apachecrunch 0.4 → 0.5
Sign up to get free protection for your applications and to get access to all the features.
- data/bin/apachecrunch +1 -1
- data/lib/apachecrunch.rb +5 -15
- data/lib/cast.rb +21 -0
- data/lib/derivation.rb +113 -0
- data/lib/element.rb +16 -0
- data/lib/element_value_fetcher.rb +72 -0
- data/lib/entry.rb +64 -54
- data/lib/format.rb +21 -63
- data/lib/format_token.rb +114 -0
- data/lib/format_token_definition.rb +183 -0
- data/lib/log_parser.rb +39 -31
- data/lib/procedure_dsl.rb +254 -244
- data/lib/progress.rb +1 -1
- data/test/mock.rb +37 -0
- data/test/runner.rb +13 -1
- data/test/stub.rb +66 -36
- data/test/test_derived_value_fetcher.rb +36 -0
- data/test/test_element.rb +18 -0
- data/test/test_element_value_fetcher.rb +45 -0
- data/test/test_entry_parser.rb +39 -0
- data/test/test_format.rb +13 -51
- data/test/test_format_parser.rb +22 -13
- data/test/test_log_parser.rb +88 -0
- data/test/test_raw_value_fetcher.rb +36 -0
- data/test/test_regex_token.rb +17 -0
- data/test/test_req_firstline_derivation_rule.rb +41 -0
- data/test/test_reqheader_token.rb +26 -0
- data/test/test_string_token.rb +27 -0
- data/test/test_time_derivation_rule.rb +29 -0
- metadata +23 -18
- data/lib/log_element.rb +0 -351
- data/test/test_entry.rb +0 -28
data/lib/format_token.rb
ADDED
@@ -0,0 +1,114 @@
|
|
1
|
+
require 'format_token_definition'
|
2
|
+
require 'derivation'
|
3
|
+
|
4
|
+
class ApacheCrunch
|
5
|
+
# Abstract for a token in a log format
|
6
|
+
class FormatToken
|
7
|
+
# Performs whatever initial population is necessary for the token.
|
8
|
+
def populate!; raise NotImplementedError; end
|
9
|
+
|
10
|
+
def name; raise NotImplementedError; end
|
11
|
+
def regex; raise NotImplementedError; end
|
12
|
+
def captured?; raise NotImplementedError; end
|
13
|
+
def derivation_rule; raise NotImplementedError; end
|
14
|
+
end
|
15
|
+
|
16
|
+
|
17
|
+
# A predefined token like %q or %r from the Apache log.
|
18
|
+
class PredefinedToken < FormatToken
|
19
|
+
def populate!(token_definition)
|
20
|
+
@token_definition = token_definition
|
21
|
+
end
|
22
|
+
|
23
|
+
def name; @token_definition.name; end
|
24
|
+
def regex; @token_definition.regex; end
|
25
|
+
def captured?; @token_definition.captured; end
|
26
|
+
def derivation_rule; @token_definition.derivation_rule; end
|
27
|
+
end
|
28
|
+
|
29
|
+
|
30
|
+
# A bare string in a log format.
|
31
|
+
class StringToken < FormatToken
|
32
|
+
# Initializes the instance given the string it represents
|
33
|
+
def populate!(string_value)
|
34
|
+
@_string_value = string_value
|
35
|
+
end
|
36
|
+
|
37
|
+
def name; nil; end
|
38
|
+
|
39
|
+
def regex
|
40
|
+
# Make sure there aren't any regex special characters in the string that will confuse
|
41
|
+
# the parsing later.
|
42
|
+
Regexp.escape(@_string_value)
|
43
|
+
end
|
44
|
+
|
45
|
+
def captured?; false; end
|
46
|
+
def derivation_rule; NullDerivationRule.new; end
|
47
|
+
end
|
48
|
+
|
49
|
+
|
50
|
+
# A token based on a request header.
|
51
|
+
class ReqheaderToken < FormatToken
|
52
|
+
def populate!(header_name)
|
53
|
+
@_name = _header_name_to_token_name(header_name)
|
54
|
+
end
|
55
|
+
|
56
|
+
def name; @_name; end
|
57
|
+
def regex; '[^"]*'; end
|
58
|
+
def captured?; true; end
|
59
|
+
def derivation_rule; NullDerivationRule.new; end
|
60
|
+
|
61
|
+
# Lowercases header name and turns hyphens into underscores
|
62
|
+
def _header_name_to_token_name(header_name)
|
63
|
+
("reqheader_" + header_name.downcase().gsub("-", "_")).to_sym
|
64
|
+
end
|
65
|
+
end
|
66
|
+
|
67
|
+
|
68
|
+
# A token based on an arbitrary regular expression.
|
69
|
+
class RegexToken < FormatToken
|
70
|
+
def populate!(regex_name, regex_text)
