turbot-runner-morph 0.0.7 → 0.0.8
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/README.md +18 -0
- data/lib/turbot_runner/version.rb +1 -1
- data/schema/schemas/gazette-notice-schema.json +331 -0
- metadata +43 -45
checksums.yaml
ADDED
@@ -0,0 +1,7 @@
|
|
1
|
+
---
|
2
|
+
SHA1:
|
3
|
+
metadata.gz: 584bcbcb31bf05e6542b4c63afd1c04c3e657212
|
4
|
+
data.tar.gz: 0fe570e207baadd32ad7c79ec7fe9123913eca55
|
5
|
+
SHA512:
|
6
|
+
metadata.gz: e75a0177b33557245a3db3f804db9eca7fec26101edafbd82d7c767c8764c0da92ee5d0e290cfe7f814e8bd2242f7a1503858b399ce0245135bb061d2fcf9be1
|
7
|
+
data.tar.gz: 2f2280ae94d4738172414b60358452b2309c19dc8e1122d6fcc35ee5c640810c71f120616aa53cde8892ce6cc8c7a5c30d8ffc681eab859f2dae2d2d8908d05f
|
data/README.md
ADDED
@@ -0,0 +1,18 @@
|
|
1
|
+
# turbot-runner-morph
|
2
|
+
|
3
|
+
## Getting started
|
4
|
+
|
5
|
+
git submodule update --init
|
6
|
+
cd schema && git checkout master && cd ..
|
7
|
+
|
8
|
+
## Updating the schema
|
9
|
+
|
10
|
+
cd schema && git pull --rebase && cd ..
|
11
|
+
git commit schema -m 'Pull in new schema'
|
12
|
+
|
13
|
+
## Releasing a new version
|
14
|
+
|
15
|
+
Bump the version in `lib/turbot_runner/version.rb`, then:
|
16
|
+
|
17
|
+
git commit lib/turbot_runner/version.rb -m 'Release new version'
|
18
|
+
bundle exec rake release # requires Rubygems credentials
|
@@ -0,0 +1,331 @@
|
|
1
|
+
{
|
2
|
+
"$schema": "http://json-schema.org/draft-04/schema#",
|
3
|
+
"title": "Notice",
|
4
|
+
"description": "A gazette notice",
|
5
|
+
"type": "object",
|
6
|
+
"properties": {
|
7
|
+
"issue": {
|
8
|
+
"description": "The issue in which the notice is published",
|
9
|
+
"$ref": "#/definitions/issue"
|
10
|
+
},
|
11
|
+
"document": {
|
12
|
+
"$ref": "#/definitions/document"
|
13
|
+
},
|
14
|
+
"uid": {
|
15
|
+
"description": "An internal, constructed, unique identifier",
|
16
|
+
"type": "string"
|
17
|
+
},
|
18
|
+
"identifier": {
|
19
|
+
"description": "The publisher's identifier for the notice",
|
20
|
+
"type": [
|
21
|
+
"integer",
|
22
|
+
"string"
|
23
|
+
]
|
24
|
+
},
|
25
|
+
"title": {
|
26
|
+
"description": "The title of the notice",
|
27
|
+
"type": "string"
|
28
|
+
},
|
29
|
+
"description": {
|
30
|
+
"description": "A description of the notice",
|
31
|
+
"type": "string"
|
32
|
+
},
|
33
|
+
"body": {
|
34
|
+
"description": "The content of the notice",
|
35
|
+
"$ref": "#/definitions/body"
|
36
|
+
},
|
37
|
+
"date_published": {
|
38
|
+
"description": "The date of publication of the notice",
|
39
|
+
"type": "string",
|
40
|
+
"format": "date"
|
41
|
+
},
|
42
|
+
"classification": {
|
43
|
+
"description": "The type of the notice",
|
44
|
+
"type": "array",
|
45
|
+
"items": {
|
46
|
+
"$ref": "#/definitions/classification"
|
47
|
+
}
|
48
|
+
},
|
49
|
+
"subjects": {
|
50
|
+
"description": "The entities that the notice is about",
|
51
|
+
"type": "array",
|
52
|
+
"items": {
|
53
|
+
"$ref": "includes/entity.json"
|
54
|
+
}
|
55
|
+
},
|
56
|
+
"about": {
|
57
|
+
"description": "The subject matter of the notice",
|
58
|
+
"type": "object",
|
59
|
+
"properties": {
|
60
|
+
"kind": {
|
61
|
+
"description": "The class of subject matter",
|
62
|
+
"type": "string"
|
63
|
+
},
|
64
|
+
"classification": {
|
65
|
+
"description": "The type of subject matter",
|
66
|
+
"type": "array",
|
67
|
+
"items": {
|
68
|
+
"$ref": "#/definitions/classification"
|
69
|
+
}
|
70
|
+
},
|
71
|
+
"body": {
|
72
|
+
"description": "A description of the item",
|
73
|
+
"$ref": "#/definitions/body"
|
74
|
+
}
|
75
|
+
}
|
76
|
+
},
|
77
|
+
"url": {
|
78
|
+
"description": "The URL of the notice",
