xml_col_finder 0.1.1 → 0.2.3
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- checksums.yaml.gz.sig +0 -0
- data/lib/xml_col_finder.rb +135 -4
- data.tar.gz.sig +0 -0
- metadata +2 -2
- metadata.gz.sig +0 -0
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 4229c7009179e903163ef10f64f0917a7f67e29d588172d184bed94f716123cf
|
4
|
+
data.tar.gz: f543fcbacf5ca517995d28fc6050671b48f18c8ae9df539aad6133486d7b74f5
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 22cd4abb94f055f100d0f56993de883fa87d1bcb2e59ef9edaf0f3f46846f49aa2c73f7e0a38f4779d40314ad5a93bfe91b22738d347bf5866afc1b2ed58c105
|
7
|
+
data.tar.gz: 10c48be3a91d355aa826ebaf1e5b6329e2f3ec7acabc795e8800eb4758fe0abed45cfaee761a91190f72b4407e7b0a0a9accc19b563ca770057a98aa7dfb148b
|
checksums.yaml.gz.sig
CHANGED
Binary file
|
data/lib/xml_col_finder.rb
CHANGED
@@ -12,10 +12,10 @@ class XMLColFinder
|
|
12
12
|
def initialize(s, debug: false)
|
13
13
|
|
14
14
|
@debug = debug
|
15
|
-
doc = Rexle.new(s)
|
15
|
+
@doc = Rexle.new(s)
|
16
16
|
|
17
17
|
a = []
|
18
|
-
doc.root.each_recursive do |node|
|
18
|
+
@doc.root.each_recursive do |node|
|
19
19
|
|
20
20
|
if node.text then
|
21
21
|
a << [BacktrackXPath.new(node, ignore_id: true).to_xpath.split('/'),
|
@@ -30,8 +30,83 @@ class XMLColFinder
|
|
30
30
|
|
31
31
|
end
|
32
32
|
|
33
|
+
def to_code(nametip: true)
|
34
|
+
|
35
|
+
@nametip = nametip
|
36
|
+
@tags = {}
|
37
|
+
|
38
|
+
xpath, remaining = @to_a
|
39
|
+
|
40
|
+
eid = getid(xpath)
|
41
|
+
linex = formatline('doc', eid, xpath)
|
42
|
+
a = scan(remaining, eid)
|
43
|
+
|
44
|
+
lines = a.flatten.compact.prepend linex
|
45
|
+
lines.join("\n").lines\
|
46
|
+
.map {|line| line =~ /.text$/ ? 'puts ' + line : line }.join
|
47
|
+
|
48
|
+
end
|
49
|
+
|
50
|
+
def to_doc()
|
51
|
+
@doc
|
52
|
+
end
|
53
|
+
|
33
54
|
private
|
34
55
|
|
56
|
+
def formatline(pid, eid=nil, key=nil, tail=nil, index: nil)
|
57
|
+
|
58
|
+
if eid then
|
59
|
+
|
60
|
+
nametip = @nametip && tail.is_a?(String)
|
61
|
+
klass = nametip ? key.scan(/@class=['"]([^'"]+)/).last : nil
|
62
|
+
|
63
|
+
line = if klass then
|
64
|
+
desc = klass[0][/^[^\-]+/].gsub(/(?=[A-Z])/,' ').downcase
|
65
|
+
desc += " (e.g. %s)" % [tail.length < 50 ? tail : tail[0..46] + '...']
|
66
|
+
"\n# " + desc + "\n"
|
67
|
+
elsif nametip
|
68
|
+
"\n# e.g. %s\n" % [tail.length < 50 ? tail : tail[0..46] + '...']
|
69
|
+
else
|
70
|
+
''
|
71
|
+
end
|
72
|
+
|
73
|
+
key.gsub!("[@class='']",'') # Rexle XPath bug solution!
