mindee 1.2.0 → 2.0.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.gitignore +1 -1
- data/.rubocop.yml +2 -2
- data/.yardopts +4 -0
- data/CHANGELOG.md +21 -0
- data/README.md +46 -23
- data/Rakefile +6 -1
- data/bin/mindee.rb +70 -61
- data/docs/ruby-api-builder.md +131 -0
- data/docs/ruby-getting-started.md +265 -0
- data/docs/ruby-invoice-ocr.md +261 -0
- data/docs/ruby-passport-ocr.md +156 -0
- data/docs/ruby-receipt-ocr.md +170 -0
- data/lib/mindee/client.rb +128 -93
- data/lib/mindee/document_config.rb +22 -154
- data/lib/mindee/geometry.rb +105 -8
- data/lib/mindee/http/endpoint.rb +80 -0
- data/lib/mindee/input/pdf_processing.rb +106 -0
- data/lib/mindee/input/sources.rb +97 -0
- data/lib/mindee/input.rb +3 -0
- data/lib/mindee/parsing/document.rb +31 -0
- data/lib/mindee/parsing/error.rb +22 -0
- data/lib/mindee/parsing/inference.rb +53 -0
- data/lib/mindee/parsing/page.rb +46 -0
- data/lib/mindee/parsing/prediction/base.rb +30 -0
- data/lib/mindee/{fields → parsing/prediction/common_fields}/amount.rb +5 -1
- data/lib/mindee/{fields → parsing/prediction/common_fields}/base.rb +16 -5
- data/lib/mindee/{fields → parsing/prediction/common_fields}/company_registration.rb +0 -0
- data/lib/mindee/{fields/datefield.rb → parsing/prediction/common_fields/date.rb} +0 -0
- data/lib/mindee/{fields → parsing/prediction/common_fields}/locale.rb +0 -0
- data/lib/mindee/{fields → parsing/prediction/common_fields}/payment_details.rb +0 -0
- data/lib/mindee/parsing/prediction/common_fields/position.rb +39 -0
- data/lib/mindee/{fields → parsing/prediction/common_fields}/tax.rb +7 -2
- data/lib/mindee/parsing/prediction/common_fields/text.rb +12 -0
- data/lib/mindee/parsing/prediction/common_fields.rb +11 -0
- data/lib/mindee/parsing/prediction/custom/custom_v1.rb +58 -0
- data/lib/mindee/{fields/custom_docs.rb → parsing/prediction/custom/fields.rb} +5 -5
- data/lib/mindee/parsing/prediction/eu/license_plate/license_plate_v1.rb +34 -0
- data/lib/mindee/parsing/prediction/fr/bank_account_details/bank_account_details_v1.rb +40 -0
- data/lib/mindee/parsing/prediction/fr/carte_vitale/carte_vitale_v1.rb +49 -0
- data/lib/mindee/parsing/prediction/fr/id_card/id_card_v1.rb +84 -0
- data/lib/mindee/parsing/prediction/invoice/invoice_line_item.rb +58 -0
- data/lib/mindee/parsing/prediction/invoice/invoice_v4.rb +216 -0
- data/lib/mindee/parsing/prediction/passport/passport_v1.rb +184 -0
- data/lib/mindee/parsing/prediction/receipt/receipt_v4.rb +84 -0
- data/lib/mindee/parsing/prediction/shipping_container/shipping_container_v1.rb +38 -0
