blackwinter-athena 0.0.6
Sign up to get free protection for your applications and to get access to all the features.
- data/COPYING +676 -0
- data/ChangeLog +5 -0
- data/README +33 -0
- data/Rakefile +22 -0
- data/bin/athena +183 -0
- data/example/config.yaml +72 -0
- data/example/example.xml +26 -0
- data/example/sisis-ex.txt +90 -0
- data/lib/athena/formats/dbm.rb +66 -0
- data/lib/athena/formats/ferret.rb +106 -0
- data/lib/athena/formats/lingo.rb +141 -0
- data/lib/athena/formats/sisis.rb +79 -0
- data/lib/athena/formats/xml.rb +274 -0
- data/lib/athena/formats.rb +88 -0
- data/lib/athena/parser.rb +90 -0
- data/lib/athena/record.rb +107 -0
- data/lib/athena/util.rb +46 -0
- data/lib/athena/version.rb +51 -0
- data/lib/athena.rb +75 -0
- metadata +98 -0
@@ -0,0 +1,141 @@
|
|
1
|
+
#--
|
2
|
+
###############################################################################
|
3
|
+
# #
|
4
|
+
# A component of athena, the database file converter. #
|
5
|
+
# #
|
6
|
+
# Copyright (C) 2007-2008 University of Cologne, #
|
7
|
+
# Albertus-Magnus-Platz, #
|
8
|
+
# 50932 Cologne, Germany #
|
9
|
+
# #
|
10
|
+
# Authors: #
|
11
|
+
# Jens Wille <jens.wille@uni-koeln.de> #
|
12
|
+
# #
|
13
|
+
# athena is free software; you can redistribute it and/or modify it under the #
|
14
|
+
# terms of the GNU General Public License as published by the Free Software #
|
15
|
+
# Foundation; either version 3 of the License, or (at your option) any later #
|
16
|
+
# version. #
|
17
|
+
# #
|
18
|
+
# athena is distributed in the hope that it will be useful, but WITHOUT ANY #
|
19
|
+
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS #
|
20
|
+
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more #
|
21
|
+
# details. #
|
22
|
+
# #
|
23
|
+
# You should have received a copy of the GNU General Public License along #
|
24
|
+
# with athena. If not, see <http://www.gnu.org/licenses/>. #
|
25
|
+
# #
|
26
|
+
###############################################################################
|
27
|
+
#++
|
28
|
+
|
29
|
+
require 'iconv'
|
30
|
+
require 'enumerator'
|
31
|
+
|
32
|
+
class Athena::Formats
|
33
|
+
|
34
|
+
module Lingo
|
35
|
+
|
36
|
+
class Base < Athena::Formats
|
37
|
+
|
38
|
+
class << self
|
39
|
+
|
40
|
+
def convert(record)
|
41
|
+
record.struct.inject([]) { |terms, (field, struct)|
|
42
|
+
terms << struct[:elements].inject([]) { |array, element|
|
43
|
+
array += (struct[:values][element] || []).map { |v|
|
44
|
+
(v || '').strip.gsub(/(?:\r?\n)+/, ' ')
|
45
|
+
}.reject { |v| v.empty? }
|
46
|
+
}
|
47
|
+
}
|
48
|
+
end
|
49
|
+
|
50
|
+
def deferred?
