turbot-runner 0.1.9 → 0.1.10
Sign up to get free protection for your applications and to get access to all the features.
data/lib/turbot_runner/runner.rb
CHANGED
@@ -16,24 +16,20 @@ module TurbotRunner
|
|
16
16
|
|
17
17
|
def run
|
18
18
|
FileUtils.mkdir_p(@output_directory)
|
19
|
-
|
20
|
-
|
21
|
-
|
19
|
+
succeeded = run_script(scraper_config)
|
20
|
+
# Run the transformers even if the scraper fails
|
22
21
|
transformers.each do |transformer_config|
|
23
|
-
|
22
|
+
succeeded = run_script(transformer_config, input_file=scraper_output_file) && succeeded
|
24
23
|
end
|
25
|
-
|
26
|
-
true
|
24
|
+
succeeded
|
27
25
|
end
|
28
26
|
|
29
27
|
def process_output
|
30
|
-
|
28
|
+
process_script_output(scraper_config)
|
31
29
|
|
32
30
|
transformers.each do |transformer_config|
|
33
|
-
|
31
|
+
process_script_output(transformer_config)
|
34
32
|
end
|
35
|
-
|
36
|
-
true
|
37
33
|
end
|
38
34
|
|
39
35
|
private
|
@@ -53,7 +49,7 @@ module TurbotRunner
|
|
53
49
|
def run_script(script_config, input_file=nil)
|
54
50
|
command = build_command(script_config[:file], input_file)
|
55
51
|
|
56
|
-
|
52
|
+
script_runner = ScriptRunner.new(
|
57
53
|
command,
|
58
54
|
output_file(script_config[:file]),
|
59
55
|
script_config,
|
@@ -61,7 +57,7 @@ module TurbotRunner
|
|
61
57
|
:timeout => @timeout
|
62
58
|
)
|
63
59
|
|
64
|
-
|
60
|
+
script_runner.run # returns boolean indicating success
|
65
61
|
end
|
66
62
|
|
67
63
|
def process_script_output(script_config)
|
@@ -4,5 +4,12 @@
|
|
4
4
|
"language": "ruby",
|
5
5
|
"data_type": "primary data",
|
6
6
|
"identifying_fields": ["licence_number"],
|
7
|
-
"files": ["scraper.rb"]
|
7
|
+
"files": ["scraper.rb", "transformer1.rb"],
|
8
|
+
"transformers": [
|
9
|
+
{
|
10
|
+
"file": "transformer1.rb",
|
11
|
+
"data_type": "simple-licence",
|
12
|
+
"identifying_fields": ["licence_number"]
|
13
|
+
}
|
14
|
+
]
|
8
15
|
}
|
@@ -0,0 +1,15 @@
|
|
1
|
+
require 'json'
|
2
|
+
|
3
|
+
STDIN.each_line do |line|
|
4
|
+
raw_record = JSON.parse(line)
|
5
|
+
|
6
|
+
transformed_record = {
|
7
|
+
:company_name => 'Foo Widgets',
|
8
|
+
:company_jurisdiction => 'gb',
|
9
|
+
:licence_number => raw_record['licence_number'],
|
10
|
+
:source_url => raw_record['source_url'],
|
11
|
+
:sample_date => raw_record['sample_date'],
|
12
|
+
}
|
13
|
+
|
14
|
+
puts transformed_record.to_json
|
15
|
+
end
|
data/spec/lib/runner_spec.rb
CHANGED
@@ -126,6 +126,14 @@ describe TurbotRunner::Runner do
|
|
126
126
|
@runner.run
|
127
127
|
expect([@runner, 'scraper']).to have_error_output_matching(/Oh no/)
|
128
128
|
end
|
129
|
+
|
130
|
+
it 'still runs the transformers' do
|
131
|
+
expect(@runner).to receive(:run_script).once.with(
|
132
|
+
hash_including(:file=>"scraper.rb"))
|
133
|
+
expect(@runner).to receive(:run_script).once.with(
|
134
|
+
hash_including(:file=>"transformer1.rb"), anything)
|
135
|
+
@runner.run
|
136
|
+
end
|
129
137
|
end
|
130
138
|
|
131
139
|
context 'with a bot that crashes in transformer' do
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: turbot-runner
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.1.
|
4
|
+
version: 0.1.10
|
5
5
|
prerelease:
|
6
6
|
platform: ruby
|
7
7
|
authors:
|
@@ -9,7 +9,7 @@ authors:
|
|
9
9
|
autorequire:
|
10
10
|
bindir: bin
|
11
11
|
cert_chain: []
|
12
|
-
date: 2014-11-
|
12
|
+
date: 2014-11-19 00:00:00.000000000 Z
|
13
13
|
dependencies:
|
14
14
|
- !ruby/object:Gem::Dependency
|
15
15
|
name: json-schema
|
@@ -44,6 +44,7 @@ files:
|
|
44
44
|
- lib/turbot_runner/version.rb
|
45
45
|
- spec/bots/bot-that-crashes-in-scraper/manifest.json
|
46
46
|
- spec/bots/bot-that-crashes-in-scraper/scraper.rb
|
47
|
+
- spec/bots/bot-that-crashes-in-scraper/transformer1.rb
|
47
48
|
- spec/bots/bot-that-crashes-in-transformer/manifest.json
|
48
49
|
- spec/bots/bot-that-crashes-in-transformer/scraper.rb
|
49
50
|
- spec/bots/bot-that-crashes-in-transformer/transformer1.rb
|