|
71
|
+
@_name = "regex_#{regex_name}".to_sym
|
72
|
+
@_regex = regex_text
|
73
|
+
end
|
74
|
+
|
75
|
+
def name; @_name; end
|
76
|
+
def regex; @_regex; end
|
77
|
+
def captured?; true; end
|
78
|
+
def derivation_rule; NullDerivationRule.new; end
|
79
|
+
end
|
80
|
+
|
81
|
+
|
82
|
+
# Generates FormatToken instances.
|
83
|
+
#
|
84
|
+
# This class does the work of figuring out which FormatToken subclass to make.
|
85
|
+
class FormatTokenFactory
|
86
|
+
# Takes an Apache log format abbreviation and returns a corresponding FormatToken
|
87
|
+
def self.from_abbrev(abbrev)
|
88
|
+
token_def = TokenDictionary.fetch(abbrev)
|
89
|
+
if token_def
|
90
|
+
# We found it in the dictionary, so just return a Token based on it
|
91
|
+
tok = PredefinedToken.new
|
92
|
+
tok.populate!(token_def)
|
93
|
+
elsif abbrev !~ /^%/
|
94
|
+
tok = StringToken.new
|
95
|
+
tok.populate!(abbrev)
|
96
|
+
elsif abbrev == "%%"
|
97
|
+
tok = StringToken.new
|
98
|
+
tok.populate!("%")
|
99
|
+
elsif abbrev =~ /^%\{([A-Za-z0-9-]+)\}i/
|
100
|
+
# HTTP request header
|
101
|
+
tok = ReqheaderToken.new
|
102
|
+
tok.populate!($1)
|
103
|
+
elsif abbrev =~ /^%\{(.*?):([^}]+)\}r/
|
104
|
+
# Arbitrary regex
|
105
|
+
tok = RegexToken.new
|
106
|
+
tok.populate!($1, $2)
|
107
|
+
else
|
108
|
+
raise "Unable to parse format definition starting at '#{abbrev}'"
|
109
|
+
end
|
110
|
+
|
111
|
+
tok
|
112
|
+
end
|
113
|
+
end
|
114
|
+
end
|
@@ -0,0 +1,183 @@
|
|
1
|
+
require 'cast'
|
2
|
+
require 'derivation'
|
3
|
+
|
4
|
+
class ApacheCrunch
|
5
|
+
# Defines the properties of a known Apache log format token (like %q or %h)
|
6
|
+
class FormatTokenDefinition
|
7
|
+
class << self; attr_accessor :name, :abbrev, :regex, :caster, :derivation_rule, :captured; end
|
8
|
+
end
|
9
|
+
|
10
|
+
|
11
|
+
class RemoteHostTokenDefinition < FormatTokenDefinition
|
12
|
+
@name = :remote_host
|
13
|
+
@abbrev = "%h"
|
14
|
+
@regex = %q![A-Za-z0-9.-]+!
|
15
|
+
@caster = nil
|
16
|
+
@derivation_rule = NullDerivationRule.new
|
17
|
+
@captured = true
|
18
|
+
end
|
19
|
+
|
20
|
+
|
21
|
+
class LogNameTokenDefinition < FormatTokenDefinition
|
22
|
+
@name = :log_name
|
23
|
+
@abbrev = "%l"
|
24
|
+
@regex = %q!\S+!
|
25
|
+
@caster = nil
|
26
|
+
@derivation_rule = NullDerivationRule.new
|
27
|
+
@captured = true
|
28
|
+
end
|
29
|
+
|
30
|
+
|
31
|
+
class RemoteUserTokenDefinition < FormatTokenDefinition
|
32
|
+
@name = :remote_user
|
33
|
+
@abbrev = "%u"
|
34
|
+
@regex = %q![^:]+!
|
35
|
+
@caster = nil
|
36
|
+
@derivation_rule = NullDerivationRule.new
|
37
|
+
@captured = true
|
38
|
+
end
|
39
|
+
|
40
|
+
|
41
|
+
class TimeTokenDefinition < FormatTokenDefinition
|
42
|
+
@name = :time
|
43
|
+
@abbrev = "%t"
|
44
|
+
@regex = %q!\[\d\d/[A-Za-z]{3}/\d\d\d\d:\d\d:\d\d:\d\d [-+]\d\d\d\d\]!
|
45
|
+
@caster = nil
|
46
|
+
@derivation_rule = TimeDerivationRule.new
|
47
|
+
@captured = true
|
48
|
+
end
|
49
|
+
|
50
|
+
|
51
|
+
class ReqFirstlineTokenDefinition < FormatTokenDefinition
|
52
|
+
@name = :req_firstline
|
53
|
+
@abbrev = "%r"
|
54
|
+
@regex = %q![^"]+!
|
55
|
+
@caster = nil
|
56
|
+
@derivation_rule = ReqFirstlineDerivationRule.new
|
57
|
+
@captured = true
|
58
|
+
end
|
59
|
+
|
60
|
+
|
61
|
+
class StatusTokenDefinition < FormatTokenDefinition
|
62
|
+
@name = :status
|
63
|
+
@abbrev = "%s"
|
64
|
+
@regex = %q!\d+|-!
|
65
|
+
@caster = nil
|
66
|
+
@derivation_rule = NullDerivationRule.new
|
67
|
+
@captured = true
|
68
|
+
end
|
69
|
+
|
70
|
+
|
71
|
+
class BytesSentTokenDefinition < FormatTokenDefinition
|
72
|
+
@name = :bytes_sent
|
73
|
+
@abbrev = "%b"
|
74
|
+
@regex = %q!\d+!
