1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
|
# frozen_string_literal: true
require 'spec_helper'
RSpec.describe BulkImports::Groups::Pipelines::LabelsPipeline do
let(:user) { create(:user) }
let(:group) { create(:group) }
let(:cursor) { 'cursor' }
let(:entity) do
create(
:bulk_import_entity,
source_full_path: 'source/full/path',
destination_name: 'My Destination Group',
destination_namespace: group.full_path,
group: group
)
end
let(:context) { BulkImports::Pipeline::Context.new(entity) }
subject { described_class.new(context) }
def extractor_data(title:, has_next_page:, cursor: nil)
data = [
{
'title' => title,
'description' => 'desc',
'color' => '#428BCA'
}
]
page_info = {
'end_cursor' => cursor,
'has_next_page' => has_next_page
}
BulkImports::Pipeline::ExtractedData.new(data: data, page_info: page_info)
end
describe '#run' do
it 'imports a group labels' do
first_page = extractor_data(title: 'label1', has_next_page: true, cursor: cursor)
last_page = extractor_data(title: 'label2', has_next_page: false)
allow_next_instance_of(BulkImports::Common::Extractors::GraphqlExtractor) do |extractor|
allow(extractor)
.to receive(:extract)
.and_return(first_page, last_page)
end
expect { subject.run }.to change(Label, :count).by(2)
label = group.labels.order(:created_at).last
expect(label.title).to eq('label2')
expect(label.description).to eq('desc')
expect(label.color).to eq('#428BCA')
end
end
describe '#after_run' do
context 'when extracted data has next page' do
it 'updates tracker information and runs pipeline again' do
data = extractor_data(title: 'label', has_next_page: true, cursor: cursor)
expect(subject).to receive(:run)
subject.after_run(data)
tracker = entity.trackers.find_by(relation: :labels)
expect(tracker.has_next_page).to eq(true)
expect(tracker.next_page).to eq(cursor)
end
end
context 'when extracted data has no next page' do
it 'updates tracker information and does not run pipeline' do
data = extractor_data(title: 'label', has_next_page: false)
expect(subject).not_to receive(:run)
subject.after_run(data)
tracker = entity.trackers.find_by(relation: :labels)
expect(tracker.has_next_page).to eq(false)
expect(tracker.next_page).to be_nil
end
end
end
describe 'pipeline parts' do
it { expect(described_class).to include_module(BulkImports::Pipeline) }
it { expect(described_class).to include_module(BulkImports::Pipeline::Runner) }
it 'has extractors' do
expect(described_class.get_extractor)
.to eq(
klass: BulkImports::Common::Extractors::GraphqlExtractor,
options: {
query: BulkImports::Groups::Graphql::GetLabelsQuery
}
)
end
it 'has transformers' do
expect(described_class.transformers)
.to contain_exactly(
{ klass: BulkImports::Common::Transformers::ProhibitedAttributesTransformer, options: nil }
)
end
it 'has loaders' do
expect(described_class.get_loader).to eq(klass: BulkImports::Groups::Loaders::LabelsLoader, options: nil)
end
end
end
|