|
79
|
+
"type": "string",
|
80
|
+
"format": "uri"
|
81
|
+
},
|
82
|
+
"media_type": {
|
83
|
+
"description": "The media type of the URL resource",
|
84
|
+
"type": "string"
|
85
|
+
},
|
86
|
+
"creator": {
|
87
|
+
"description": "The entity placing the notice",
|
88
|
+
"$ref": "includes/entity.json"
|
89
|
+
},
|
90
|
+
"publisher": {
|
91
|
+
"description": "The publisher of the notice",
|
92
|
+
"$ref": "#/definitions/publisher"
|
93
|
+
},
|
94
|
+
"update_action": {
|
95
|
+
"description": "Whether this notice corrects or cancels an earlier notice",
|
96
|
+
"$ref": "#/definitions/update_action"
|
97
|
+
},
|
98
|
+
"source_url": {
|
99
|
+
"description": "A URL where this fact may be verified",
|
100
|
+
"type": "string",
|
101
|
+
"format": "uri"
|
102
|
+
},
|
103
|
+
"sample_date": {
|
104
|
+
"description": "A date on which this fact is known to be true",
|
105
|
+
"type": "string",
|
106
|
+
"format": "date"
|
107
|
+
},
|
108
|
+
"retrieved_at": {
|
109
|
+
"description": "The time at which the notice was retrieved",
|
110
|
+
"type": "string",
|
111
|
+
"format": "date-time"
|
112
|
+
},
|
113
|
+
"confidence": {
|
114
|
+
"description": "The aggregator's confidence in the accuracy of the data",
|
115
|
+
"type": "string",
|
116
|
+
"enum": [
|
117
|
+
"HIGH",
|
118
|
+
"MEDIUM",
|
119
|
+
"LOW"
|
120
|
+
]
|
121
|
+
},
|
122
|
+
"other_attributes": {
|
123
|
+
"description": "Any other structured data from the notice ",
|
124
|
+
"type": "object"
|
125
|
+
}
|
126
|
+
},
|
127
|
+
"additionalProperties": false,
|
128
|
+
"required": [
|
129
|
+
"issue",
|
130
|
+
"uid",
|
131
|
+
"identifier",
|
132
|
+
"source_url",
|
133
|
+
"sample_date",
|
134
|
+
"retrieved_at",
|
135
|
+
"confidence"
|
136
|
+
],
|
137
|
+
"definitions": {
|
138
|
+
"publisher": {
|
139
|
+
"title": "Publisher",
|
140
|
+
"description": "A gazette publisher",
|
141
|
+
"type": "object",
|
142
|
+
"properties": {
|
143
|
+
"name": {
|
144
|
+
"description": "The name of the publisher",
|
145
|
+
"type": "string"
|
146
|
+
},
|
147
|
+
"identifier": {
|
148
|
+
"description": "The identifier of the publisher",
|
149
|
+
"type": [
|
150
|
+
"string",
|
151
|
+
"integer"
|
152
|
+
]
|
153
|
+
},
|
154
|
+
"url": {
|
155
|
+
"description": "The URL of the publisher",
|
156
|
+
"type": "string",
|
157
|
+
"format": "uri"
|
158
|
+
}
|
159
|
+
},
|
160
|
+
"additionalProperties": false,
|
161
|
+
"required": [
|
162
|
+
"name"
|
163
|
+
]
|
164
|
+
},
|
165
|
+
"publication": {
|
166
|
+
"title": "Publication",
|
167
|
+
"description": "A gazette",
|
168
|
+
"type": "object",
|
169
|
+
"properties": {
|
170
|
+
"publisher": {
|
171
|
+
"description": "The publisher of the gazette",
|
172
|
+
"$ref": "#/definitions/publisher"
|
173
|
+
},
|
174
|
+
"jurisdiction_code": {
|
175
|
+
"description": "The jurisdiction covered by the gazette",
|
176
|
+
"type": "string"
|
177
|
+
},
|
178
|
+
"title": {
|
179
|
+
"description": "The title of the gazette",
|
180
|
+
"type": "string"
|
181
|
+
},
|
182
|
+
"url": {
|
183
|
+
"description": "The URL of the gazette",
|
184
|
+
"type": "string",
|
185
|
+
"format": "uri"
|
186
|
+
}
|
187
|
+
},
|
188
|
+
"additionalProperties": false,
|
189
|
+
"required": [
|
190
|
+
"publisher"
|
191
|
+
]
|
192
|
+
},
|
193
|
+
"issue": {
|
194
|
+
"title": "Issue",
|
195
|
+
"description": "An issue of a gazette",
|
196
|
+
"type": "object",
|
197
|
+
"properties": {
|
198
|
+
"publication": {
|
199
|
+
"description": "The gazette of which the issue is a part",
|
200
|
+
"$ref": "#/definitions/publication"
|
201
|
+
},