- data/lib/mindee/parsing/prediction/us/bank_check/bank_check_v1.rb +70 -0
- data/lib/mindee/parsing/prediction.rb +12 -0
- data/lib/mindee/parsing.rb +4 -0
- data/lib/mindee/version.rb +1 -1
- data/mindee.gemspec +2 -1
- metadata +57 -24
- data/lib/mindee/documents/base.rb +0 -35
- data/lib/mindee/documents/custom.rb +0 -65
- data/lib/mindee/documents/financial_doc.rb +0 -135
- data/lib/mindee/documents/invoice.rb +0 -162
- data/lib/mindee/documents/passport.rb +0 -163
- data/lib/mindee/documents/receipt.rb +0 -109
- data/lib/mindee/documents.rb +0 -7
- data/lib/mindee/endpoint.rb +0 -105
- data/lib/mindee/fields/orientation.rb +0 -26
- data/lib/mindee/fields.rb +0 -11
- data/lib/mindee/inputs.rb +0 -153
- data/lib/mindee/response.rb +0 -27
data/lib/mindee/geometry.rb
CHANGED
@@ -3,19 +3,116 @@
|
|
3
3
|
module Mindee
|
4
4
|
# Various helper functions for geometry.
|
5
5
|
module Geometry
|
6
|
+
# A relative set of coordinates (X, Y) on the document.
|
7
|
+
class Point
|
8
|
+
# @return [Float]
|
9
|
+
attr_accessor :x
|
10
|
+
# @return [Float]
|
11
|
+
attr_accessor :y
|
12
|
+
|
13
|
+
# @param x [Float]
|
14
|
+
# @param y [Float]
|
15
|
+
# rubocop:disable Naming/MethodParameterName
|
16
|
+
def initialize(x, y)
|
17
|
+
@x = x
|
18
|
+
@y = y
|
19
|
+
end
|
20
|
+
# rubocop:enable Naming/MethodParameterName
|
21
|
+
|
22
|
+
# @return [Float]
|
23
|
+
def [](key)
|
24
|
+
case key
|
25
|
+
when 0
|
26
|
+
@x
|
27
|
+
when 1
|
28
|
+
@y
|
29
|
+
else
|
30
|
+
throw '0 or 1 only'
|
31
|
+
end
|
32
|
+
end
|
33
|
+
end
|
34
|
+
|
35
|
+
# Contains exactly 4 relative vertices coordinates (Points).
|
36
|
+
class Quadrilateral
|
37
|
+
# @return [Mindee::Geometry::Point]
|
38
|
+
attr_accessor :top_left
|
39
|
+
# @return [Mindee::Geometry::Point]
|
40
|
+
attr_accessor :top_right
|
41
|
+
# @return [Mindee::Geometry::Point]
|
42
|
+
attr_accessor :bottom_right
|
43
|
+
# @return [Mindee::Geometry::Point]
|
44
|
+
attr_accessor :bottom_left
|
45
|
+
|
46
|
+
# @param top_left [Mindee::Geometry::Point]
|
47
|
+
# @param top_right [Mindee::Geometry::Point]
|
48
|
+
# @param bottom_right [Mindee::Geometry::Point]
|
49
|
+
# @param bottom_left [Mindee::Geometry::Point]
|
50
|
+
def initialize(top_left, top_right, bottom_right, bottom_left)
|
51
|
+
@top_left = top_left
|
52
|
+
@top_right = top_right
|
53
|
+
@bottom_right = bottom_right
|
54
|
+
@bottom_left = bottom_left
|
55
|
+
end
|
56
|
+
|
57
|
+
# @return [Mindee::Geometry::Point]
|
58
|
+
def [](key)
|
59
|
+
case key
|
60
|
+
when 0
|
61
|
+
@top_left
|
62
|
+
when 1
|
63
|
+
@top_right
|
64
|
+
when 2
|
65
|
+
@bottom_right
|
66
|
+
when 3
|
67
|
+
@bottom_left
|
68
|
+
else
|
69
|
+
throw '0, 1, 2, 3 only'
|
70
|
+
end
|
71
|
+
end
|
72
|
+
end
|
73
|
+
|
74
|
+
class Polygon < Array
|
75
|
+
end
|
76
|
+
|
77
|
+
# Transform a prediction into a Quadrilateral.
|
78
|
+
def self.quadrilateral_from_prediction(prediction)
|
79
|
+
throw "Prediction must have exactly 4 points, found #{prediction.size}" if prediction.size != 4
|
80
|
+
|
81
|
+
Quadrilateral.new(
|
82
|
+
Point.new(prediction[0][0], prediction[0][1]),
|
83
|
+
Point.new(prediction[1][0], prediction[1][1]),
|
84
|
+
Point.new(prediction[2][0], prediction[2][1]),
|
85
|
+
Point.new(prediction[3][0], prediction[3][1])
|
86
|
+
)
|
87
|
+
end
|
88
|
+
|
89
|
+
# Transform a prediction into a Polygon.
|
90
|
+
def self.polygon_from_prediction(prediction)
|
91
|
+
polygon = Polygon.new
|
92
|
+
return polygon if prediction.nil?
