polyrex-parser 0.2.8 → 0.3.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/lib/polyrex-parser.rb +64 -13
- metadata +13 -11
data/lib/polyrex-parser.rb
CHANGED
@@ -13,48 +13,99 @@ class PolyrexParser
|
|
13
13
|
def to_a()
|
14
14
|
@a
|
15
15
|
end
|
16
|
+
|
17
|
+
def to_s()
|
18
|
+
name, value, attributes, *remaining = @a
|
19
|
+
[value.strip, scan_a(remaining)].flatten.join(' ')
|
20
|
+
end
|
16
21
|
|
17
22
|
private
|
18
23
|
|
19
|
-
def
|
24
|
+
def scan_a(a)
|
25
|
+
a.inject([]) do |r, x|
|
26
|
+
name, value, attributes, *remaining = x
|
27
|
+
text_remaining = scan_a remaining if remaining
|
20
28
|
|
29
|
+
value = '' if name == 'format_mask' or name == 'schema' or name == 'recordx_type'
|
30
|
+
r << value.strip << text_remaining if value
|
31
|
+
end
|
32
|
+
end
|
33
|
+
|
34
|
+
def parse(s)
|
21
35
|
s.instance_eval{
|
22
36
|
def fetch_summary()
|
23
|
-
|
24
|
-
|
25
|
-
|
37
|
+
name = 'summary'
|
38
|
+
self.slice(((self =~ /<#{name}>/) + name.length + 2) .. \
|
39
|
+
(self =~ /<\/#{name}>/m) - 1) if self[/<#{name}>/]
|
26
40
|
end
|
27
41
|
def fetch_records()
|
28
|
-
|
29
|
-
|
30
|
-
|
42
|
+
name = 'records'
|
43
|
+
self.slice(((self =~ /<#{name}/) + name.length + 2) .. \
|
44
|
+
(self.rindex(/<\/#{name}>/m)) - 1) if self[/<\/#{name}/]
|
31
45
|
end
|
32
46
|
}
|
33
47
|
|
34
48
|
root_name, raw_attributes = s.match(/<(\w+)(\s[^\/>]+)?/).captures
|
35
49
|
attributes = get_attributes(raw_attributes) if raw_attributes
|
50
|
+
raw_summary = s.fetch_summary
|
36
51
|
|
37
|
-
summary = RexleParser.new("<summary>#{
|
52
|
+
summary = RexleParser.new("<summary>#{raw_summary}</summary>").to_a
|
38
53
|
|
39
54
|
raw_records = s.fetch_records
|
40
55
|
records = nil
|
41
56
|
|
57
|
+
|
42
58
|
if raw_records then
|
59
|
+
|
43
60
|
node_name = raw_records[/<(\w+)/,1]
|
44
|
-
|
45
|
-
|
61
|
+
|
62
|
+
#record_threads = raw_records.strip.split(/(?=<#{node_name}[^>]*>)/).map do |x|
|
63
|
+
|
64
|
+
a = []
|
65
|
+
i = 0
|
66
|
+
|
67
|
+
while i < raw_records.strip.length do
|
68
|
+
i = scan_s(raw_records, node_name, i) + 1
|
69
|
+
a << i
|
70
|
+
end
|
71
|
+
|
72
|
+
|
73
|
+
record_threads = ([0] + a).each_cons(2).map do |s1, s2|
|
74
|
+
raw_s = raw_records[s1...s2]
|
75
|
+
|
76
|
+
Thread.new{ Thread.current[:record] = parse(raw_s) }
|
46
77
|
end
|
47
78
|
records = record_threads.map{|x| x.join; x[:record]}
|
79
|
+
|
48
80
|
end
|
49
81
|
|
50
82
|
[root_name, "", attributes ||= {}, [*summary], ['records', "",{}, *records]]
|
83
|
+
|
51
84
|
end
|
52
85
|
|
53
|
-
|
86
|
+
|
87
|
+
def scan_s(s, node_name, instances=0, i=0)
|
88
|
+
|
89
|
+
r = s[i..-1] =~ /<\/?#{node_name}/
|
90
|
+
l = node_name.length + 1
|
91
|
+
return s.length if r.nil?
|
92
|
+
|
93
|
+
if s.slice(i + r,l) == "<#{node_name}" then
|
94
|
+
scan_s(s, node_name, instances+1, i + r + l)
|
95
|
+
else
|
96
|
+
if instances > 1 then
|
97
|
+
scan_s(s, node_name, instances - 1, i + r + node_name.length + 3)
|
98
|
+
else
|
99
|
+
return i + r + node_name.length + 2
|
100
|
+
end
|
101
|
+
end
|
102
|
+
|
103
|
+
end
|
104
|
+
|
105
|
+
def get_attributes(raw_attributes)
|
54
106
|
raw_attributes.scan(/(\w+\='[^']+')|(\w+\="[^"]+")/).map(&:compact).flatten.inject({}) do |r, x|
|
55
107
|
attr_name, val = x.split(/=/)
|
56
108
|
r.merge(attr_name.to_sym => val[1..-2])
|
57
109
|
end
|
58
|
-
end
|
59
|
-
|
110
|
+
end
|
60
111
|
end
|
metadata
CHANGED
@@ -1,27 +1,29 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: polyrex-parser
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
|
4
|
+
prerelease:
|
5
|
+
version: 0.3.0
|
5
6
|
platform: ruby
|
6
|
-
authors:
|
7
|
-
|
7
|
+
authors:
|
8
|
+
- James Robertson
|
8
9
|
autorequire:
|
9
10
|
bindir: bin
|
10
11
|
cert_chain: []
|
11
12
|
|
12
|
-
date: 2011-
|
13
|
+
date: 2011-11-20 00:00:00 +00:00
|
13
14
|
default_executable:
|
14
15
|
dependencies:
|
15
16
|
- !ruby/object:Gem::Dependency
|
16
17
|
name: recordx-parser
|
17
|
-
|
18
|
-
|
19
|
-
|
18
|
+
prerelease: false
|
19
|
+
requirement: &id001 !ruby/object:Gem::Requirement
|
20
|
+
none: false
|
20
21
|
requirements:
|
21
22
|
- - ">="
|
22
23
|
- !ruby/object:Gem::Version
|
23
24
|
version: "0"
|
24
|
-
|
25
|
+
type: :runtime
|
26
|
+
version_requirements: *id001
|
25
27
|
description:
|
26
28
|
email:
|
27
29
|
executables: []
|
@@ -42,21 +44,21 @@ rdoc_options: []
|
|
42
44
|
require_paths:
|
43
45
|
- lib
|
44
46
|
required_ruby_version: !ruby/object:Gem::Requirement
|
47
|
+
none: false
|
45
48
|
requirements:
|
46
49
|
- - ">="
|
47
50
|
- !ruby/object:Gem::Version
|
48
51
|
version: "0"
|
49
|
-
version:
|
50
52
|
required_rubygems_version: !ruby/object:Gem::Requirement
|
53
|
+
none: false
|
51
54
|
requirements:
|
52
55
|
- - ">="
|
53
56
|
- !ruby/object:Gem::Version
|
54
57
|
version: "0"
|
55
|
-
version:
|
56
58
|
requirements: []
|
57
59
|
|
58
60
|
rubyforge_project:
|
59
|
-
rubygems_version: 1.
|
61
|
+
rubygems_version: 1.5.2
|
60
62
|
signing_key:
|
61
63
|
specification_version: 3
|
62
64
|
summary: polyrex-parser
|