datacatalog-importer 0.1.16 → 0.1.17
Sign up to get free protection for your applications and to get access to all the features.
- data/VERSION +1 -1
- data/datacatalog-importer.gemspec +5 -5
- data/lib/pusher.rb +34 -4
- data/lib/utility.rb +14 -3
- data/spec/utility_spec.rb +21 -0
- metadata +4 -4
- data/spec/datacatalog-importer_spec.rb +0 -7
data/VERSION
CHANGED
@@ -1 +1 @@
|
|
1
|
-
0.1.
|
1
|
+
0.1.17
|
@@ -5,7 +5,7 @@
|
|
5
5
|
|
6
6
|
Gem::Specification.new do |s|
|
7
7
|
s.name = %q{datacatalog-importer}
|
8
|
-
s.version = "0.1.
|
8
|
+
s.version = "0.1.17"
|
9
9
|
|
10
10
|
s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
|
11
11
|
s.authors = ["David James"]
|
@@ -32,9 +32,9 @@ Gem::Specification.new do |s|
|
|
32
32
|
"lib/sort_yaml_hash.rb",
|
33
33
|
"lib/tasks.rb",
|
34
34
|
"lib/utility.rb",
|
35
|
-
"spec/datacatalog-importer_spec.rb",
|
36
35
|
"spec/spec.opts",
|
37
|
-
"spec/spec_helper.rb"
|
36
|
+
"spec/spec_helper.rb",
|
37
|
+
"spec/utility_spec.rb"
|
38
38
|
]
|
39
39
|
s.homepage = %q{http://github.com/djsun/datacatalog-importer}
|
40
40
|
s.rdoc_options = ["--charset=UTF-8"]
|
@@ -42,8 +42,8 @@ Gem::Specification.new do |s|
|
|
42
42
|
s.rubygems_version = %q{1.3.6}
|
43
43
|
s.summary = %q{A framework to write National Data Catalog importers}
|
44
44
|
s.test_files = [
|
45
|
-
"spec/
|
46
|
-
"spec/
|
45
|
+
"spec/spec_helper.rb",
|
46
|
+
"spec/utility_spec.rb"
|
47
47
|
]
|
48
48
|
|
49
49
|
if s.respond_to? :specification_version then
|
data/lib/pusher.rb
CHANGED
@@ -87,7 +87,14 @@ module DataCatalog
|
|
87
87
|
n = docs.length
|
88
88
|
if n == 0
|
89
89
|
puts "Creating Organization: #{name}"
|
90
|
-
|
90
|
+
begin
|
91
|
+
DataCatalog::Organization.create(data)
|
92
|
+
rescue DataCatalog::BadRequest => e
|
93
|
+
error("Cannot create Organization", {
|
94
|
+
:params => data,
|
95
|
+
:errors => e.errors,
|
96
|
+
})
|
97
|
+
end
|
91
98
|
else
|
92
99
|
if n > 1
|
93
100
|
warning("Cannot find unique Source with url : #{url}", {
|
@@ -96,7 +103,14 @@ module DataCatalog
|
|
96
103
|
})
|
97
104
|
end
|
98
105
|
puts "Updating Organization: #{name}"
|
99
|
-
|
106
|
+
begin
|
107
|
+
DataCatalog::Organization.update(docs[0].id, data)
|
108
|
+
rescue DataCatalog::BadRequest => e
|
109
|
+
error("Cannot update Organization with id : #{docs[0].id}", {
|
110
|
+
:params => data,
|
111
|
+
:errors => e.errors,
|
112
|
+
})
|
113
|
+
end
|
100
114
|
end
|
101
115
|
end
|
102
116
|
|
@@ -161,10 +175,24 @@ module DataCatalog
|
|
161
175
|
case n
|
162
176
|
when 0
|
163
177
|
puts "- Creating Download: #{data[:format]}"
|
164
|
-
|
178
|
+
begin
|
179
|
+
DataCatalog::Download.create(data)
|
180
|
+
rescue DataCatalog::BadRequest => e
|
181
|
+
error("Cannot create Download", {
|
182
|
+
:params => data,
|
183
|
+
:errors => e.errors,
|
184
|
+
})
|
185
|
+
end
|
165
186
|
when 1
|
166
187
|
puts "- Updating Download: #{data[:format]}"
|
167
|
-
|
188
|
+
begin
|
189
|
+
DataCatalog::Download.update(docs[0].id, data)
|
190
|
+
rescue DataCatalog::BadRequest => e
|
191
|
+
error("Cannot update Download with id : #{docs[0].id}", {
|
192
|
+
:params => data,
|
193
|
+
:errors => e.errors,
|
194
|
+
})
|
195
|
+
end
|
168
196
|
else
|
169
197
|
error("Cannot find unique Download with source_id : #{source.id} and format : #{data[:format]}", {
|
170
198
|
:error => "#{n} matches: " + docs.map { |x| x.id }.join(" ")
|
@@ -232,6 +260,8 @@ module DataCatalog
|
|
232
260
|
:object => object,
|
233
261
|
})
|
234
262
|
nil
|
263
|
+
rescue DataCatalog::BadRequest => e
|
264
|
+
raise Error, "Could not upload Report to API: #{e.errors.inspect}"
|
235
265
|
end
|
236
266
|
|
237
267
|
end
|
data/lib/utility.rb
CHANGED
@@ -9,7 +9,15 @@ module DataCatalog
|
|
9
9
|
# == URLs ==
|
10
10
|
|
11
11
|
def self.absolute_url(page_url, url)
|
12
|
-
|
12
|
+
plain_string(URI.parse(page_url).merge(url).to_s)
|
13
|
+
end
|
14
|
+
|
15
|
+
def self.normalize_url(url)
|
16
|
+
uri = URI.parse(url).normalize
|
17
|
+
unless uri.scheme
|
18
|
+
uri = URI.parse("http://#{url}").normalize
|
19
|
+
end
|
20
|
+
uri.to_s
|
13
21
|
end
|
14
22
|
|
15
23
|
# == Cleaning ==
|
@@ -42,7 +50,7 @@ module DataCatalog
|
|
42
50
|
|
43
51
|
def self.headers
|
44
52
|
{
|
45
|
-
"UserAgent" => "National Data Catalog Importer/0.1.