|
202
|
+
"identifier": {
|
203
|
+
"description": "The identifier of the issue",
|
204
|
+
"type": [
|
205
|
+
"string",
|
206
|
+
"integer"
|
207
|
+
]
|
208
|
+
},
|
209
|
+
"title": {
|
210
|
+
"description": "The title of the issue",
|
211
|
+
"type": "string"
|
212
|
+
},
|
213
|
+
"edition_id": {
|
214
|
+
"description": "The identifier of the edition of which the issue is a part",
|
215
|
+
"type": [
|
216
|
+
"string",
|
217
|
+
"integer"
|
218
|
+
]
|
219
|
+
},
|
220
|
+
"page_range": {
|
221
|
+
"description": "The page range of the issue",
|
222
|
+
"type": "string",
|
223
|
+
"pattern": "^[0-9]*(-[0-9]*)?$"
|
224
|
+
},
|
225
|
+
"url": {
|
226
|
+
"description": "The URL of the issue",
|
227
|
+
"type": "string",
|
228
|
+
"format": "uri"
|
229
|
+
},
|
230
|
+
"media_type": {
|
231
|
+
"description": "The media type of the URL resource",
|
232
|
+
"type": "string"
|
233
|
+
}
|
234
|
+
},
|
235
|
+
"additionalProperties": false,
|
236
|
+
"required": [
|
237
|
+
"publication"
|
238
|
+
]
|
239
|
+
},
|
240
|
+
"document": {
|
241
|
+
"title": "Document",
|
242
|
+
"description": "A section of an issue",
|
243
|
+
"type": "object",
|
244
|
+
"properties": {
|
245
|
+
"page_range": {
|
246
|
+
"description": "The page range of the document",
|
247
|
+
"type": "string",
|
248
|
+
"pattern": "^[0-9]*(-[0-9]*)?$"
|
249
|
+
},
|
250
|
+
"url": {
|
251
|
+
"description": "The URL of the document",
|
252
|
+
"type": "string",
|
253
|
+
"format": "uri"
|
254
|
+
},
|
255
|
+
"media_type": {
|
256
|
+
"description": "The media type of the URL resource",
|
257
|
+
"type": "string"
|
258
|
+
}
|
259
|
+
},
|
260
|
+
"additionalProperties": false,
|
261
|
+
"required": [
|
262
|
+
"url"
|
263
|
+
]
|
264
|
+
},
|
265
|
+
"classification": {
|
266
|
+
"title": "Classification",
|
267
|
+
"description": "A classification",
|
268
|
+
"type": "object",
|
269
|
+
"properties": {
|
270
|
+
"scheme": {
|
271
|
+
"description": "The scheme of the classification",
|
272
|
+
"type": "string"
|
273
|
+
},
|
274
|
+
"value": {
|
275
|
+
"description": "The classification",
|
276
|
+
"type": "string"
|
277
|
+
}
|
278
|
+
},
|
279
|
+
"additionalProperties": false,
|
280
|
+
"required": [
|
281
|
+
"value"
|
282
|
+
]
|
283
|
+
},
|
284
|
+
"body": {
|
285
|
+
"title": "Body",
|
286
|
+
"description": "Content",
|
287
|
+
"type": "object",
|
288
|
+
"properties": {
|
289
|
+
"value": {
|
290
|
+
"description": "The content",
|
291
|
+
"type": "string"
|
292
|
+
},
|
293
|
+
"media_type": {
|
294
|
+
"description": "The media type of the content",
|
295
|
+
"type": "string",
|
296
|
+
"enum": [
|
297
|
+
"text/plain",
|
298
|
+
"text/html"
|
299
|
+
]
|
300
|
+
}
|
301
|
+
},
|
302
|
+
"additionalProperties": false,
|
303
|
+
"required": [
|
304
|
+
"value"
|
305
|
+
]
|
306
|
+
},
|
307
|
+
"update_action": {
|
308
|
+
"title": "Update action",
|
309
|
+
"description": "An act of changing an object",
|
310
|
+
"type": "object",
|
311
|
+
"properties": {
|
312
|
+
"type": {
|
313
|
+
"description": "The type of change",
|
314
|
+
"type": "string",
|
315
|
+
"enum": [
|
316
|
+
"cancellation",
|
317
|
+
"correction"
|
318
|
+
]
|
319
|
+
},
|
320
|
+
"object": {
|
321
|
+
"description": "The object that was changed",
|
322
|
+
"type": "object"
|
323
|
+
}
|
324
|
+
},
|
325
|
+
"additionalProperties": false,
|
326
|
+
"required": [
|
327
|
+
"object"
|
328
|
+
]
|
329
|
+
}
|
330
|
+
}
|
331
|
+
}
|
metadata
CHANGED
@@ -1,20 +1,18 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: turbot-runner-morph
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.0.