|
74
|
+
line += "%s = %s.element(\"%s\")" % [eid, pid, key]
|
75
|
+
if tail.is_a? String
|
76
|
+
line += '.text'
|
77
|
+
#line += "\n" if nametip
|
78
|
+
end
|
79
|
+
|
80
|
+
else
|
81
|
+
line = index ? ("%s[%d].text" % [pid, index]) : ("%s.text" % pid)
|
82
|
+
end
|
83
|
+
|
84
|
+
return line
|
85
|
+
end
|
86
|
+
|
87
|
+
|
88
|
+
def getid(rawtag)
|
89
|
+
|
90
|
+
rawtagx = rawtag.split('/').last[/\w+/]
|
91
|
+
|
92
|
+
tag = case rawtagx.to_sym
|
93
|
+
when :a
|
94
|
+
'link'
|
95
|
+
when :p
|
96
|
+
'para'
|
97
|
+
else
|
98
|
+
rawtagx
|
99
|
+
end
|
100
|
+
|
101
|
+
if @tags.include?(tag) then
|
102
|
+
@tags[tag] =~ /\d+$/ ? @tags[tag].succ! : @tags[tag] += '1'
|
103
|
+
else
|
104
|
+
@tags[tag] = tag
|
105
|
+
end
|
106
|
+
|
107
|
+
end
|
108
|
+
|
109
|
+
|
35
110
|
# Groups xpath by matching branches
|
36
111
|
#
|
37
112
|
def group_by_xpath(a)
|
@@ -73,8 +148,6 @@ class XMLColFinder
|
|
73
148
|
|
74
149
|
else
|
75
150
|
|
76
|
-
puts "path.join('/'): " + path.join('/').inspect
|
77
|
-
puts 'txt:' + txt.inspect
|
78
151
|
|
79
152
|
h2[stickypath.sub(/^\//,'')] ||= []
|
80
153
|
h2[stickypath.sub(/^\//,'')] << [path.join('/'), txt]
|
@@ -99,6 +172,64 @@ class XMLColFinder
|
|
99
172
|
|
100
173
|
end
|
101
174
|
|
175
|
+
def scan(a, eid='doc', pid=eid.clone)
|
176
|
+
|
177
|
+
#puts 'a: ' + a.inspect if @debug
|
178
|
+
|
179
|
+
a.map do |row|
|
180
|
+
|
181
|
+
head, tail = row
|
182
|
+
|
183
|
+
if head.is_a? Array then
|
184
|
+
|
185
|
+
hline = scan(row, eid, pid)
|
186
|
+
|
187
|
+
elsif head
|
188
|
+
|
189
|
+
if head[0] == '/' then
|
190
|
+
|
191
|
+
key = head[1..-1]
|
192
|
+
puts 'key: ' + key.inspect if @debug
|
193
|
+
|
194
|
+
eid = getid(key)
|
195
|
+
|
196
|
+
hline = if tail.is_a? Array and tail.all? {|x| x.is_a? String } then
|
197
|
+
@prev_xpath = true
|
198
|
+
"%s = %s.xpath(\"%s\")" % [eid, pid, key]
|
199
|
+
else
|
200
|
+
@prev_xpath = false
|
201
|
+
formatline(pid, eid, key, tail)
|
202
|
+
end
|
203
|
+
|
204
|
+
else
|
205
|
+
|
206
|
+
hline = formatline(pid=eid)
|
207
|
+
end
|
208
|
+
end
|
209
|
+
|
210
|
+
if tail.is_a? Array then
|
211
|
+
|
212
|
+
if tail.compact[0].is_a? Array then
|
213
|
+
|
214
|
+
puts 'tail: ' + tail.inspect if @debug
|
215
|
+
|
216
|
+
tline = scan(tail, eid)
|
217
|
+
|
218
|
+
elsif tail.all? {|x| x.is_a? String} and tail[0][0] != '/'
|
219
|
+
puts '_tail: ' + tail.inspect if @debug
|
220
|
+
tline = tail.map.with_index do |x,i|
|
221
|
+
formatline(pid=eid, index: i)
|
222
|
+
end
|
223
|
+
|
224
|
+
end
|
225
|
+
end
|
226
|
+
|
227
|
+
[hline, tline]
|
228
|
+
end
|
229
|
+
|
230
|
+
end
|
231
|
+
|
232
|
+
|
102
233
|
def truncate_xpath(records, offset=0)
|
103
234
|
|
104
235
|
records.map do |record|
|
data.tar.gz.sig
CHANGED
Binary file
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: xml_col_finder
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.
|
4
|
+
version: 0.2.3
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- James Robertson
|
@@ -35,7 +35,7 @@ cert_chain:
|
|
35
35
|
GSlZ9ilAfm8srTjbZ2cWQyNGGxH+zHQ3Z02c4ZEtgPv/wHjptd1VeBm0P1aemsRA
|
36
36
|
ShsxXxzmzIrRENmpBp3tyR3k
|
37
37
|
-----END CERTIFICATE-----
|
38
|
-
date: 2022-01-
|
38
|
+
date: 2022-01-24 00:00:00.000000000 Z
|
39
39
|
dependencies:
|
40
40
|
- !ruby/object:Gem::Dependency
|
41
41
|
name: rexle
|
metadata.gz.sig
CHANGED
Binary file
|