|
51
|
+
true
|
52
|
+
end
|
53
|
+
|
54
|
+
private
|
55
|
+
|
56
|
+
def check_number_of_arguments(expected, actual, blow = false, &block)
|
57
|
+
return true if block ? block[actual] : expected == actual
|
58
|
+
|
59
|
+
msg = "wrong number of arguments for #{self} (#{actual} for #{expected})"
|
60
|
+
|
61
|
+
if blow
|
62
|
+
raise FormatArgumentError, msg
|
63
|
+
else
|
64
|
+
warn msg
|
65
|
+
return false
|
66
|
+
end
|
67
|
+
end
|
68
|
+
|
69
|
+
def check_number_of_arguments!(expected, actual, &block)
|
70
|
+
check_number_of_arguments(expected, actual, true, &block)
|
71
|
+
end
|
72
|
+
|
73
|
+
end
|
74
|
+
|
75
|
+
end
|
76
|
+
|
77
|
+
# "Nasenbär\n"
|
78
|
+
class SingleWord < Athena::Formats::Lingo::Base
|
79
|
+
|
80
|
+
register_formats :out, 'lingo/single_word'
|
81
|
+
|
82
|
+
def self.convert(record)
|
83
|
+
super.flatten
|
84
|
+
end
|
85
|
+
|
86
|
+
end
|
87
|
+
|
88
|
+
# "John Vorhauer*Vorhauer, John\n"
|
89
|
+
class KeyValue < Athena::Formats::Lingo::Base
|
90
|
+
|
91
|
+
register_formats :out, 'lingo/key_value'
|
92
|
+
|
93
|
+
def self.convert(record)
|
94
|
+
super.map { |terms|
|
95
|
+
next unless check_number_of_arguments(2, terms.size)
|
96
|
+
|
97
|
+
terms.join('*')
|
98
|
+
}.compact
|
99
|
+
end
|
100
|
+
|
101
|
+
end
|
102
|
+
|
103
|
+
# "Essen,essen #v Essen #s Esse #s\n"
|
104
|
+
class WordClass < Athena::Formats::Lingo::Base
|
105
|
+
|
106
|
+
register_formats :out, 'lingo/word_class'
|
107
|
+
|
108
|
+
def self.convert(record)
|
109
|
+
super.map { |terms|
|
110
|
+
next unless check_number_of_arguments('odd, > 1', terms.size) { |actual|
|
111
|
+
actual > 1 && actual % 2 == 1
|
112
|
+
}
|
113
|
+
|
114
|
+
[terms.shift, terms.to_enum(:each_slice, 2).map { |form, wc|
|
115
|
+
"#{form} ##{wc}"
|
116
|
+
}.join(' ')].join(',')
|
117
|
+
}.compact
|
118
|
+
end
|
119
|
+
|
120
|
+
end
|
121
|
+
|
122
|
+
# "Fax;Faxkopie;Telefax\n"
|
123
|
+
class MultiValue < Athena::Formats::Lingo::Base
|
124
|
+
|
125
|
+
register_formats :out, 'lingo/multi_value', 'lingo/multi_key'
|
126
|
+
|
127
|
+
def self.convert(record)
|
128
|
+
super.map { |terms|
|
129
|
+
next unless check_number_of_arguments('> 1', terms.size) { |actual|
|
130
|
+
actual > 1
|
131
|
+
}
|
132
|
+
|
133
|
+
terms.join(';')
|
134
|
+
}.compact
|
135
|
+
end
|
136
|
+
|
137
|
+
end
|
138
|
+
|
139
|
+
end
|
140
|
+
|
141
|
+
end
|
@@ -0,0 +1,79 @@
|
|
1
|
+
#--
|
2
|
+
###############################################################################
|
3
|
+
# #
|
4
|
+
# A component of athena, the database file converter. #
|
5
|
+
# #
|
6
|
+
# Copyright (C) 2007-2008 University of Cologne, #
|
7
|
+
# Albertus-Magnus-Platz, #
|
8
|
+
# 50932 Cologne, Germany #
|
9
|
+
# #
|
10
|
+
# Authors: #
|
11
|
+
# Jens Wille <jens.wille@uni-koeln.de> #
|
12
|
+
# #
|
13
|
+
# athena is free software; you can redistribute it and/or modify it under the #
|
14
|
+
# terms of the GNU General Public License as published by the Free Software #
|
15
|
+
# Foundation; either version 3 of the License, or (at your option) any later #
|
16
|
+
# version. #
|
17
|
+
# #
|
18
|
+
# athena is distributed in the hope that it will be useful, but WITHOUT ANY #
|
19
|
+
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS #
|
20
|
+
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more #
|
21
|
+
# details. #
|
22
|
+
# #
|
23
|
+
# You should have received a copy of the GNU General Public License along #
|
24
|
+
# with athena. If not, see <http://www.gnu.org/licenses/>. #
|
25
|
+
# #
|
26
|
+
###############################################################################
|
27
|
+
#++
|
28
|
+
|
29
|
+
class Athena::Formats
|
30
|
+
|
31
|
+
class Sisis < Athena::Formats
|
32
|
+
|
33
|
+
register_format :in, 'sisis'
|
34
|
+
|
35
|
+
attr_reader :record_element, :config, :parser
|
36
|
+
|
37
|
+
def initialize(parser)
|
38
|
+
config = parser.config.dup
|
39
|
+
|
40
|
+
case @record_element = config.delete(:__record_element)
