context.dev 2.18.0 → 2.20.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +4 -4
- data/CHANGELOG.md +220 -0
- data/README.md +1 -1
- data/lib/context_dev/client.rb +0 -4
- data/lib/context_dev/models/monitor_retrieve_run_params.rb +26 -0
- data/lib/context_dev/models/monitor_retrieve_run_response.rb +232 -0
- data/lib/context_dev/models/monitor_rotate_webhook_secret_params.rb +20 -0
- data/lib/context_dev/models/monitor_rotate_webhook_secret_response.rb +765 -0
- data/lib/context_dev/models/{web_web_scrape_sitemap_params.rb → web_map_urls_params.rb} +14 -14
- data/lib/context_dev/models/web_map_urls_response.rb +123 -0
- data/lib/context_dev/models/web_scrape_params.rb +1035 -0
- data/lib/context_dev/models/web_scrape_response.rb +974 -0
- data/lib/context_dev/models/web_screenshot_params.rb +15 -1
- data/lib/context_dev/models/web_screenshot_response.rb +3 -3
- data/lib/context_dev/models.rb +8 -22
- data/lib/context_dev/resources/brand.rb +0 -36
- data/lib/context_dev/resources/monitors.rb +47 -0
- data/lib/context_dev/resources/web.rb +118 -486
- data/lib/context_dev/version.rb +1 -1
- data/lib/context_dev.rb +8 -23
- data/rbi/context_dev/client.rbi +0 -3
- data/rbi/context_dev/models/monitor_retrieve_run_params.rbi +46 -0
- data/rbi/context_dev/models/monitor_retrieve_run_response.rbi +452 -0
- data/rbi/context_dev/models/monitor_rotate_webhook_secret_params.rbi +38 -0
- data/rbi/context_dev/models/monitor_rotate_webhook_secret_response.rbi +1367 -0
- data/rbi/context_dev/models/{web_web_scrape_sitemap_params.rbi → web_map_urls_params.rbi} +23 -45
- data/rbi/context_dev/models/web_map_urls_response.rbi +226 -0
- data/rbi/context_dev/models/web_scrape_params.rbi +2221 -0
- data/rbi/context_dev/models/web_scrape_response.rbi +1874 -0
- data/rbi/context_dev/models/web_screenshot_params.rbi +23 -0
- data/rbi/context_dev/models/web_screenshot_response.rbi +4 -4
- data/rbi/context_dev/models.rbi +9 -24
- data/rbi/context_dev/resources/brand.rbi +0 -35
- data/rbi/context_dev/resources/monitors.rbi +24 -0
- data/rbi/context_dev/resources/web.rbi +152 -654
- data/sig/context_dev/client.rbs +0 -2
- data/sig/context_dev/models/monitor_retrieve_run_params.rbs +28 -0
- data/sig/context_dev/models/monitor_retrieve_run_response.rbs +182 -0
- data/sig/context_dev/models/monitor_rotate_webhook_secret_params.rbs +23 -0
- data/sig/context_dev/models/monitor_rotate_webhook_secret_response.rbs +545 -0
- data/sig/context_dev/models/{web_web_scrape_sitemap_params.rbs → web_map_urls_params.rbs} +22 -22