|
93
|
+
|
94
|
+
prediction.each do |point|
|
95
|
+
polygon << Point.new(point[0], point[1])
|
96
|
+
end
|
97
|
+
polygon
|
98
|
+
end
|
99
|
+
|
6
100
|
# @return [Array<Float>]
|
7
101
|
def self.get_bbox(vertices)
|
8
|
-
|
9
|
-
|
10
|
-
|
11
|
-
y_max = vertices.map { |v| v[1] }.max
|
12
|
-
[x_min, y_min, x_max, y_max]
|
102
|
+
x_coords = vertices.map(&:x)
|
103
|
+
y_coords = vertices.map(&:y)
|
104
|
+
[x_coords.min, y_coords.min, x_coords.max, y_coords.max]
|
13
105
|
end
|
14
106
|
|
15
|
-
# @return [
|
16
|
-
def self.
|
107
|
+
# @return [Mindee::Geometry::Quadrilateral]
|
108
|
+
def self.get_bounding_box(vertices)
|
17
109
|
x_min, y_min, x_max, y_max = get_bbox(vertices)
|
18
|
-
|
110
|
+
Quadrilateral.new(
|
111
|
+
Point.new(x_min, y_min),
|
112
|
+
Point.new(x_max, y_min),
|
113
|
+
Point.new(x_max, y_max),
|
114
|
+
Point.new(x_min, y_max)
|
115
|
+
)
|
19
116
|
end
|
20
117
|
end
|
21
118
|
end
|
@@ -0,0 +1,80 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require 'net/http'
|
4
|
+
require_relative '../version'
|
5
|
+
|
6
|
+
module Mindee
|
7
|
+
module HTTP
|
8
|
+
API_KEY_ENV_NAME = 'MINDEE_API_KEY'
|
9
|
+
API_KEY_DEFAULT = nil
|
10
|
+
|
11
|
+
BASE_URL_ENV_NAME = 'MINDEE_BASE_URL'
|
12
|
+
BASE_URL_DEFAULT = 'https://api.mindee.net/v1'
|
13
|
+
|
14
|
+
REQUEST_TIMEOUT_ENV_NAME = 'MINDEE_REQUEST_TIMEOUT'
|
15
|
+
TIMEOUT_DEFAULT = 120
|
16
|
+
|
17
|
+
USER_AGENT = "mindee-api-ruby@v#{Mindee::VERSION} ruby-v#{RUBY_VERSION} #{Mindee::PLATFORM}"
|
18
|
+
|
19
|
+
# Generic API endpoint for a product.
|
20
|
+
class Endpoint
|
21
|
+
# @return [String]
|
22
|
+
attr_reader :api_key
|
23
|
+
# @return [Integer]
|
24
|
+
attr_reader :request_timeout
|
25
|
+
|
26
|
+
def initialize(owner, url_name, version, api_key: '')
|
27
|
+
@owner = owner
|
28
|
+
@url_name = url_name
|
29
|
+
@version = version
|
30
|
+
@request_timeout = ENV.fetch(REQUEST_TIMEOUT_ENV_NAME, TIMEOUT_DEFAULT).to_i
|
31
|
+
@api_key = api_key.nil? || api_key.empty? ? ENV.fetch(API_KEY_ENV_NAME, API_KEY_DEFAULT) : api_key
|
32
|
+
@url_root = "#{BASE_URL_DEFAULT}/products/#{@owner}/#{@url_name}/v#{@version}"
|
33
|
+
end
|
34
|
+
|
35
|
+
# @param input_doc [Mindee::InputDocument]
|
36
|
+
# @param include_words [Boolean]
|
37
|
+
# @param close_file [Boolean]
|
38
|
+
# @param cropper [Boolean]
|
39
|
+
# @return [Net::HTTPResponse]
|
40
|
+
def predict_req_post(input_doc, include_words: false, close_file: true, cropper: false)
|
41
|
+
uri = URI("#{@url_root}/predict")
|
42
|
+
|
43
|
+
params = {}
|
44
|
+
params[:cropper] = 'true' if cropper
|
45
|
+
uri.query = URI.encode_www_form(params)
|
46
|
+
|
47
|
+
headers = {
|
48
|
+
'Authorization' => "Token #{@api_key}",
|
49
|
+
'User-Agent' => USER_AGENT,
|
50
|
+
}
|
51
|
+
req = Net::HTTP::Post.new(uri, headers)
|
52
|
+
|
53
|
+
form_data = {
|
54
|
+
'document' => input_doc.read_document(close: close_file),
|
55
|
+
}
|
56
|
+
form_data.push ['include_mvision', 'true'] if include_words
|
57
|
+
|
58
|
+
req.set_form(form_data, 'multipart/form-data')
|
59
|
+
|
60
|
+