|
41
|
+
when String
|
42
|
+
# fine!
|
43
|
+
when nil
|
44
|
+
raise NoRecordElementError, 'no record element specified'
|
45
|
+
else
|
46
|
+
raise IllegalRecordElementError, "illegal record element #{@record_element}"
|
47
|
+
end
|
48
|
+
|
49
|
+
@config = config
|
50
|
+
@parser = parser
|
51
|
+
end
|
52
|
+
|
53
|
+
def parse(source)
|
54
|
+
record = nil
|
55
|
+
|
56
|
+
source.each { |line|
|
57
|
+
element, value = line.match(/(\d+).*?:\s*(.*)/)[1, 2]
|
58
|
+
|
59
|
+
case element
|
60
|
+
when record_element
|
61
|
+
record.close if record
|
62
|
+
record = Athena::Record.new(parser.block, value)
|
63
|
+
else
|
64
|
+
record.update(element, value, config[element])
|
65
|
+
end
|
66
|
+
}
|
67
|
+
|
68
|
+
record.close if record
|
69
|
+
end
|
70
|
+
|
71
|
+
class NoRecordElementError < StandardError
|
72
|
+
end
|
73
|
+
|
74
|
+
class IllegalRecordElementError < StandardError
|
75
|
+
end
|
76
|
+
|
77
|
+
end
|
78
|
+
|
79
|
+
end
|
@@ -0,0 +1,274 @@
|
|
1
|
+
#--
|
2
|
+
###############################################################################
|
3
|
+
# #
|
4
|
+
# A component of athena, the database file converter. #
|
5
|
+
# #
|
6
|
+
# Copyright (C) 2007-2008 University of Cologne, #
|
7
|
+
# Albertus-Magnus-Platz, #
|
8
|
+
# 50932 Cologne, Germany #
|
9
|
+
# #
|
10
|
+
# Authors: #
|
11
|
+
# Jens Wille <jens.wille@uni-koeln.de> #
|
12
|
+
# #
|
13
|
+
# athena is free software; you can redistribute it and/or modify it under the #
|
14
|
+
# terms of the GNU General Public License as published by the Free Software #
|
15
|
+
# Foundation; either version 3 of the License, or (at your option) any later #
|
16
|
+
# version. #
|
17
|
+
# #
|
18
|
+
# athena is distributed in the hope that it will be useful, but WITHOUT ANY #
|
19
|
+
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS #
|
20
|
+
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more #
|
21
|
+
# details. #
|
22
|
+
# #
|
23
|
+
# You should have received a copy of the GNU General Public License along #
|
24
|
+
# with athena. If not, see <http://www.gnu.org/licenses/>. #
|
25
|
+
# #
|
26
|
+
###############################################################################
|
27
|
+
#++
|
28
|
+
|
29
|
+
require 'forwardable'
|
30
|
+
|
31
|
+
require 'rubygems'
|
32
|
+
|
33
|
+
require 'xmlstreamin'
|
34
|
+
require 'nuggets/hash/insert'
|
35
|
+
|
36
|
+
class Athena::Formats
|
37
|
+
|
38
|
+
class XML < Athena::Formats
|
39
|
+
|
40
|
+
include Athena::Util
|
41
|
+
|
42
|
+
register_format :in, 'xml'
|
43
|
+
|
44
|
+
attr_reader :spec, :listener
|
45
|
+
|
46
|
+
def initialize(parser)
|
47
|
+
@spec = build_spec(parser)
|
48
|
+
@listener = XMLStreamin::XMLStreamListener.new(@spec)
|
49
|
+
end
|
50
|
+
|
51
|
+
def parse(source)
|
52
|
+
REXML::Document.parse_stream(source, listener)
|
53
|
+
end
|
54
|
+
|
55
|
+
private
|
56
|
+
|
57
|
+
def build_spec(parser)
|
58
|
+
config = parser.config.dup
|
59
|
+
|
60
|
+
case record_element = config.delete(:__record_element)