|
75
|
+
@caster = IntegerCast.new
|
76
|
+
@derivation_rule = NullDerivationRule.new
|
77
|
+
@captured = true
|
78
|
+
end
|
79
|
+
|
80
|
+
|
81
|
+
class BytesSentTokenDefinition < FormatTokenDefinition
|
82
|
+
@name = :bytes_sent
|
83
|
+
@abbrev = "%b"
|
84
|
+
@regex = %q![\d-]+!
|
85
|
+
@caster = CLFIntegerCast.new
|
86
|
+
@derivation_rule = NullDerivationRule.new
|
87
|
+
@captured = true
|
88
|
+
end
|
89
|
+
|
90
|
+
|
91
|
+
class BytesSentWithHeadersTokenDefinition < FormatTokenDefinition
|
92
|
+
@name = :bytes_sent_with_headers
|
93
|
+
@abbrev = "%O"
|
94
|
+
@regex = %q!\d+!
|
95
|
+
@caster = IntegerCast.new
|
96
|
+
@derivation_rule = NullDerivationRule.new
|
97
|
+
@captured = true
|
98
|
+
end
|
99
|
+
|
100
|
+
|
101
|
+
class ServeTimeMicroTokenDefinition < FormatTokenDefinition
|
102
|
+
@name = :serve_time_micro
|
103
|
+
@abbrev = "%D"
|
104
|
+
@regex = %q!\d+!
|
105
|
+
@caster = IntegerCast.new
|
106
|
+
@derivation_rule = NullDerivationRule.new
|
107
|
+
@captured = true
|
108
|
+
end
|
109
|
+
|
110
|
+
|
111
|
+
class UrlPathTokenDefinition < FormatTokenDefinition
|
112
|
+
@name = :url_path
|
113
|
+
@abbrev = "%U"
|
114
|
+
@regex = %q!/[^?]*!
|
115
|
+
@caster = nil
|
116
|
+
@derivation_rule = NullDerivationRule.new
|
117
|
+
@captured = true
|
118
|
+
end
|
119
|
+
|
120
|
+
|
121
|
+
class QueryStringTokenDefinition < FormatTokenDefinition
|
122
|
+
@name = :query_string
|
123
|
+
@abbrev = "%q"
|
124
|
+
@regex = %q!\??\S*!
|
125
|
+
@caster = nil
|
126
|
+
@derivation_rule = NullDerivationRule.new
|
127
|
+
@captured = true
|
128
|
+
end
|
129
|
+
|
130
|
+
|
131
|
+
class ReqMethodTokenDefinition < FormatTokenDefinition
|
132
|
+
@name = :req_method
|
133
|
+
@abbrev = "%m"
|
134
|
+
@regex = %q![A-Z]+!
|
135
|
+
@caster = nil
|
136
|
+
@derivation_rule = NullDerivationRule.new
|
137
|
+
@captured = true
|
138
|
+
end
|
139
|
+
|
140
|
+
|
141
|
+
class ProtocolTokenDefinition < FormatTokenDefinition
|
142
|
+
@name = :protocol
|
143
|
+
@abbrev = "%H"
|
144
|
+
@regex = %q!\S+!
|
145
|
+
@caster = nil
|
146
|
+
@derivation_rule = NullDerivationRule.new
|
147
|
+
@captured = true
|
148
|
+
end
|
149
|
+
|
150
|
+
|
151
|
+
# Finds log format elements given information about them.
|
152
|
+
class TokenDictionary
|
153
|
+
@@_defs = [
|
154
|
+
RemoteHostTokenDefinition,
|
155
|
+
LogNameTokenDefinition,
|
156
|
+
RemoteUserTokenDefinition,
|
157
|
+
TimeTokenDefinition,
|
158
|
+
ReqFirstlineTokenDefinition,
|
159
|
+
StatusTokenDefinition,
|
160
|
+
BytesSentTokenDefinition,
|
161
|
+
BytesSentTokenDefinition,
|
162
|
+
BytesSentWithHeadersTokenDefinition,
|
163
|
+
ServeTimeMicroTokenDefinition,
|
164
|
+
UrlPathTokenDefinition,
|
165
|
+
QueryStringTokenDefinition,
|
166
|
+
ReqMethodTokenDefinition,
|
167
|
+
ProtocolTokenDefinition
|
168
|
+
]
|
169
|
+
|
170
|
+
# Returns the FormatToken subclass with the given abbreviation.
|
171
|
+
#
|
172
|
+
# If none exists, returns nil.
|
173
|
+
def self.fetch(abbrev)