Net::HTTP.start(uri.hostname, uri.port, use_ssl: true, read_timeout: @request_timeout) do |http|
|
61
|
+
http.request(req)
|
62
|
+
end
|
63
|
+
end
|
64
|
+
end
|
65
|
+
|
66
|
+
# Receipt API endpoint
|
67
|
+
class StandardEndpoint < Endpoint
|
68
|
+
def initialize(endpoint_name, version, api_key)
|
69
|
+
super('mindee', endpoint_name, version, api_key: api_key)
|
70
|
+
end
|
71
|
+
end
|
72
|
+
|
73
|
+
# Custom (constructed) API endpoint
|
74
|
+
class CustomEndpoint < Endpoint
|
75
|
+
def initialize(account_name, endpoint_name, version, api_key)
|
76
|
+
super(account_name, endpoint_name, version, api_key: api_key)
|
77
|
+
end
|
78
|
+
end
|
79
|
+
end
|
80
|
+
end
|
@@ -0,0 +1,106 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require 'set'
|
4
|
+
require 'origami'
|
5
|
+
|
6
|
+
# Monkey-patching for Origami
|
7
|
+
module PDFTools
|
8
|
+
def to_io_stream(params = {})
|
9
|
+
options = {
|
10
|
+
delinearize: true,
|
11
|
+
recompile: true,
|
12
|
+
decrypt: false,
|
13
|
+
}
|
14
|
+
options.update(params)
|
15
|
+
|
16
|
+
if frozen? # incompatible flags with frozen doc (signed)
|
17
|
+
options[:recompile] = nil
|
18
|
+
options[:rebuild_xrefs] = nil
|
19
|
+
options[:noindent] = nil
|
20
|
+
options[:obfuscate] = false
|
21
|
+
end
|
22
|
+
load_all_objects unless @loaded
|
23
|
+
|
24
|
+
intents_as_pdfa1 if options[:intent] =~ %r{pdf[/-]?A1?/i}
|
25
|
+
delinearize! if options[:delinearize] && linearized?
|
26
|
+
compile(options) if options[:recompile]
|
27
|
+
|
28
|
+
io_stream = StringIO.new(output(options))
|
29
|
+
io_stream.set_encoding Encoding::BINARY
|
30
|
+
io_stream
|
31
|
+
end
|
32
|
+
end
|
33
|
+
|
34
|
+
Origami::PDF.class_eval { include PDFTools }
|
35
|
+
|
36
|
+
module Mindee
|
37
|
+
module Input
|
38
|
+
# Class for PDF documents
|
39
|
+
module PdfProcessor
|
40
|
+
DEFAULT_OPTIONS = {
|
41
|
+
page_indexes: [0],
|
42
|
+
operation: :KEEP_ONLY,
|
43
|
+
on_min_pages: 0,
|
44
|
+
}.freeze
|
45
|
+
|
46
|
+
# @param io_stream [StreamIO]
|
47
|
+
# @param options [Hash]
|
48
|
+
def self.parse(io_stream, options)
|
49
|
+
options = DEFAULT_OPTIONS.merge(options)
|
50
|
+
|
51
|
+
current_pdf = open_pdf(io_stream)
|
52
|
+
pages_count = current_pdf.pages.size
|
53
|
+
return if options[:on_min_pages] > pages_count
|
54
|
+
|
55
|
+
all_pages = (0..pages_count - 1).to_a
|
56
|
+
|
57
|
+
case options[:operation]
|
58
|
+
when :KEEP_ONLY
|
59
|
+
pages_to_remove = indexes_from_keep(options[:page_indexes], all_pages)
|
60
|
+
when :REMOVE
|
61
|
+
pages_to_remove = indexes_from_remove(options[:page_indexes], all_pages)
|
62
|
+
else
|
63
|
+
raise "operation must be one of :KEEP_ONLY or :REMOVE, sent '#{behavior}'"
|
64
|
+
end
|
65
|
+
|
66
|
+
current_pdf.delete_pages_at(pages_to_remove) if pages_to_remove.to_a != all_pages.to_a
|
67
|
+
current_pdf.to_io_stream
|
68
|
+
end
|
69
|
+
|
70
|
+
# @param page_indexes [Array]
|
71
|
+
# @param all_pages [Array]
|
72
|
+
def self.indexes_from_keep(page_indexes, all_pages)
|
73
|
+
pages_to_keep = Set.new
|
74
|
+
page_indexes.each do |idx|
|
75
|
+
idx = (all_pages.length - (idx + 2)) if idx.negative?