|
61
|
+
when String
|
62
|
+
# fine!
|
63
|
+
when nil
|
64
|
+
raise NoRecordElementError, 'no record element specified'
|
65
|
+
else
|
66
|
+
raise IllegalRecordElementError, "illegal record element #{record_element}"
|
67
|
+
end
|
68
|
+
|
69
|
+
element_specs = config.inject({}) { |specs, (element, element_spec)|
|
70
|
+
element_spec.each { |field, c|
|
71
|
+
element.split('/').reverse.inject({}) { |hash, part|
|
72
|
+
s = define_spec(element, field, c, hash.empty? ? :default : hash)
|
73
|
+
merge_specs(hash, part, s)
|
74
|
+
}.each { |key, s|
|
75
|
+
merge_specs(specs, key, s)
|
76
|
+
}
|
77
|
+
}
|
78
|
+
|
79
|
+
specs
|
80
|
+
}
|
81
|
+
|
82
|
+
record_spec = RecordSpec.new(parser)
|
83
|
+
record_spec.specs!(element_specs)
|
84
|
+
|
85
|
+
root_spec = BaseSpec.new
|
86
|
+
root_spec.specs!(record_element => record_spec)
|
87
|
+
|
88
|
+
spec = BaseSpec.new
|
89
|
+
spec.default!(root_spec)
|
90
|
+
|
91
|
+
verbose(:spec, BaseSpec) do
|
92
|
+
spec.inspect_spec
|
93
|
+
end
|
94
|
+
|
95
|
+
spec
|
96
|
+
end
|
97
|
+
|
98
|
+
def define_spec(element, field, config, arg)
|
99
|
+
spec = ElementSpec.new(element, field, config)
|
100
|
+
|
101
|
+
case arg
|
102
|
+
when Hash
|
103
|
+
spec.specs!(arg)
|
104
|
+
else
|
105
|
+
spec.default!(SubElementSpec.new(spec))
|
106
|
+
end
|
107
|
+
|
108
|
+
spec
|
109
|
+
end
|
110
|
+
|
111
|
+
def merge_specs(container, key, spec)
|
112
|
+
container.insert!(key, spec) { |s1, s2|
|
113
|
+
if s1.respond_to?(:specs!)
|
114
|
+
s1.specs!(s2.respond_to?(:specs) ? s2.specs : s2)
|
115
|
+
s1
|
116
|
+
else
|
117
|
+
s1.merge(s2)
|
118
|
+
end
|
119
|
+
}
|
120
|
+
end
|
121
|
+
|
122
|
+
class BaseSpec < XMLStreamin::XMLSpec
|
123
|
+
|
124
|
+
include Athena::Util
|
125
|
+
|
126
|
+
@level = 0
|
127
|
+
|
128
|
+
def start(context, name, attrs)
|
129
|
+
verbose(:xml) do
|
130
|
+
spit "#{indent(level)}<#{name}>"
|
131
|
+
step :down
|
132
|
+
|
133
|
+
attrs.each { |attr|
|
134
|
+
spit "#{indent(level + 1)}[#{attr[0]} = #{attr[1]}]"
|
135
|
+
}
|
136
|
+
end
|
137
|
+
|
138
|
+
return context
|
139
|
+
end
|
140
|
+
|
141
|
+
def text(context, data)
|
142
|
+
verbose(:xml) do
|
143
|
+
content = data.strip
|
144
|
+
spit "#{indent(level)}#{content}" unless content.empty?