|
174
|
+
@@_defs.each do |token_def|
|
175
|
+
if token_def.abbrev == abbrev
|
176
|
+
return token_def
|
177
|
+
end
|
178
|
+
end
|
179
|
+
|
180
|
+
nil
|
181
|
+
end
|
182
|
+
end
|
183
|
+
end
|
data/lib/log_parser.rb
CHANGED
@@ -2,54 +2,58 @@ class ApacheCrunch
|
|
2
2
|
# Parses a log file given a path and a Format instance
|
3
3
|
class LogParser
|
4
4
|
# Initializes the parser with the path to a log file and a EntryParser.
|
5
|
-
def initialize(
|
6
|
-
@
|
7
|
-
@
|
5
|
+
def initialize(entry_parser)
|
6
|
+
@_entry_parser = entry_parser
|
7
|
+
@_log_file = nil
|
8
8
|
|
9
|
-
@
|
10
|
-
@_file = nil
|
9
|
+
@_File = File
|
11
10
|
end
|
12
11
|
|
13
|
-
#
|
14
|
-
|
15
|
-
|
16
|
-
|
17
|
-
def next_entry
|
18
|
-
@_file = @_file_cls.open(@path) if @_file.nil?
|
12
|
+
# Handles dependency injection
|
13
|
+
def dep_inject!(file_cls)
|
14
|
+
@_File = file_cls
|
15
|
+
end
|
19
16
|
|
20
|
-
|
17
|
+
# Returns the next parsed line in the log file as an Entry, or nil if we've reached EOF.
|
18
|
+
def next_entry
|
19
|
+
while line_text = @_log_file.gets
|
20
|
+
# This is if we've reached EOF:
|
21
21
|
return nil if line_text.nil?
|
22
|
-
logline = @entry_parser.from_text(line_text)
|
23
22
|
|
23
|
+
entry = @_entry_parser.parse(@_format, line_text)
|
24
24
|
# The EntryParser returns nil and writes a warning if the line text doesn't
|
25
25
|
# match our expected format.
|
26
|
-
next if
|
26
|
+
next if entry.nil?
|
27
27
|
|
28
|
-
return
|
28
|
+
return entry
|
29
29
|
end
|
30
30
|
end
|
31
31
|
|
32
32
|
# Resets the LogParser's filehandle so we can start over.
|
33
|
-
def
|
34
|
-
@
|
33
|
+
def reset_file!
|
34
|
+
@_log_file.close
|
35
|
+
@_log_file = @_File.open(@_log_file.path)
|
35
36
|
end
|
36
37
|
|
37
|
-
# Makes the LogParser
|
38
|
+
# Makes the LogParser start parsing a new log file
|
38
39
|
#
|
39
40
|
# `new_target` is a writable file object that the parser should start parsing, and if
|
40
|
-
# in_place is true, we actually replace the contents of the current target with those
|
41
|
+
# `in_place` is true, we actually replace the contents of the current target with those
|
41
42
|
# of the new target.
|
42
|
-
def
|
43
|
-
|
43
|
+
def set_file!(new_file)
|
44
|
+
@_log_file.close unless @_log_file.nil?
|
45
|
+
@_log_file = new_file
|
46
|
+
end
|
44
47
|
|
45
|
-
|
46
|
-
|
47
|
-
|
48
|
-
|
49
|
-
|
50
|
-
|
48
|
+
# Replaces the LogParser current file with another. Like, for real, on the filesystem.
|
49
|
+
def replace_file!(new_file)
|
50
|
+
@_log_file.close
|
51
|
+
@_File.rename(new_file.path, @_log_file.path)
|
52
|
+
@_log_file = @_File.open(@_log_file.path)
|
53
|
+
end
|
51
54
|
|
52
|
-
|
55
|
+
def set_format!(format)
|
56
|
+
@_format = format
|
53
57
|
end
|
54
58
|
end
|
55
59
|
|
@@ -64,11 +68,15 @@ class ApacheCrunch
|
|
64
68
|
# First we generate a Format instance based on the format definition we were given
|
65
69
|
log_format = FormatFactory.from_format_def(format_def)
|
66
70
|
|
67
|
-
# Now we generate a
|
68
|
-
|
71
|
+
# Now we generate a parser for the individual entries
|
72
|
+
entry_parser = EntryParser.new
|
73
|
+
entry_parser.add_progress_meter!(progress_meter)
|
69
74
|
|
70
75
|
# And now we can instantiate and return a LogParser
|
71
|
-
|
76
|
+
log_parser = LogParser.new(entry_parser)
|
77
|
+
log_parser.set_file!(open(path, "r"))
|
78
|
+
log_parser.set_format!(log_format)
|
79
|
+
log_parser
|
72
80
|
end
|
73
81
|
end
|
74
82
|
end
|
data/lib/procedure_dsl.rb
CHANGED
@@ -1,308 +1,318 @@
|
|
1
|
-
|
2
|
-
|
3
|
-
|
4
|
-
|
5
|
-
|
6
|
-
|
1
|
+
require 'element_value_fetcher'
