indico 0.10.3 → 0.10.4
Sign up to get free protection for your applications and to get access to all the features.
- data/lib/indico.rb +5 -0
- data/lib/indico/pdf.rb +26 -0
- data/lib/indico/version.rb +1 -1
- data/spec/data/test.pdf +0 -0
- data/spec/pdf_extraction_spec.rb +43 -0
- metadata +5 -2
data/lib/indico.rb
CHANGED
@@ -1,6 +1,7 @@
|
|
1
1
|
require 'indico/version'
|
2
2
|
require 'indico/helper'
|
3
3
|
require 'indico/image'
|
4
|
+
require 'indico/pdf'
|
4
5
|
require 'indico/multi'
|
5
6
|
require 'indico/settings'
|
6
7
|
require 'indico/errors'
|
@@ -127,6 +128,10 @@ module Indico
|
|
127
128
|
api_handler(text, "summarization", config)
|
128
129
|
end
|
129
130
|
|
131
|
+
def self.pdf_extraction(pdf, config = {})
|
132
|
+
api_handler(preprocess_pdf(pdf), "pdfextraction", config)
|
133
|
+
end
|
134
|
+
|
130
135
|
def self.fer(image, config = nil)
|
131
136
|
size = (config != nil and config["detect"] == true) ? false : 48
|
132
137
|
api_handler(preprocess(image, size, false), 'fer', config)
|
data/lib/indico/pdf.rb
ADDED
@@ -0,0 +1,26 @@
|
|
1
|
+
require 'base64'
|
2
|
+
|
3
|
+
module Indico
|
4
|
+
def self.preprocess_pdf(pdf)
|
5
|
+
if pdf.class == Array
|
6
|
+
# Batch Request
|
7
|
+
pdf_array = Array.new
|
8
|
+
|
9
|
+
# process each image
|
10
|
+
pdf.each do |_pdf|
|
11
|
+
pdf_array.push(preprocess_pdf(_pdf))
|
12
|
+
end
|
13
|
+
|
14
|
+
return pdf_array
|
15
|
+
elsif pdf.class != String
|
16
|
+
raise Exception.new("PDF input must be filename, url or base64 string")
|
17
|
+
end
|
18
|
+
|
19
|
+
begin
|
20
|
+
return Base64.encode64(File.read(pdf))
|
21
|
+
rescue
|
22
|
+
# likely a url or a base64 encoded string already
|
23
|
+
return pdf
|
24
|
+
end
|
25
|
+
end
|
26
|
+
end
|
data/lib/indico/version.rb
CHANGED
data/spec/data/test.pdf
ADDED
Binary file
|
@@ -0,0 +1,43 @@
|
|
1
|
+
require 'spec_helper'
|
2
|
+
|
3
|
+
describe Indico do
|
4
|
+
before do
|
5
|
+
api_key = ENV['INDICO_API_KEY']
|
6
|
+
private_cloud = 'indico-test'
|
7
|
+
@config = { api_key: api_key, cloud: private_cloud}
|
8
|
+
end
|
9
|
+
|
10
|
+
it 'should properly handle pdf urls' do
|
11
|
+
pdf_url = "https://s3-us-west-2.amazonaws.com/indico-test-data/test.pdf"
|
12
|
+
|
13
|
+
response = Indico.pdf_extraction(pdf_url, @config)
|
14
|
+
expected_keys = Set.new(%w(metadata text))
|
15
|
+
|
16
|
+
expect(Set.new(response.keys)).to eql(expected_keys)
|
17
|
+
end
|
18
|
+
|
19
|
+
it 'should properly handle local pdf files' do
|
20
|
+
pdf_path = File.expand_path(
|
21
|
+
File.join(File.dirname(__FILE__), "data", "test.pdf")
|
22
|
+
)
|
23
|
+
|
24
|
+
response = Indico.pdf_extraction(pdf_path, @config)
|
25
|
+
expected_keys = Set.new(%w(metadata text))
|
26
|
+
|
27
|
+
expect(Set.new(response.keys)).to eql(expected_keys)
|
28
|
+
end
|
29
|
+
|
30
|
+
it 'should properly handle array of local pdf files' do
|
31
|
+
pdf_path = File.expand_path(
|
32
|
+
File.join(File.dirname(__FILE__), "data", "test.pdf")
|
33
|
+
)
|
34
|
+
arr = Array.new()
|
35
|
+
arr.push(pdf_path)
|
36
|
+
arr.push(pdf_path)
|
37
|
+
|
38
|
+
response = Indico.pdf_extraction(arr, @config)
|
39
|
+
expected_keys = Set.new(%w(metadata text))
|
40
|
+
|
41
|
+
expect(Set.new(response[0].keys)).to eql(expected_keys)
|
42
|
+
end
|
43
|
+
end
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: indico
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.10.
|
4
|
+
version: 0.10.4
|
5
5
|
prerelease:
|
6
6
|
platform: ruby
|
7
7
|
authors:
|
@@ -12,7 +12,7 @@ authors:
|
|
12
12
|
autorequire:
|
13
13
|
bindir: bin
|
14
14
|
cert_chain: []
|
15
|
-
date: 2016-
|
15
|
+
date: 2016-12-16 00:00:00.000000000 Z
|
16
16
|
dependencies:
|
17
17
|
- !ruby/object:Gem::Dependency
|
18
18
|
name: inifile
|
@@ -117,6 +117,7 @@ files:
|
|
117
117
|
- lib/indico/helper.rb
|
118
118
|
- lib/indico/image.rb
|
119
119
|
- lib/indico/multi.rb
|
120
|
+
- lib/indico/pdf.rb
|
120
121
|
- lib/indico/settings.rb
|
121
122
|
- lib/indico/version.rb
|
122
123
|
- spec/config/.indicorc.test
|
@@ -125,10 +126,12 @@ files:
|
|
125
126
|
- spec/data/dog.jpg
|
126
127
|
- spec/data/happy.png
|
127
128
|
- spec/data/happy64.txt
|
129
|
+
- spec/data/test.pdf
|
128
130
|
- spec/imagerecognition_spec.rb
|
129
131
|
- spec/indico_batch_spec.rb
|
130
132
|
- spec/indico_spec.rb
|
131
133
|
- spec/keywords_v2_spec.rb
|
134
|
+
- spec/pdf_extraction_spec.rb
|
132
135
|
- spec/political_spec.rb
|
133
136
|
- spec/settings_spec.rb
|
134
137
|
- spec/spec_helper.rb
|