|
53
|
+
"UserAgent" => "National Data Catalog Importer/0.1.17",
|
46
54
|
}
|
47
55
|
end
|
48
56
|
|
@@ -51,7 +59,9 @@ module DataCatalog
|
|
51
59
|
# For background on rescuing net/http errors, see:
|
52
60
|
# * http://jerith.livejournal.com/40063.html
|
53
61
|
# * http://lindsaar.net/2007/12/9/rbuf_filltimeout-error
|
54
|
-
def self.fetch(uri,
|
62
|
+
def self.fetch(uri, options={})
|
63
|
+
max_attempts = options[:max_attempts] || 3
|
64
|
+
retry_delay = options[:retry_delay] || 5
|
55
65
|
attempts = 0
|
56
66
|
loop do
|
57
67
|
begin
|
@@ -63,6 +73,7 @@ module DataCatalog
|
|
63
73
|
puts " Attempt ##{attempts} failed."
|
64
74
|
puts " Error: #{e.inspect}"
|
65
75
|
break if attempts >= max_attempts
|
76
|
+
sleep(retry_delay)
|
66
77
|
end
|
67
78
|
end
|
68
79
|
end
|
@@ -0,0 +1,21 @@
|
|
1
|
+
require File.expand_path(File.dirname(__FILE__) + '/spec_helper')
|
2
|
+
|
3
|
+
U = DataCatalog::ImporterFramework::Utility
|
4
|
+
|
5
|
+
describe "Utility" do
|
6
|
+
|
7
|
+
describe "normalize_url" do
|
8
|
+
it "add trailing slash if missing" do
|
9
|
+
U.normalize_url("sunlightlabs.com").should == "http://sunlightlabs.com/"
|
10
|
+
end
|
11
|
+
|
12
|
+
it "lowercases" do
|
13
|
+
U.normalize_url("http://SunlightLabs.com/").should == "http://sunlightlabs.com/"
|
14
|
+
end
|
15
|
+
|
16
|
+
it "adds http if missing" do
|
17
|
+
U.normalize_url("sunlightlabs.com/").should == "http://sunlightlabs.com/"
|
18
|
+
end
|
19
|
+
end
|
20
|
+
|
21
|
+
end
|
metadata
CHANGED
@@ -5,8 +5,8 @@ version: !ruby/object:Gem::Version
|
|
5
5
|
segments:
|
6
6
|
- 0
|
7
7
|
- 1
|
8
|
-
-
|
9
|
-
version: 0.1.
|
8
|
+
- 17
|
9
|
+
version: 0.1.17
|
10
10
|
platform: ruby
|
11
11
|
authors:
|
12
12
|
- David James
|
@@ -84,9 +84,9 @@ files:
|
|
84
84
|
- lib/sort_yaml_hash.rb
|
85
85
|
- lib/tasks.rb
|
86
86
|
- lib/utility.rb
|
87
|
-
- spec/datacatalog-importer_spec.rb
|
88
87
|
- spec/spec.opts
|
89
88
|
- spec/spec_helper.rb
|
89
|
+
- spec/utility_spec.rb
|
90
90
|
has_rdoc: true
|
91
91
|
homepage: http://github.com/djsun/datacatalog-importer
|
92
92
|
licenses: []
|
@@ -118,5 +118,5 @@ signing_key:
|
|
118
118
|
specification_version: 3
|
119
119
|
summary: A framework to write National Data Catalog importers
|
120
120
|
test_files:
|
121
|
-
- spec/datacatalog-importer_spec.rb
|
122
121
|
- spec/spec_helper.rb
|
122
|
+
- spec/utility_spec.rb
|