|
2
|
+
|
3
|
+
class ApacheCrunch
|
4
|
+
# Abstract for a procedure routine.
|
5
|
+
class ProcedureRoutine
|
6
|
+
def initialize(log_parser)
|
7
|
+
@_log_parser = log_parser
|
8
|
+
@_current_entry = nil
|
9
|
+
end
|
7
10
|
|
8
|
-
|
9
|
-
|
10
|
-
|
11
|
-
|
11
|
+
# Allows blocks passed to a DSL routine to access parameters from the current log entry
|
12
|
+
def method_missing(sym, *args)
|
13
|
+
@_current_entry.fetch(sym)
|
14
|
+
end
|
12
15
|
|
13
|
-
|
14
|
-
|
15
|
-
|
16
|
-
|
17
|
-
|
18
|
-
|
16
|
+
# Executes the DSL routine using the given block
|
17
|
+
#
|
18
|
+
# Abstract method
|
19
|
+
def execute(&blk)
|
20
|
+
raise "Not implemented"
|
21
|
+
end
|
19
22
|
|
20
|
-
|
21
|
-
|
22
|
-
|
23
|
-
|
23
|
+
# Anything that needs to happen after the routine completes but before it returns its
|
24
|
+
# result can go in here.
|
25
|
+
def finish
|
26
|
+
@_log_parser.reset_file!
|
27
|
+
end
|
24
28
|
end
|
25
|
-
end
|
26
29
|
|
27
30
|
|
28
|
-
# DSL routine that returns the number of log entries where the block evaluates to true
|
29
|
-
class CountWhere < ProcedureRoutine
|
30
|
-
|
31
|
-
|
32
|
-
|
33
|
-
|
34
|
-
|
31
|
+
# DSL routine that returns the number of log entries where the block evaluates to true
|
32
|
+
class CountWhere < ProcedureRoutine
|
33
|
+
def execute(&blk)
|
34
|
+
count = 0
|
35
|
+
while @_current_entry = @_log_parser.next_entry
|
36
|
+
if instance_eval(&blk)
|
37
|
+
count += 1
|
38
|
+
end
|
35
39
|
end
|
40
|
+
count
|
36
41
|
end
|
37
|
-
count
|
38
42
|
end
|
39
|
-
end
|
40
43
|
|
41
44
|
|
42
|
-
# DSL routine that executes the block for every log entry
|
43
|
-
class Each < ProcedureRoutine
|
44
|
-
|
45
|
-
|
46
|
-
|
45
|
+
# DSL routine that executes the block for every log entry
|
46
|
+
class Each < ProcedureRoutine
|
47
|
+
def execute(&blk)
|
48
|
+
while @_current_entry = @_log_parser.next_entry
|
49
|
+
instance_eval(&blk)
|
50
|
+
end
|
47
51
|
end
|
48
52
|
end
|
49
|
-
end
|
50
53
|
|
51
54
|
|
52
|
-
# DSL routine(s) that filter(s) for entries for which the given block evaluates to true
|
53
|
-
#
|
54
|
-
# This can be called as 'filter()', which means the filtering happens in a temporary file, or
|
55
|
-
# as 'filter(path)', which means the filtering happens in the given file. It can also be called
|
56
|
-
# as 'filter!()', which means the filtering happens in place, clobbering what's in apachecrunch's
|
57
|
-
# target file.
|
58
|
-
class Filter < ProcedureRoutine
|
59
|
-
|
60
|
-
|
61
|
-
|
62
|
-
|
63
|
-
|
64
|
-
|
65
|
-
|
55
|
+
# DSL routine(s) that filter(s) for entries for which the given block evaluates to true
|
56
|
+
#
|
57
|
+
# This can be called as 'filter()', which means the filtering happens in a temporary file, or
|
58
|
+
# as 'filter(path)', which means the filtering happens in the given file. It can also be called
|
59
|
+
# as 'filter!()', which means the filtering happens in place, clobbering what's in apachecrunch's
|
60
|
+
# target file.
|
61
|
+
class Filter < ProcedureRoutine
|
62
|
+
def execute(path=nil, in_place=false, &blk)
|
63
|
+
@_in_place = in_place
|
64
|
+
@_results_file = _make_results_file(path, in_place)
|
65
|
+
|
66
|
+
while @_current_entry = @_log_parser.next_entry
|
67
|
+
if instance_eval(&blk)
|
68
|
+
@_results_file.write(@_current_entry.fetch(:text))
|
69
|
+
end
|
66
70
|
end
|
67
71
|
end
|
68
|
-
end
|
69
72
|
|
70
|
-
|
71
|
-
|
72
|
-
|
73
|
+
def finish
|
74
|
+
@_results_file.close
|
75
|
+
@_results_file = open(@_results_file.path)
|
76
|
+
if @_in_place
|
77
|
+
@_log_parser.replace_file!(@_results_file)
|
78
|
+
else
|
79
|
+
@_log_parser.set_file!(@_results_file)
|
80
|
+
end
|
81
|
+
end
|
73
82
|
|
74
|
-
|
75
|
-
|
76
|
-
|
77
|
-
|
78
|
-
|
79
|
-
|
80
|
-
|
81
|
-
|
83
|
+
# Returns a writable file object to which the results of the filter should be written.
|
84
|
+
def _make_results_file(path, in_place)
|
85
|
+
if path.nil?