|
76
|
+
page = all_pages[idx]
|
77
|
+
next if page.nil?
|
78
|
+
|
79
|
+
pages_to_keep << page
|
80
|
+
end
|
81
|
+
all_pages.to_set - pages_to_keep
|
82
|
+
end
|
83
|
+
|
84
|
+
# @param page_indexes [Array]
|
85
|
+
# @param all_pages [Array]
|
86
|
+
def self.indexes_from_remove(page_indexes, all_pages)
|
87
|
+
pages_to_remove = Set.new
|
88
|
+
page_indexes.each do |idx|
|
89
|
+
idx = (all_pages.length - (idx + 2)) if idx.negative?
|
90
|
+
page = all_pages[idx]
|
91
|
+
next if page.nil?
|
92
|
+
|
93
|
+
pages_to_remove << page
|
94
|
+
end
|
95
|
+
end
|
96
|
+
|
97
|
+
# @param io_stream [StringIO]
|
98
|
+
# @return [Origami::PDF]
|
99
|
+
def self.open_pdf(io_stream)
|
100
|
+
pdf_parser = Origami::PDF::LinearParser.new({ verbosity: Origami::Parser::VERBOSE_QUIET })
|
101
|
+
io_stream.seek(0)
|
102
|
+
pdf_parser.parse(io_stream)
|
103
|
+
end
|
104
|
+
end
|
105
|
+
end
|
106
|
+
end
|
@@ -0,0 +1,97 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require 'stringio'
|
4
|
+
require 'marcel'
|
5
|
+
|
6
|
+
require_relative 'pdf_processing'
|
7
|
+
|
8
|
+
module Mindee
|
9
|
+
module Input
|
10
|
+
ALLOWED_MIME_TYPES = [
|
11
|
+
'application/pdf',
|
12
|
+
'image/heic',
|
13
|
+
'image/png',
|
14
|
+
'image/jpeg',
|
15
|
+
'image/tiff',
|
16
|
+
'image/webp',
|
17
|
+
].freeze
|
18
|
+
|
19
|
+
# Base class for loading documents.
|
20
|
+
class InputDocument
|
21
|
+
# @return [String]
|
22
|
+
attr_reader :filename
|
23
|
+
# @return [String]
|
24
|
+
attr_reader :file_mimetype
|
25
|
+
# @return [StreamIO]
|
26
|
+
attr_reader :io_stream
|
27
|
+
|
28
|
+
# @param io_stream [StreamIO]
|
29
|
+
def initialize(io_stream, filename)
|
30
|
+
@io_stream = io_stream
|
31
|
+
@filename = filename
|
32
|
+
@file_mimetype = Marcel::MimeType.for @io_stream, name: @filename
|
33
|
+
|
34
|
+
return if ALLOWED_MIME_TYPES.include? @file_mimetype
|
35
|
+
|
36
|
+
raise "File type not allowed, must be one of #{ALLOWED_MIME_TYPES.join(', ')}"
|
37
|
+
end
|
38
|
+
|
39
|
+
def pdf?