|
5
|
-
prerelease:
|
4
|
+
version: 0.0.8
|
6
5
|
platform: ruby
|
7
6
|
authors:
|
8
7
|
- OpenCorporates
|
9
8
|
autorequire:
|
10
9
|
bindir: bin
|
11
10
|
cert_chain: []
|
12
|
-
date: 2016-01-
|
11
|
+
date: 2016-01-15 00:00:00.000000000 Z
|
13
12
|
dependencies:
|
14
13
|
- !ruby/object:Gem::Dependency
|
15
14
|
name: openc-json_schema
|
16
15
|
requirement: !ruby/object:Gem::Requirement
|
17
|
-
none: false
|
18
16
|
requirements:
|
19
17
|
- - '='
|
20
18
|
- !ruby/object:Gem::Version
|
@@ -22,7 +20,6 @@ dependencies:
|
|
22
20
|
type: :runtime
|
23
21
|
prerelease: false
|
24
22
|
version_requirements: !ruby/object:Gem::Requirement
|
25
|
-
none: false
|
26
23
|
requirements:
|
27
24
|
- - '='
|
28
25
|
- !ruby/object:Gem::Version
|
@@ -33,6 +30,7 @@ executables: []
|
|
33
30
|
extensions: []
|
34
31
|
extra_rdoc_files: []
|
35
32
|
files:
|
33
|
+
- README.md
|
36
34
|
- bin/rspec
|
37
35
|
- lib/turbot_runner.rb
|
38
36
|
- lib/turbot_runner/base_handler.rb
|
@@ -44,6 +42,41 @@ files:
|
|
44
42
|
- lib/turbot_runner/utils.rb
|
45
43
|
- lib/turbot_runner/validator.rb
|
46
44
|
- lib/turbot_runner/version.rb
|
45
|
+
- schema/schemas/accounts-statement-schema.json
|
46
|
+
- schema/schemas/company-schema.json
|
47
|
+
- schema/schemas/filing-schema.json
|
48
|
+
- schema/schemas/financial-payment-schema.json
|
49
|
+
- schema/schemas/gazette-notice-schema.json
|
50
|
+
- schema/schemas/includes/accounts_element.json
|
51
|
+
- schema/schemas/includes/address.json
|
52
|
+
- schema/schemas/includes/alternative_name.json
|
53
|
+
- schema/schemas/includes/company-for-nesting.json
|
54
|
+
- schema/schemas/includes/company.json
|
55
|
+
- schema/schemas/includes/entity.json
|
56
|
+
- schema/schemas/includes/filing.json
|
57
|
+
- schema/schemas/includes/filing_document.json
|
58
|
+
- schema/schemas/includes/financial-payment-data-object.json
|
59
|
+
- schema/schemas/includes/identifier.json
|
60
|
+
- schema/schemas/includes/industry_code.json
|
61
|
+
- schema/schemas/includes/licence-data-object.json
|
62
|
+
- schema/schemas/includes/officer.json
|
63
|
+
- schema/schemas/includes/organisation.json
|
64
|
+
- schema/schemas/includes/permission.json
|
65
|
+
- schema/schemas/includes/person.json
|
66
|
+
- schema/schemas/includes/person_name.json
|
67
|
+
- schema/schemas/includes/previous_name.json
|
68
|
+
- schema/schemas/includes/share-parcel-data.json
|
69
|
+
- schema/schemas/includes/share-parcel.json
|
70
|
+
- schema/schemas/includes/subsidiary-relationship-data.json
|
71
|
+
- schema/schemas/includes/total-shares.json
|
72
|
+
- schema/schemas/includes/unknown_entity_type.json
|
73
|
+
- schema/schemas/licence-schema.json
|
74
|
+
- schema/schemas/primary-data-schema.json
|
75
|
+
- schema/schemas/share-parcel-schema.json
|
76
|
+
- schema/schemas/simple-financial-payment-schema.json
|
77
|
+
- schema/schemas/simple-licence-schema.json
|
78
|
+
- schema/schemas/simple-subsidiary-schema.json
|
79
|
+
- schema/schemas/subsidiary-relationship-schema.json
|
47
80
|
- spec/bots/bot-that-crashes-immediately/manifest.json
|
48
81
|