|
86
|
+
# If no path passed (this includes the case where the filter is being performed
|
87
|
+
# in place), we want a temp file.
|
88
|
+
return Tempfile.new("apachecrunch")
|
89
|
+
else
|
90
|
+
return open(path, "w")
|
91
|
+
end
|
82
92
|
end
|
83
93
|
end
|
84
|
-
end
|
85
94
|
|
86
|
-
# DSL routine that returns the count of entries with each found value of the given block
|
87
|
-
#
|
88
|
-
# You might for instance run this with the block { status }, and you'd get back something like
|
89
|
-
# {"200" => 941, "301" => 41, "404" => 2, "500" => 0}
|
90
|
-
class CountBy < ProcedureRoutine
|
91
|
-
|
92
|
-
|
93
|
-
|
94
|
-
|
95
|
-
|
96
|
-
|
97
|
-
|
98
|
-
|
95
|
+
# DSL routine that returns the count of entries with each found value of the given block
|
96
|
+
#
|
97
|
+
# You might for instance run this with the block { status }, and you'd get back something like
|
98
|
+
# {"200" => 941, "301" => 41, "404" => 2, "500" => 0}
|
99
|
+
class CountBy < ProcedureRoutine
|
100
|
+
def execute(&blk)
|
101
|
+
counts = {}
|
102
|
+
while @_current_entry = @_log_parser.next_entry
|
103
|
+
val = instance_eval(&blk)
|
104
|
+
if counts.key?(val)
|
105
|
+
counts[val] += 1
|
106
|
+
else
|
107
|
+
counts[val] = 1
|
108
|
+
end
|
99
109
|
end
|
110
|
+
return counts
|
100
111
|
end
|
101
|
-
return counts
|
102
112
|
end
|
103
|
-
end
|
104
113
|
|
105
114
|
|
106
|
-
# DSL routine that finds the distribution of (numeric) values to which the given block evaluates
|
107
|
-
#
|
108
|
-
# For example,
|
109
|
-
#
|
110
|
-
# distribution 100 do
|
111
|
-
# bytes_sent
|
112
|
-
# end
|
113
|
-
#
|
114
|
-
# would return a hash with keys from 0 up by multiples of 100, the value of each being the number
|
115
|
-
# of entries for which bytes_sent is between that key and the next key.
|
116
|
-
class Distribution < ProcedureRoutine
|
117
|
-
|
118
|
-
|
119
|
-
|
120
|
-
|
121
|
-
|
122
|
-
|
123
|
-
|
124
|
-
|
125
|
-
|
115
|
+
# DSL routine that finds the distribution of (numeric) values to which the given block evaluates
|
116
|
+
#
|
117
|
+
# For example,
|
118
|
+
#
|
119
|
+
# distribution 100 do
|
120
|
+
# bytes_sent
|
121
|
+
# end
|
122
|
+
#
|
123
|
+
# would return a hash with keys from 0 up by multiples of 100, the value of each being the number
|
124
|
+
# of entries for which bytes_sent is between that key and the next key.
|
125
|
+
class Distribution < ProcedureRoutine
|
126
|
+
def execute(bucket_width, &blk)
|
127
|
+
dist = {}
|
128
|
+
while @_current_entry = @_log_parser.next_entry
|
129
|
+
val = instance_eval(&blk)
|
130
|
+
k = _key_for(val, bucket_width)
|
131
|
+
if dist.key?(k)
|
132
|
+
dist[k] += 1
|
133
|
+
else
|
134
|
+
dist[k] = 1
|
135
|
+
end
|
126
136
|
end
|
127
|
-
end
|
128
137
|
|
129
|
-
|
130
|
-
|
131
|
-
|
138
|
+
# Backfill keys for which we didn't find a value
|
139
|
+
0.step(dist.keys.max, bucket_width).each do |k|
|
140
|
+
dist[k] = 0 unless dist.key?(k)
|
141
|
+
end
|
142
|
+
|
143
|
+
dist
|
132
144
|
end
|
133
145
|
|
134
|
-
|
146
|
+
# Determines the key for the distribution hash given the value and step
|
147
|
+
def _key_for(val, bucket_width)
|
148
|
+
(val.to_i / bucket_width) * bucket_width
|
149
|
+
end
|
135
150
|
end
|
136
151
|
|
137
|
-
# Determines the key for the distribution hash given the value and step
|
138
|
-
def _key_for(val, bucket_width)
|
139
|
-
(val.to_i / bucket_width) * bucket_width
|
140
|
-
end
|
141
|
-
end
|
142
152
|
|
153
|
+
# Same as Distribution, but the buckets get expenentially wider
|
154
|
+
class LogDistribution < ProcedureRoutine
|
155
|
+
def execute(width_base, &blk)
|
156
|
+
dist = {}
|
157
|
+
while @_current_entry = @_log_parser.next_entry
|
158
|
+
val = instance_eval(&blk)
|
159
|
+
k = _key_for(val, width_base)
|
160
|
+
if dist.key?(k)
|
161