|
145
|
+
end
|
146
|
+
|
147
|
+
return context
|
148
|
+
end
|
149
|
+
|
150
|
+
def done(context, name)
|
151
|
+
verbose(:xml) do
|
152
|
+
step :up
|
153
|
+
spit "#{indent(level)}</#{name}>"
|
154
|
+
end
|
155
|
+
|
156
|
+
return context
|
157
|
+
end
|
158
|
+
|
159
|
+
def empty(context)
|
160
|
+
verbose(:xml) do
|
161
|
+
step :up
|
162
|
+
end
|
163
|
+
|
164
|
+
return context
|
165
|
+
end
|
166
|
+
|
167
|
+
def inspect_spec(element = nil, level = 0)
|
168
|
+
if respond_to?(:field)
|
169
|
+
msg = "#{indent(level)}[#{element}] #{field.to_s.upcase} -> #{name}"
|
170
|
+
respond_to?(:spit) ? spit(msg) : warn(msg)
|
171
|
+
specs.each { |e, s|
|
172
|
+
s.inspect_spec(e, level + 1)
|
173
|
+
}
|
174
|
+
else
|
175
|
+
if specs.empty?
|
176
|
+
specs.default.inspect_spec('?', level)
|
177
|
+
else
|
178
|
+
specs.each { |e, s|
|
179
|
+
s.inspect_spec(e, level)
|
180
|
+
}
|
181
|
+
end
|
182
|
+
end
|
183
|
+
end
|
184
|
+
|
185
|
+
private
|
186
|
+
|
187
|
+
def level
|
188
|
+
BaseSpec.instance_variable_get(:@level)
|
189
|
+
end
|
190
|
+
|
191
|
+
def step(direction)
|
192
|
+
steps = { :down => 1, :up => -1 }
|
193
|
+
BaseSpec.instance_variable_set(:@level, level + steps[direction])
|
194
|
+
end
|
195
|
+
|
196
|
+
end
|
197
|
+
|
198
|
+
class RecordSpec < BaseSpec
|
199
|
+
|
200
|
+
attr_reader :parser
|
201
|
+
attr_accessor :record
|
202
|
+
|
203
|
+
def initialize(parser)
|
204
|
+
super()
|
205
|
+
|
206
|
+
@parser = parser
|
207
|
+
end
|
208
|
+
|
209
|
+
def start(context, name, attrs)
|
210
|
+
super
|
211
|
+
|
212
|
+
self.record = Athena::Record.new(parser.block)
|
213
|
+
end
|
214
|
+
|
215
|
+
def done(context, name)
|
216
|
+
super
|
217
|
+
|
218
|
+
record.close
|
219
|
+
end
|
220
|
+
|
221
|
+
end
|
222
|
+
|
223
|
+
class ElementSpec < BaseSpec
|
224
|
+
|
225
|
+
attr_reader :name, :field, :config
|
226
|
+
attr_accessor :record
|
227
|
+
|
228
|
+
def initialize(name, field, config)
|
229
|
+
super()
|
230
|
+
|
231
|
+
@name = name
|
232
|
+
@field = field
|
233
|
+
@config = config
|
234
|
+
end
|
235
|
+
|
236
|
+
def start(context, name, attrs)
|
237
|
+
super
|
238
|
+
|
239
|
+
self.record = Athena::Record[field, config]
|
240
|
+
end
|
241
|
+
|
242
|
+
def text(context, data)
|
243
|
+
super
|
244
|
+
|
245
|
+
record.update(name, data)
|
246
|
+
end
|
247
|
+
|
248
|
+
end
|
249
|
+
|
250
|
+
class SubElementSpec < BaseSpec
|
251
|
+
|
252
|
+
extend Forwardable
|
253
|
+
|
254
|
+
# Forward to parent element; need to specify *all* its attributes and methods
|
255
|
+
def_delegators :@parent, :name, :field, :config, :record, :start, :text
|
256
|
+
|
257
|
+
def initialize(parent)
|
258
|
+
super()
|
259
|
+
|
260
|
+
@parent = parent
|
261
|
+
default!(self)
|
262
|
+
end
|
263
|
+
|
264
|
+
end
|
265
|
+
|
266
|
+
class NoRecordElementError < StandardError
|
267
|
+
end
|
268
|
+
|