- data/sig/context_dev/models/web_map_urls_response.rbs +125 -0
- data/sig/context_dev/models/web_scrape_params.rbs +925 -0
- data/sig/context_dev/models/web_scrape_response.rbs +782 -0
- data/sig/context_dev/models/web_screenshot_params.rbs +7 -0
- data/sig/context_dev/models.rbs +8 -22
- data/sig/context_dev/resources/brand.rbs +0 -9
- data/sig/context_dev/resources/monitors.rbs +11 -0
- data/sig/context_dev/resources/web.rbs +33 -128
- metadata +26 -71
- data/lib/context_dev/models/ai_extract_product_params.rb +0 -125
- data/lib/context_dev/models/ai_extract_product_response.rb +0 -402
- data/lib/context_dev/models/ai_extract_products_params.rb +0 -233
- data/lib/context_dev/models/ai_extract_products_response.rb +0 -370
- data/lib/context_dev/models/brand_retrieve_simplified_params.rb +0 -120
- data/lib/context_dev/models/brand_retrieve_simplified_response.rb +0 -441
- data/lib/context_dev/models/web_extract_fonts_params.rb +0 -114
- data/lib/context_dev/models/web_extract_fonts_response.rb +0 -291
- data/lib/context_dev/models/web_extract_params.rb +0 -419
- data/lib/context_dev/models/web_extract_response.rb +0 -287
- data/lib/context_dev/models/web_web_scrape_bytes_params.rb +0 -346
- data/lib/context_dev/models/web_web_scrape_bytes_response.rb +0 -135
- data/lib/context_dev/models/web_web_scrape_html_params.rb +0 -812
- data/lib/context_dev/models/web_web_scrape_html_response.rb +0 -628
- data/lib/context_dev/models/web_web_scrape_images_params.rb +0 -373
- data/lib/context_dev/models/web_web_scrape_images_response.rb +0 -409
- data/lib/context_dev/models/web_web_scrape_md_params.rb +0 -670
- data/lib/context_dev/models/web_web_scrape_md_response.rb +0 -567
- data/lib/context_dev/models/web_web_scrape_screenshot_params.rb +0 -471
- data/lib/context_dev/models/web_web_scrape_screenshot_response.rb +0 -169
- data/lib/context_dev/models/web_web_scrape_sitemap_response.rb +0 -144
- data/lib/context_dev/resources/ai.rb +0 -71
- data/rbi/context_dev/models/ai_extract_product_params.rbi +0 -253
- data/rbi/context_dev/models/ai_extract_product_response.rbi +0 -791
- data/rbi/context_dev/models/ai_extract_products_params.rbi +0 -490
- data/rbi/context_dev/models/ai_extract_products_response.rbi +0 -718
- data/rbi/context_dev/models/brand_retrieve_simplified_params.rbi +0 -256
- data/rbi/context_dev/models/brand_retrieve_simplified_response.rbi +0 -1040
- data/rbi/context_dev/models/web_extract_fonts_params.rbi +0 -223
- data/rbi/context_dev/models/web_extract_fonts_response.rbi +0 -563
- data/rbi/context_dev/models/web_extract_params.rbi +0 -781
- data/rbi/context_dev/models/web_extract_response.rbi +0 -512
- data/rbi/context_dev/models/web_web_scrape_bytes_params.rbi +0 -702
- data/rbi/context_dev/models/web_web_scrape_bytes_response.rbi +0 -238
- data/rbi/context_dev/models/web_web_scrape_html_params.rbi +0 -1690
- data/rbi/context_dev/models/web_web_scrape_html_response.rbi +0 -1287
- data/rbi/context_dev/models/web_web_scrape_images_params.rbi +0 -728
- data/rbi/context_dev/models/web_web_scrape_images_response.rbi +0 -910
- data/rbi/context_dev/models/web_web_scrape_md_params.rbi +0 -1052
- data/rbi/context_dev/models/web_web_scrape_md_response.rbi +0 -1088
- data/rbi/context_dev/models/web_web_scrape_screenshot_params.rbi +0 -1580
- data/rbi/context_dev/models/web_web_scrape_screenshot_response.rbi +0 -332
- data/rbi/context_dev/models/web_web_scrape_sitemap_response.rbi +0 -258
- data/rbi/context_dev/resources/ai.rbi +0 -63
- data/sig/context_dev/models/ai_extract_product_params.rbs +0 -106
- data/sig/context_dev/models/ai_extract_product_response.rbs +0 -314
- data/sig/context_dev/models/ai_extract_products_params.rbs +0 -202
- data/sig/context_dev/models/ai_extract_products_response.rbs +0 -293
- data/sig/context_dev/models/brand_retrieve_simplified_params.rbs +0 -104
- data/sig/context_dev/models/brand_retrieve_simplified_response.rbs +0 -413
- data/sig/context_dev/models/web_extract_fonts_params.rbs +0 -93
- data/sig/context_dev/models/web_extract_fonts_response.rbs +0 -230
- data/sig/context_dev/models/web_extract_params.rbs +0 -324
- data/sig/context_dev/models/web_extract_response.rbs +0 -239
- data/sig/context_dev/models/web_web_scrape_bytes_params.rbs +0 -531
- data/sig/context_dev/models/web_web_scrape_bytes_response.rbs +0 -108
- data/sig/context_dev/models/web_web_scrape_html_params.rbs +0 -878
- data/sig/context_dev/models/web_web_scrape_html_response.rbs +0 -525
- data/sig/context_dev/models/web_web_scrape_images_params.rbs +0 -281
- data/sig/context_dev/models/web_web_scrape_images_response.rbs +0 -371
- data/sig/context_dev/models/web_web_scrape_md_params.rbs +0 -758