|
40
|
+
@file_mimetype == 'application/pdf'
|
41
|
+
end
|
42
|
+
|
43
|
+
def process_pdf(options)
|
44
|
+
@io_stream.seek(0)
|
45
|
+
@io_stream = PdfProcessor.parse(@io_stream, options)
|
46
|
+
end
|
47
|
+
|
48
|
+
# @param close [Boolean]
|
49
|
+
def read_document(close: true)
|
50
|
+
@io_stream.seek(0)
|
51
|
+
data = @io_stream.read
|
52
|
+
@io_stream.close if close
|
53
|
+
[data].pack('m')
|
54
|
+
end
|
55
|
+
end
|
56
|
+
|
57
|
+
# Load a document from a path.
|
58
|
+
class PathDocument < InputDocument
|
59
|
+
# @param filepath [String]
|
60
|
+
def initialize(filepath)
|
61
|
+
io_stream = File.open(filepath, 'rb')
|
62
|
+
super(io_stream, File.basename(filepath))
|
63
|
+
end
|
64
|
+
end
|
65
|
+
|
66
|
+
# Load a document from a base64 string.
|
67
|
+
class Base64Document < InputDocument
|
68
|
+
# @param base64_string [String]
|
69
|
+
# @param filename [String]
|
70
|
+
def initialize(base64_string, filename)
|
71
|
+
io_stream = StringIO.new(base64_string.unpack1('m*'))
|
72
|
+
io_stream.set_encoding Encoding::BINARY
|
73
|
+
super(io_stream, filename)
|
74
|
+
end
|
75
|
+
end
|
76
|
+
|
77
|
+
# Load a document from raw bytes.
|
78
|
+
class BytesDocument < InputDocument
|
79
|
+
# @param raw_bytes [String]
|
80
|
+
# @param filename [String]
|
81
|
+
def initialize(raw_bytes, filename)
|
82
|
+
io_stream = StringIO.new(raw_bytes)
|
83
|
+
io_stream.set_encoding Encoding::BINARY
|
84
|
+
super(io_stream, filename)
|
85
|
+
end
|
86
|
+
end
|
87
|
+
|
88
|
+
# Load a document from a file handle.
|
89
|
+
class FileDocument < InputDocument
|
90
|
+
# @param filename [String]
|
91
|
+
def initialize(file_handle, filename)
|
92
|
+
io_stream = file_handle
|
93
|
+
super(io_stream, filename)
|
94
|
+
end
|
95
|
+
end
|
96
|
+
end
|
97
|
+
end
|
data/lib/mindee/input.rb
ADDED
@@ -0,0 +1,31 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require_relative 'inference'
|
4
|
+
|
5
|
+
module Mindee
|
6
|
+
# Stores all response attributes.
|
7
|
+
class Document
|
8
|
+
# @return [Mindee::Inference]
|
9
|
+
attr_reader :inference
|
10
|
+
# @return [String] Filename sent to the API
|
11
|
+
attr_reader :name
|
12
|
+
# @return [String] Mindee ID of the document
|
13
|
+
attr_reader :id
|
14
|
+
|
15
|
+
# @param prediction_class [Class<Mindee::Prediction::Prediction>]
|
16
|
+
# @param http_response [Hash]
|
17
|
+
def initialize(prediction_class, http_response)
|
18
|
+
@id = http_response['id']
|
19
|
+
@name = http_response['name']
|
20
|
+
@inference = Mindee::Inference.new(prediction_class, http_response['inference'])
|
21
|
+
end
|
22
|
+
|
23
|
+
def to_s
|
24
|
+
out_str = String.new
|
25
|
+
out_str << "########\nDocument\n########"
|
26
|
+
out_str << "\n:Mindee ID: #{@id}"
|
27
|
+
out_str << "\n:Filename: #{@name}"
|
28
|
+
out_str << "\n\n#{@inference}"
|
29
|
+
end
|
30
|
+
end
|
31
|
+
end
|
@@ -0,0 +1,22 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Mindee
|
4
|
+
module Parsing
|
5
|
+
# API Error
|
6
|
+
class Error < StandardError
|
7
|
+
# @return [String]
|
8
|
+
attr_reader :api_code
|
9
|
+
# @return [String]
|
10
|
+
attr_reader :api_details
|
11
|
+
# @return [String]
|
12
|
+
attr_reader :api_message
|
13
|
+
|
14
|
+
def initialize(error)
|
15
|
+
@api_code = error['code']
|
16
|
+
@api_details = error['details']
|
17
|
+
@api_message = error['message']
|
18
|
+