269
|
+
class IllegalRecordElementError < StandardError
|
270
|
+
end
|
271
|
+
|
272
|
+
end
|
273
|
+
|
274
|
+
end
|
@@ -0,0 +1,88 @@
|
|
1
|
+
#--
|
2
|
+
###############################################################################
|
3
|
+
# #
|
4
|
+
# A component of athena, the database file converter. #
|
5
|
+
# #
|
6
|
+
# Copyright (C) 2007-2008 University of Cologne, #
|
7
|
+
# Albertus-Magnus-Platz, #
|
8
|
+
# 50932 Cologne, Germany #
|
9
|
+
# #
|
10
|
+
# Authors: #
|
11
|
+
# Jens Wille <jens.wille@uni-koeln.de> #
|
12
|
+
# #
|
13
|
+
# athena is free software; you can redistribute it and/or modify it under the #
|
14
|
+
# terms of the GNU General Public License as published by the Free Software #
|
15
|
+
# Foundation; either version 3 of the License, or (at your option) any later #
|
16
|
+
# version. #
|
17
|
+
# #
|
18
|
+
# athena is distributed in the hope that it will be useful, but WITHOUT ANY #
|
19
|
+
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS #
|
20
|
+
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more #
|
21
|
+
# details. #
|
22
|
+
# #
|
23
|
+
# You should have received a copy of the GNU General Public License along #
|
24
|
+
# with athena. If not, see <http://www.gnu.org/licenses/>. #
|
25
|
+
# #
|
26
|
+
###############################################################################
|
27
|
+
#++
|
28
|
+
|
29
|
+
class Athena::Formats
|
30
|
+
|
31
|
+
@formats = { :in => {}, :out => {} }
|
32
|
+
|
33
|
+
class << self
|
34
|
+
|
35
|
+
def formats
|
36
|
+
Athena::Formats.instance_variable_get(:@formats)
|
37
|
+
end
|
38
|
+
|
39
|
+
def [](direction, format)
|
40
|
+
formats[direction][format]
|
41
|
+
end
|
42
|
+
|
43
|
+
def valid_format?(direction, format)
|
44
|
+
formats[direction].has_key?(format)
|
45
|
+
end
|
46
|
+
|
47
|
+
def deferred?
|
48
|
+
false
|
49
|
+
end
|
50
|
+
|
51
|
+
def convert(*args)
|
52
|
+
raise NotImplementedError, 'must be defined by sub-class'
|
53
|
+
end
|
54
|
+
|
55
|
+
private
|
56
|
+
|
57
|
+
def register_format(direction, format)
|
58
|
+
if existing = formats[direction][format]
|
59
|
+
raise DuplicateFormatDefinitionError,
|
60
|
+
"format already defined (#{direction}): #{format} = #{existing}"
|
61
|
+
end
|
62
|
+
|
63
|
+
formats[direction][format] = self
|
64
|
+
end
|
65
|
+
|
66
|
+
def register_formats(direction, *formats)
|
67
|
+
formats.each { |format|
|
68
|
+
register_format(direction, format)
|
69
|
+
}
|
70
|
+
end
|
71
|
+
|
72
|
+
end
|
73
|
+
|
74
|
+
def parse(*args)
|
75
|
+
raise NotImplementedError, 'must be defined by sub-class'
|
76
|
+
end
|
77
|
+
|
78
|
+
class DuplicateFormatDefinitionError < StandardError
|
79
|
+
end
|
80
|
+
|
81
|
+
class FormatArgumentError < ArgumentError
|
82
|
+
end
|
83
|
+
|
84
|
+
end
|
85
|
+
|
86
|
+
Dir[__FILE__.sub(/\.rb$/, '/**/*.rb')].each { |rb|
|
87
|
+
require rb
|
88
|
