
gem.email = "daniel@thepolfers.com"
|
11
11
|
gem.homepage = "http://github.com/polfer/linesource"
|
12
12
|
gem.authors = ["Daniel Polfer"]
|
13
|
+
gem.has_rdoc = true
|
13
14
|
# gem is a Gem::Specification... see http://www.rubygems.org/read/chapter/20 for additional settings
|
14
15
|
end
|
15
16
|
Jeweler::GemcutterTasks.new
|
data/VERSION
CHANGED
@@ -1 +1 @@
|
|
1
|
-
0.0
|
1
|
+
0.1.0
|
data/lib/linesource.rb
CHANGED
@@ -1,17 +1,27 @@
|
|
1
1
|
require 'stringio'
|
2
2
|
require 'zlib'
|
3
3
|
|
4
|
-
#
|
5
|
-
#
|
6
|
-
#
|
7
|
-
#
|
8
|
-
|
4
|
+
# The LineSource class is initialized with a pattern for files that
|
5
|
+
# need to be read and processed line by line. For example, you could
|
6
|
+
# pass a pattern to all the log files in a date hierarchy ('**/*.log').
|
7
|
+
# Every line in every file will be read and the transition from file
|
8
|
+
# to file is handled automatically (with callbacks if desired).
|
9
9
|
class LineSource
|
10
10
|
|
11
11
|
include Enumerable
|
12
12
|
|
13
|
+
# Linenum represents the line number of the current file (1 based).
|
13
14
|
attr_reader :linenum
|
14
15
|
|
16
|
+
# Constructor can take a number of different options including
|
17
|
+
# the pattern of files to be read (Dir.glob style). The files can
|
18
|
+
# be regular text files or gzipped files (detected through a .gz)
|
19
|
+
# extension. The skiplines parameter can be used to specify a certain
|
20
|
+
# number of lines at the head of each file that should be skipped
|
21
|
+
# (usually header files, for example on a csv file). The filenew
|
22
|
+
# and filedone params represent a callback that can be fired at the
|
23
|
+
# start or end of processing for each of the files that match pattern.
|
24
|
+
# TODO: We should be able to pass multiple pattenrs to initialize.
|
15
25
|
def initialize( pattern, skiplines = 0, filenew = nil, filedone = nil )
|
16
26
|
@skiplines = skiplines
|
17
27
|
@filenames = Dir.glob(pattern).sort!.freeze
|
@@ -24,18 +34,24 @@ class LineSource
|
|
24
34
|
nextfile
|
25
35
|
end
|
26
36
|
|
37
|
+
# Allows the prefile callback to be set after LineSource creation.
|
27
38
|
def setfilenew( filenew )
|
28
39
|
@filenew = filenew
|
29
40
|
end
|
30
41
|
|
42
|
+
# Allows the postfile callback to be set after LineSource creation.
|
31
43
|
def setfiledone( filedone )
|
32
44
|
@filedone = filedone
|
33
45
|
end
|
34
46
|
|
47
|
+
# Tests if the end of _all_ files in the original pattern are processed.
|
35
48
|
def eof?