|
+
dist[k] += 1
|
162
|
+
else
|
163
|
+
dist[k] = 1
|
164
|
+
end
|
165
|
+
end
|
143
166
|
|
144
|
-
#
|
145
|
-
|
146
|
-
|
147
|
-
|
148
|
-
|
149
|
-
val = instance_eval(&blk)
|
150
|
-
k = _key_for(val, width_base)
|
151
|
-
if dist.key?(k)
|
152
|
-
dist[k] += 1
|
153
|
-
else
|
154
|
-
dist[k] = 1
|
167
|
+
# Backfill keys for which we didn't find a value
|
168
|
+
k = dist.keys.min
|
169
|
+
max_key = dist.keys.max
|
170
|
+
while k *= width_base and k < max_key
|
171
|
+
dist[k] = 0 unless dist.key?(k)
|
155
172
|
end
|
156
|
-
end
|
157
173
|
|
158
|
-
|
159
|
-
k = dist.keys.min
|
160
|
-
max_key = dist.keys.max
|
161
|
-
while k *= width_base and k < max_key
|
162
|
-
dist[k] = 0 unless dist.key?(k)
|
174
|
+
dist
|
163
175
|
end
|
164
176
|
|
165
|
-
|
166
|
-
|
167
|
-
|
168
|
-
|
169
|
-
|
170
|
-
|
171
|
-
exp = (Math.log(val) / Math.log(width_base)).to_i
|
172
|
-
width_base ** exp
|
177
|
+
# Determines the key for the distribution hash given the value and logarithmic base for
|
178
|
+
# the bucket width
|
179
|
+
def _key_for(val, width_base)
|
180
|
+
exp = (Math.log(val) / Math.log(width_base)).to_i
|
181
|
+
width_base ** exp
|
182
|
+
end
|
173
183
|
end
|
174
|
-
end
|
175
184
|
|
176
185
|
|
177
|
-
# DSL routine that determines a confidence interval for the values to which the block evaluates
|
178
|
-
#
|
179
|
-
# For example,
|
180
|
-
#
|
181
|
-
# confidence_interval 95 do
|
182
|
-
# time_to_serve
|
183
|
-
# end
|
184
|
-
#
|
185
|
-
# would return two numbers, the lower and upper bound of a 95% confidence interval for the values
|
186
|
-
# of time_to_serve.
|
187
|
-
class ConfidenceInterval < ProcedureRoutine
|
188
|
-
|
189
|
-
|
190
|
-
|
191
|
-
|
192
|
-
|
193
|
-
|
194
|
-
|
186
|
+
# DSL routine that determines a confidence interval for the values to which the block evaluates
|
187
|
+
#
|
188
|
+
# For example,
|
189
|
+
#
|
190
|
+
# confidence_interval 95 do
|
191
|
+
# time_to_serve
|
192
|
+
# end
|
193
|
+
#
|
194
|
+
# would return two numbers, the lower and upper bound of a 95% confidence interval for the values
|
195
|
+
# of time_to_serve.
|
196
|
+
class ConfidenceInterval < ProcedureRoutine
|
197
|
+
def execute(confidence, &blk)
|
198
|
+
# Build a list of all the values found
|
199
|
+
values = []
|
200
|
+
while @_current_entry = @_log_parser.next_entry
|
201
|
+
values << instance_eval(&blk)
|
202
|
+
end
|
203
|
+
values.sort!
|
195
204
|
|
196
|
-
|
197
|
-
|
205
|
+
# Determine how many values are outside the bounds of the CI
|
206
|
+
count_outside = (values.length * (1.0 - confidence/100.0)).to_i
|
198
207
|
|
199
|
-
|
200
|
-
|
208
|
+
# Find the bounds of the confidence interval
|
209
|
+
return values[count_outside / 2], values[-count_outside / 2]
|
210
|
+
end
|
201
211
|
end
|
202
|
-
end
|
203
|
-
|
204
212
|
|
205
|
-
# DSL routine that finds the most common n values for the given block.
|
206
|
-
#
|
207
|
-
# Returns a list of lists, each of which is [value, count]. This list is sorted by count.
|
208
|
-
class MostCommon < ProcedureRoutine
|
209
|
-
def execute(n, &blk)
|
210
|
-
counts = CountBy.new(@_log_parser).execute(&blk)
|
211
213
|
|
212
|
-
|
213
|
-
|
214
|
-
|
215
|
-
|
214
|
+
# DSL routine that finds the most common n values for the given block.
|
215
|
+
#
|
216
|
+
# Returns a list of lists, each of which is [value, count]. This list is sorted by count.
|
217
|
+
class MostCommon < ProcedureRoutine
|
218
|
+
def execute(n, &blk)
|
219
|
+
counts = CountBy.new(@_log_parser).execute(&blk)
|
220
|
+
|
221
|
+
# Sort the block values descending
|
222
|
+
sorted_vals = counts.keys.sort do |val_a,val_b|
|
223
|
+
- (counts[val_a] <=> counts[val_b])
|
224
|
+
end
|
216
225
|
|
217
|
-
|
218
|
-
|
226
|
+
sorted_vals[0..n].map do |val|
|
227
|
+
[val, counts[val]]