super("#{@api_code}: #{@api_details} - #{@api_message}")
|
19
|
+
end
|
20
|
+
end
|
21
|
+
end
|
22
|
+
end
|
@@ -0,0 +1,53 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
require_relative 'page'
|
4
|
+
|
5
|
+
module Mindee
|
6
|
+
# Product information
|
7
|
+
class Product
|
8
|
+
attr_reader :name, :type, :version
|
9
|
+
|
10
|
+
# @param http_response [Hash]
|
11
|
+
def initialize(http_response)
|
12
|
+
@name = http_response['name']
|
13
|
+
@type = http_response['type']
|
14
|
+
@version = http_response['version']
|
15
|
+
end
|
16
|
+
end
|
17
|
+
|
18
|
+
# Inference holds all predictions
|
19
|
+
class Inference
|
20
|
+
# @return [Boolean]
|
21
|
+
attr_reader :is_rotation_applied
|
22
|
+
# @return [Array<Mindee::Page>]
|
23
|
+
attr_reader :pages
|
24
|
+
# @return [Mindee::Prediction]
|
25
|
+
attr_reader :prediction
|
26
|
+
# @return [Mindee::Product]
|
27
|
+
attr_reader :product
|
28
|
+
|
29
|
+
# @param prediction_class [Class<Mindee::Prediction::Prediction>]
|
30
|
+
# @param http_response [Hash]
|
31
|
+
def initialize(prediction_class, http_response)
|
32
|
+
@is_rotation_applied = http_response['is_rotation_applied']
|
33
|
+
@prediction = prediction_class.new(http_response['prediction'], nil)
|
34
|
+
@product = Product.new(http_response['product'])
|
35
|
+
@pages = []
|
36
|
+
http_response['pages'].each do |page|
|
37
|
+
@pages.push(Page.new(prediction_class, page))
|
38
|
+
end
|
39
|
+
end
|
40
|
+
|
41
|
+
def to_s
|
42
|
+
is_rotation_applied = @is_rotation_applied ? 'Yes' : 'No'
|
43
|
+
out_str = String.new
|
44
|
+
out_str << "Inference\n#########"
|
45
|
+
out_str << "\n:Product: #{@product.name} v#{@product.version}"
|
46
|
+
out_str << "\n:Rotation applied: #{is_rotation_applied}"
|
47
|
+
out_str << "\n\nPrediction\n=========="
|
48
|
+
out_str << "\n#{@prediction}"
|
49
|
+
out_str << "\n\nPage Predictions\n================\n\n"
|
50
|
+
out_str << @pages.map(&:to_s).join("\n\n")
|
51
|
+
end
|
52
|
+
end
|
53
|
+
end
|
@@ -0,0 +1,46 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Mindee
|
4
|
+
# Page orientation
|
5
|
+
class Orientation
|
6
|
+
# @return [Integer]
|
7
|
+
attr_reader :page_id
|
8
|
+
# A prediction among these 3 possible outputs:
|
9
|
+
# * 0 degrees: the page is already upright
|
10
|
+
# * 90 degrees: the page must be rotated clockwise to be upright
|
11
|
+
# * 270 degrees: the page must be rotated counterclockwise to be upright
|
12
|
+
# @return [Integer, nil]
|
13
|
+
attr_reader :value
|
14
|
+
|
15
|
+
# @param prediction [Hash]
|
16
|
+
# @param page_id [Integer]
|
17
|
+
def initialize(prediction, page_id)
|
18
|
+
@value = prediction['value']
|
19
|
+
@page_id = page_id
|
20
|
+
end
|
21
|
+
end
|
22
|
+
|
23
|
+
# Page object
|
24
|
+
class Page
|
25
|
+
# @return [Integer]
|
26
|
+
attr_reader :id
|
27
|
+
# @return [Mindee::Prediction]
|
28
|
+
attr_reader :prediction
|
29
|
+
# @return [Mindee::Orientation]
|
30
|
+
attr_reader :orientation
|
31
|
+
|
32
|
+
def initialize(prediction_class, page)
|
33
|
+
@id = page['id'].to_i
|
34
|
+
@prediction = prediction_class.new(page['prediction'], @id)
|
35
|
+
@orientation = Orientation.new(page['orientation'], @id)
|
36
|
+
end
|
37
|
+
|
38
|
+
def to_s
|
39
|
+
out_str = String.new
|
40
|
+
title = "Page #{@id}"
|
41
|
+
out_str << "#{title}\n"
|
42
|
+
out_str << ('-' * title.size)
|
43
|
+
out_str << "\n#{@prediction}"