|
36
49
|
@content ? @content.eof? : true
|
37
50
|
end
|
38
51
|
|
52
|
+
# Closes all files held open on the current LineSource.
|
53
|
+
# Effectively ends the use of this LineSource.
|
54
|
+
# TODO: We could make it so that a close is like a "reset".
|
39
55
|
def close
|
40
56
|
if @content
|
41
57
|
@filedone.call(self) if @filedone
|
@@ -47,6 +63,7 @@ class LineSource
|
|
47
63
|
@linenum = -1
|
48
64
|
end
|
49
65
|
|
66
|
+
# Skips the current file being processed and sets to the next.
|
50
67
|
def nextfile
|
51
68
|
begin
|
52
69
|
if @content
|
@@ -78,6 +95,7 @@ class LineSource
|
|
78
95
|
end
|
79
96
|
end
|
80
97
|
|
98
|
+
# Returns an individual line from the LineSource.
|
81
99
|
def gets
|
82
100
|
@lastline = nil
|
83
101
|
if @content
|
@@ -91,18 +109,23 @@ class LineSource
|
|
91
109
|
@lastline
|
92
110
|
end
|
93
111
|
|
112
|
+
# Re-returns the last line fetched from a LineSource.
|
94
113
|
def lastline
|
95
114
|
@lastline
|
96
115
|
end
|
97
116
|
|
117
|
+
# Returns the name of the file currently being processed.
|
98
118
|
def filename
|
99
119
|
@index >= 0 ? @filenames[@index] : nil
|
100
120
|
end
|
101
121
|
|
122
|
+
# Returns the size of the file currently being processed. Note
|
123
|
+
# that in the case of a GZipped file this will be the compressed size.
|
102
124
|
def filesize
|
103
125
|
(@index >= 0 && @filenames[@index]) ? File.stat(@filenames[@index]).size : -1
|
104
126
|
end
|
105
127
|
|
128
|
+
# Calls a block of code for each line from the LineSource.
|
106
129
|
def each
|
107
130
|
while gets
|
108
131
|
yield @lastline
|
data/test/test_linesource.rb
CHANGED
@@ -1,7 +1,60 @@
|
|
1
1
|
require 'helper'
|
2
2
|
|
3
3
|
class TestLinesource < Test::Unit::TestCase
|
4
|
-
def
|
5
|
-
|
4
|
+
def setup
|
5
|
+
@testlines = IO.readlines(__FILE__)
|
6
6
|
end
|
7
|
+
|
8
|
+
def test_single_file
|
9
|
+
linenum = 0
|
10
|
+
ls = LineSource.new(__FILE__)
|
11
|
+
ls.each do |line|
|
12
|
+
linenum += 1
|
13
|
+
assert_equal(linenum, ls.linenum)
|
14
|
+
assert_equal(@testlines.shift,line)
|
15
|
+
end
|
16
|
+
end
|
17
|
+
|
18
|
+
def test_line_skip
|
19
|
+
linenum = 4
|
20
|
+
4.times { @testlines.shift }
|
21
|
+
ls = LineSource.new( __FILE__, 4 )
|
22
|
+
ls.each do |line|
|
23
|
+
linenum += 1
|
24
|
+
assert_equal(linenum, ls.linenum)
|
25
|
+
assert_equal(line, @testlines.shift)
|
26
|
+
end
|
27
|
+
end
|
28
|
+
|
29
|
+
def test_single_prefile_callback
|
30
|
+
prefilehit = false
|
31
|
+
progresscallback = lambda do |ls|
|
32
|
+
assert_equal(__FILE__, ls.filename)
|
33
|
+
prefilehit = true
|
34
|
+
end
|
35
|
+
ls = LineSource.new( __FILE__, 0, progresscallback )
|
36
|
+
assert_equal( prefilehit, true )
|
37
|
+
end
|
38
|
+
|
39
|
+
def test_single_postfile_no_callback
|
40
|
+
postfilehit = false
|
41
|
+
progresscallback = lambda do |ls|
|
42
|
+
assert_equal(__FILE__, ls.filename)
|
43
|
+
postfilehit = true
|
44
|
+
end
|
45
|
+
ls = LineSource.new( __FILE__, 0, nil, progresscallback )
|
46
|
+
assert_equal( postfilehit, false )
|
47
|
+
end
|
48
|
+
|
49
|
+
def test_single_postfile_callback
|
50
|
+
postfilehit = false
|
51
|
+
progresscallback = lambda do |ls|
|
52
|
+
assert_equal(__FILE__, ls.filename)
|
53
|
+
postfilehit = true
|
54
|
+
end
|
55
|
+
ls = LineSource.new( __FILE__, 0, nil, progresscallback )
|
56
|
+
ls.each { |line| assert_equal(line, @testlines.shift) }
|
57
|
+
assert_equal( postfilehit, true )
|
58
|
+
end
|
59
|
+
|
7
60
|
end
|