- spec/bots/bot-that-crashes-immediately/scraper.rb
|
49
82
|
- spec/bots/bot-that-crashes-immediately/transformer1.rb
|
@@ -95,63 +128,28 @@ files:
|
|
95
128
|
- spec/outputs/full-transformer.out
|
96
129
|
- spec/outputs/truncated-scraper.out
|
97
130
|
- spec/spec_helper.rb
|
98
|
-
- schema/schemas/accounts-statement-schema.json
|
99
|
-
- schema/schemas/company-schema.json
|
100
|
-
- schema/schemas/filing-schema.json
|
101
|
-
- schema/schemas/financial-payment-schema.json
|
102
|
-
- schema/schemas/includes/accounts_element.json
|
103
|
-
- schema/schemas/includes/address.json
|
104
|
-
- schema/schemas/includes/alternative_name.json
|
105
|
-
- schema/schemas/includes/company-for-nesting.json
|
106
|
-
- schema/schemas/includes/company.json
|
107
|
-
- schema/schemas/includes/entity.json
|
108
|
-
- schema/schemas/includes/filing.json
|
109
|
-
- schema/schemas/includes/filing_document.json
|
110
|
-
- schema/schemas/includes/financial-payment-data-object.json
|
111
|
-
- schema/schemas/includes/identifier.json
|
112
|
-
- schema/schemas/includes/industry_code.json
|
113
|
-
- schema/schemas/includes/licence-data-object.json
|
114
|
-
- schema/schemas/includes/officer.json
|
115
|
-
- schema/schemas/includes/organisation.json
|
116
|
-
- schema/schemas/includes/permission.json
|
117
|
-
- schema/schemas/includes/person.json
|
118
|
-
- schema/schemas/includes/person_name.json
|
119
|
-
- schema/schemas/includes/previous_name.json
|
120
|
-
- schema/schemas/includes/share-parcel-data.json
|
121
|
-
- schema/schemas/includes/share-parcel.json
|
122
|
-
- schema/schemas/includes/subsidiary-relationship-data.json
|
123
|
-
- schema/schemas/includes/total-shares.json
|
124
|
-
- schema/schemas/includes/unknown_entity_type.json
|
125
|
-
- schema/schemas/licence-schema.json
|
126
|
-
- schema/schemas/primary-data-schema.json
|
127
|
-
- schema/schemas/share-parcel-schema.json
|
128
|
-
- schema/schemas/simple-financial-payment-schema.json
|
129
|
-
- schema/schemas/simple-licence-schema.json
|
130
|
-
- schema/schemas/simple-subsidiary-schema.json
|
131
|
-
- schema/schemas/subsidiary-relationship-schema.json
|
132
131
|
homepage: http://turbot.opencorporates.com/
|
133
132
|
licenses:
|
134
133
|
- MIT
|
134
|
+
metadata: {}
|
135
135
|
post_install_message:
|
136
136
|
rdoc_options: []
|
137
137
|
require_paths:
|
138
138
|
- lib
|
139
139
|
required_ruby_version: !ruby/object:Gem::Requirement
|
140
|
-
none: false
|
141
140
|
requirements:
|
142
|
-
- -
|
141
|
+
- - ">="
|
143
142
|
- !ruby/object:Gem::Version
|
144
143
|
version: 1.9.2
|
145
144
|
required_rubygems_version: !ruby/object:Gem::Requirement
|
146
|
-
none: false
|
147
145
|
requirements:
|
148
|
-
- -
|
146
|
+
- - ">="
|
149
147
|
- !ruby/object:Gem::Version
|
150
148
|
version: '0'
|
151
149
|
requirements: []
|
152
150
|
rubyforge_project:
|
153
|
-
rubygems_version:
|
151
|
+
rubygems_version: 2.4.5
|
154
152
|
signing_key:
|
155
|
-
specification_version:
|
153
|
+
specification_version: 4
|
156
154
|
summary: Utilities for running bots with Turbot
|
157
155
|
test_files: []
|