|
44
|
+
end
|
45
|
+
end
|
46
|
+
end
|
@@ -0,0 +1,30 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
module Mindee
|
4
|
+
module Prediction
|
5
|
+
# Base document object.
|
6
|
+
class Prediction
|
7
|
+
# document type
|
8
|
+
# @return [String]
|
9
|
+
attr_reader :document_type
|
10
|
+
# Validation checks for the document
|
11
|
+
# @return [Hash<Symbol, Boolean>]
|
12
|
+
attr_reader :checklist
|
13
|
+
# Original filename of the document
|
14
|
+
# @return [String, nil]
|
15
|
+
attr_reader :filename
|
16
|
+
# Detected MIME type of the document
|
17
|
+
# @return [String, nil]
|
18
|
+
attr_reader :file_mimetype
|
19
|
+
|
20
|
+
def initialize(_prediction, _page_id)
|
21
|
+
@checklist = {}
|
22
|
+
end
|
23
|
+
|
24
|
+
# @return [Boolean]
|
25
|
+
def all_checks
|
26
|
+
@checklist.all? { |_, value| value == true }
|
27
|
+
end
|
28
|
+
end
|
29
|
+
end
|
30
|
+
end
|
@@ -4,7 +4,7 @@ require_relative 'base'
|
|
4
4
|
|
5
5
|
module Mindee
|
6
6
|
# Represents tax information.
|
7
|
-
class
|
7
|
+
class AmountField < Field
|
8
8
|
# Amount value as 3 decimal float
|
9
9
|
# @return [Float, nil]
|
10
10
|
attr_reader :value
|
@@ -13,5 +13,9 @@ module Mindee
|
|
13
13
|
super
|
14
14
|
@value = @value.round(3) unless @value.nil?
|
15
15
|
end
|
16
|
+
|
17
|
+
def to_s
|
18
|
+
Field.float_to_string(@value)
|
19
|
+
end
|
16
20
|
end
|
17
21
|
end
|
@@ -1,6 +1,6 @@
|
|
1
1
|
# frozen_string_literal: true
|
2
2
|
|
3
|
-
require_relative '
|
3
|
+
require_relative '../../../geometry'
|
4
4
|
|
5
5
|
module Mindee
|
6
6
|
# Base field object.
|
@@ -8,8 +8,8 @@ module Mindee
|
|
8
8
|
# @return [String, Float, Integer, Boolean]
|
9
9
|
attr_reader :value
|
10
10
|
# @return [Array<Array<Float>>]
|
11
|
-
attr_reader :
|
12
|
-
# @return [
|
11
|
+
attr_reader :bounding_box
|
12
|
+
# @return [Mindee::Geometry::Polygon]
|
13
13
|
attr_reader :polygon
|
14
14
|
# @return [Integer, nil]
|
15
15
|
attr_reader :page_id
|
@@ -26,8 +26,8 @@ module Mindee
|
|
26
26
|
def initialize(prediction, page_id, reconstructed: false)
|
27
27
|
@value = prediction['value']
|
28
28
|
@confidence = prediction['confidence']
|
29
|
-
@polygon = prediction['polygon']
|
30
|
-
@
|
29
|
+
@polygon = Geometry.polygon_from_prediction(prediction['polygon'])
|
30
|
+
@bounding_box = Geometry.get_bounding_box(@polygon) unless @polygon.nil? || @polygon.empty?
|
31
31
|
@page_id = page_id || prediction['page_id']
|
32
32
|
@reconstructed = reconstructed
|
33
33
|
end
|
@@ -57,5 +57,16 @@ module Mindee
|
|
57
57
|
end
|
58
58
|
arr_sum.to_f
|
59
59
|
end
|
60
|
+
|
61
|
+
# @param value [Float]
|
62
|
+
# @param min_precision [Integer]
|
63
|
+
def self.float_to_string(value, min_precision = 2)
|
64
|
+
return String.new if value.nil?
|
65
|
+
|
66
|
+
precision = value.to_f.to_s.split('.')[1].size
|
67
|
+
precision = [precision, min_precision].max
|
68
|
+
format_string = "%.#{precision}f"
|
69
|
+
format(format_string, value)
|
70
|
+
end
|
60
71
|
end
|
61
72
|
end
|
File without changes
|
File without changes
|
File without changes
|
File without changes
|