+
}
|
@@ -0,0 +1,90 @@
|
|
1
|
+
#--
|
2
|
+
###############################################################################
|
3
|
+
# #
|
4
|
+
# A component of athena, the database file converter. #
|
5
|
+
# #
|
6
|
+
# Copyright (C) 2007-2008 University of Cologne, #
|
7
|
+
# Albertus-Magnus-Platz, #
|
8
|
+
# 50932 Cologne, Germany #
|
9
|
+
# #
|
10
|
+
# Authors: #
|
11
|
+
# Jens Wille <jens.wille@uni-koeln.de> #
|
12
|
+
# #
|
13
|
+
# athena is free software; you can redistribute it and/or modify it under the #
|
14
|
+
# terms of the GNU General Public License as published by the Free Software #
|
15
|
+
# Foundation; either version 3 of the License, or (at your option) any later #
|
16
|
+
# version. #
|
17
|
+
# #
|
18
|
+
# athena is distributed in the hope that it will be useful, but WITHOUT ANY #
|
19
|
+
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS #
|
20
|
+
# FOR A PARTICULAR PURPOSE. See the GNU General Public License for more #
|
21
|
+
# details. #
|
22
|
+
# #
|
23
|
+
# You should have received a copy of the GNU General Public License along #
|
24
|
+
# with athena. If not, see <http://www.gnu.org/licenses/>. #
|
25
|
+
# #
|
26
|
+
###############################################################################
|
27
|
+
#++
|
28
|
+
|
29
|
+
class Athena::Parser
|
30
|
+
|
31
|
+
include Athena::Util
|
32
|
+
|
33
|
+
DEFAULT_SEPARATOR = ', '
|
34
|
+
DEFAULT_EMPTY = '<<EMPTY>>'
|
35
|
+
|
36
|
+
attr_reader :config, :spec
|
37
|
+
attr_accessor :block
|
38
|
+
|
39
|
+
def initialize(config, spec)
|
40
|
+
@config = build_config(config)
|
41
|
+
@spec = Athena::Formats[:in, spec].new(self)
|
42
|
+
end
|
43
|
+
|
44
|
+
def parse(source, &block)
|
45
|
+
self.block = block
|
46
|
+
|
47
|
+
spec.parse(source)
|
48
|
+
Athena::Record.records
|
49
|
+
end
|
50
|
+
|
51
|
+
private
|
52
|
+
|
53
|
+
def build_config(config)
|
54
|
+
config.inject({}) { |hash, (field, v)|
|
55
|
+
if field.to_s =~ /^__/
|
56
|
+
hash.merge(field => v)
|
57
|
+
else
|
58
|
+
case v
|
59
|
+
when String, Array
|
60
|
+
elements = [*v]
|
61
|
+
v = {}
|
62
|
+
when Hash
|
63
|
+
elements = v[:elements] || v[:element].to_a
|
64
|
+
|
65
|
+
raise ArgumentError, "no elements specified for field #{field}" \
|
66
|
+
unless elements.is_a?(Array)
|
67
|
+
else
|
68
|
+
raise ArgumentError, "illegal value for field #{field}"
|
69
|
+
end
|
70
|
+
|
71
|
+
separator = v[:separator] || DEFAULT_SEPARATOR
|
72
|
+
|
73
|
+
elements.each { |element|
|
74
|
+
verbose(:config) do
|
75
|
+
spit "#{field.to_s.upcase} -> #{element}"
|
76
|
+
end
|
77
|
+
|
78
|
+
(hash[element] ||= {})[field] = {
|
79
|
+
:string => v[:string] || ['%s'] * elements.size * separator,
|
80
|
+
:empty => v[:empty] || DEFAULT_EMPTY,
|
81
|
+
:elements => elements
|
82
|
+
}
|
83
|
+
}
|
84
|
+
|
85
|
+
hash
|
86
|
+
end
|
87
|
+
}
|
88
|
+
end
|
89
|
+
|
90
|
+
end
|