|
228
|
+
end
|
219
229
|
end
|
220
230
|
end
|
221
|
-
end
|
222
231
|
|
223
232
|
|
224
|
-
# The environment in which a procedure file is evaluated.
|
225
|
-
#
|
226
|
-
# A procedure file is some ruby code that uses our DSL.
|
227
|
-
class ProcedureEnvironment
|
228
|
-
|
229
|
-
|
230
|
-
|
233
|
+
# The environment in which a procedure file is evaluated.
|
234
|
+
#
|
235
|
+
# A procedure file is some ruby code that uses our DSL.
|
236
|
+
class ProcedureEnvironment
|
237
|
+
def initialize(log_parser)
|
238
|
+
@_log_parser = log_parser
|
239
|
+
end
|
231
240
|
|
232
|
-
|
233
|
-
|
234
|
-
|
235
|
-
|
241
|
+
# Evaluates the given string as a procedure in our DSL
|
242
|
+
def eval_procedure(proc_string)
|
243
|
+
eval proc_string
|
244
|
+
end
|
236
245
|
|
237
|
-
|
238
|
-
|
239
|
-
|
240
|
-
|
241
|
-
|
242
|
-
|
243
|
-
|
246
|
+
# DSL routine 'count_where'
|
247
|
+
def count_where(&blk)
|
248
|
+
routine = CountWhere.new(@_log_parser)
|
249
|
+
rv = routine.execute(&blk)
|
250
|
+
routine.finish
|
251
|
+
rv
|
252
|
+
end
|
244
253
|
|
245
|
-
|
246
|
-
|
247
|
-
|
248
|
-
|
249
|
-
|
250
|
-
|
251
|
-
|
254
|
+
# DSL routine 'filter!'
|
255
|
+
def filter!(&blk)
|
256
|
+
routine = Filter.new(@_log_parser)
|
257
|
+
routine.execute(nil, true, &blk)
|
258
|
+
routine.finish
|
259
|
+
nil
|
260
|
+
end
|
252
261
|
|
253
|
-
|
254
|
-
|
255
|
-
|
256
|
-
|
257
|
-
|
258
|
-
|
259
|
-
|
262
|
+
# DSL routine 'filter'
|
263
|
+
def filter(target_path=nil, &blk)
|
264
|
+
routine = Filter.new(@_log_parser)
|
265
|
+
routine.execute(target_path, &blk)
|
266
|
+
routine.finish
|
267
|
+
nil
|
268
|
+
end
|
260
269
|
|
261
|
-
|
262
|
-
|
263
|
-
|
264
|
-
|
265
|
-
|
266
|
-
|
267
|
-
|
270
|
+
# DSL routine 'each'
|
271
|
+
def each(&blk)
|
272
|
+
routine = Each.new(@_log_parser)
|
273
|
+
routine.execute(&blk)
|
274
|
+
routine.finish
|
275
|
+
nil
|
276
|
+
end
|
268
277
|
|
269
|
-
|
270
|
-
|
271
|
-
|
272
|
-
|
273
|
-
|
274
|
-
|
275
|
-
|
278
|
+
# DSL routine 'count_by'
|
279
|
+
def count_by(&blk)
|
280
|
+
routine = CountBy.new(@_log_parser)
|
281
|
+
rv = routine.execute(&blk)
|
282
|
+
routine.finish
|
283
|
+
rv
|
284
|
+
end
|
276
285
|
|
277
|
-
|
278
|
-
|
279
|
-
|
280
|
-
|
281
|
-
|
282
|
-
|
283
|
-
|
286
|
+
# DSL routine 'distribution'
|
287
|
+
def distribution(bucket_width, &blk)
|
288
|
+
routine = Distribution.new(@_log_parser)
|
289
|
+
rv = routine.execute(bucket_width, &blk)
|
290
|
+
routine.finish
|
291
|
+
rv
|
292
|
+
end
|
284
293
|
|
285
|
-
|
286
|
-
|
287
|
-
|
288
|
-
|
289
|
-
|
290
|
-
|
291
|
-
|
294
|
+
# DSL routine 'log_distribution'
|
295
|
+
def log_distribution(width_base, &blk)
|
296
|
+
routine = LogDistribution.new(@_log_parser)
|
297
|
+
rv = routine.execute(width_base, &blk)
|
298
|
+
routine.finish
|
299
|
+
rv
|
300
|
+
end
|
292
301
|
|
293
|
-
|
294
|
-
|
295
|
-
|
296
|
-
|
297
|
-
|
298
|
-
|
299
|
-
|
302
|
+
# DSL routine 'confidence_interval'
|
303
|
+
def confidence_interval(confidence, &blk)
|
304
|
+
routine = ConfidenceInterval.new(@_log_parser)
|
305
|
+
rv = routine.execute(confidence, &blk)
|
306
|
+
routine.finish
|
307
|
+
rv
|
308
|
+
end
|
300
309
|
|
301
|
-
|
302
|
-
|
303
|
-
|
304
|
-
|
305
|
-
|
306
|
-
|
310
|
+
# DSL routine 'most_common'
|
311
|
+
def most_common(n, &blk)
|
312
|
+
routine = MostCommon.new(@_log_parser)
|
313
|
+
rv = routine.execute(n, &blk)
|
314
|
+
routine.finish
|
315
|
+
rv
|
316
|
+
end
|
307
317
|
end
|
308
318
|
end
|