- data/sig/context_dev/models/web_web_scrape_md_response.rbs +0 -465
- data/sig/context_dev/models/web_web_scrape_screenshot_params.rbs +0 -619
- data/sig/context_dev/models/web_web_scrape_screenshot_response.rbs +0 -127
- data/sig/context_dev/models/web_web_scrape_sitemap_response.rbs +0 -117
- data/sig/context_dev/resources/ai.rbs +0 -21
|
@@ -1,812 +0,0 @@
|
|
|
1
|
-
# frozen_string_literal: true
|
|
2
|
-
|
|
3
|
-
module ContextDev
|
|
4
|
-
module Models
|
|
5
|
-
# @see ContextDev::Resources::Web#web_scrape_html
|
|
6
|
-
class WebWebScrapeHTMLParams < ContextDev::Internal::Type::BaseModel
|
|
7
|
-
extend ContextDev::Internal::Type::RequestParameters::Converter
|
|
8
|
-
include ContextDev::Internal::Type::RequestParameters
|
|
9
|
-
|
|
10
|
-
# @!attribute url
|
|
11
|
-
# Full URL to scrape (must include http:// or https:// protocol)
|
|
12
|
-
#
|
|
13
|
-
# @return [String]
|
|
14
|
-
required :url, String
|
|
15
|
-
|
|
16
|
-
# @!attribute actions
|
|
17
|
-
# Optional browser actions executed in array order after the page loads and before
|
|
18
|
-
# content is captured. Requires a paid plan. Send a JSON array in the query
|
|
19
|
-
# parameter. Maximum: 5 actions.
|
|
20
|
-
#
|
|
21
|
-
# @return [Array<ContextDev::Models::WebWebScrapeHTMLParams::Action::Wait, ContextDev::Models::WebWebScrapeHTMLParams::Action::Perform, ContextDev::Models::WebWebScrapeHTMLParams::Action::Scroll>, nil]
|
|
22
|
-
optional :actions,
|
|
23
|
-
-> { ContextDev::Internal::Type::ArrayOf[union: ContextDev::WebWebScrapeHTMLParams::Action] },
|
|
24
|
-
nil?: true
|
|
25
|
-
|
|
26
|
-
# @!attribute country
|
|
27
|
-
# Fetch the target page through a residential proxy in this country (ISO 3166-1
|
|
28
|
-
# alpha-2).
|
|
29
|
-
#
|
|
30
|
-
# @return [Symbol, ContextDev::Models::WebWebScrapeHTMLParams::Country, nil]
|
|
31
|
-
optional :country, enum: -> { ContextDev::WebWebScrapeHTMLParams::Country }
|
|
32
|
-
|
|
33
|
-
# @!attribute exclude_selectors
|
|
34
|
-
# CSS selectors to remove from the result. Applied after includeSelectors.
|
|
35
|
-
# Exclusion takes precedence: an element matching both is removed. Examples:
|
|
36
|
-
# "nav", "footer", ".ad-banner", "[aria-hidden=true]".
|
|
37
|
-
#
|
|
38
|
-
# @return [Array<String>, nil]
|
|
39
|
-
optional :exclude_selectors, ContextDev::Internal::Type::ArrayOf[String], nil?: true
|
|
40
|
-
|
|
41
|
-
# @!attribute extract_rules
|
|
42
|
-
# Optional CSS extraction rules applied to the returned HTML after selector and
|
|
43
|
-
# main-content filtering. Use selector strings ("h1", "a@href") or objects with
|
|
44
|
-
# selector, type (item or list), and output (text, html, @attribute, or nested
|
|
45
|
-
# rules). Text whitespace is normalized; html includes the matched element;
|
|
46
|
-
# attributes are returned as written. Missing items are null and missing lists are
|
|
47
|
-
# empty. CSS only; XPath is not supported. Maximum: 100 fields across 5 levels.
|
|
48
|
-
# Send a JSON-encoded string in the extractRules query parameter.
|
|
49
|
-
#
|
|
50
|
-
# @return [Hash{Symbol=>String, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1}, nil]
|
|
51
|
-
optional :extract_rules,
|
|
52
|
-
-> { ContextDev::Internal::Type::HashOf[union: ContextDev::WebWebScrapeHTMLParams::ExtractRule] }
|
|
53
|
-
|
|
54
|
-
# @!attribute headers
|
|
55
|
-
# Optional outbound HTTP headers forwarded only to the target URL, sent as
|
|
56
|
-
# deep-object query params such as headers[X-Custom]=value. When provided, caching
|
|
57
|
-
# is bypassed: the result is neither read from nor written to cache.
|
|
58
|
-
#
|
|
59
|
-
# @return [Hash{Symbol=>String}, nil]
|
|
60
|
-
optional :headers, ContextDev::Internal::Type::HashOf[String]
|
|
61
|
-
|
|
62
|
-
# @!attribute include_frames
|
|
63
|
-
# When true, iframes are rendered inline into the returned HTML.
|
|
64
|
-
#
|
|
65
|
-
# @return [Boolean, nil]
|
|
66
|
-
optional :include_frames, ContextDev::Internal::Type::Boolean
|
|
67
|
-
|
|
68
|
-
# @!attribute include_selectors
|
|
69
|
-
# CSS selectors. When provided, only matching subtrees (and their descendants) are
|
|
70
|
-
# kept and everything else is dropped. When omitted, the entire document is kept.
|
|
71
|
-
# Examples: "article.main", "#content", "[role=main]".
|
|
72
|
-
#
|
|
73
|
-
# @return [Array<String>, nil]
|
|
74
|
-
optional :include_selectors, ContextDev::Internal::Type::ArrayOf[String], nil?: true
|
|
75
|
-
|
|
76
|
-
# @!attribute max_age_ms
|
|
77
|
-
# Return a cached result if a prior scrape for the same parameters exists and is
|
|
78
|
-
# younger than this many milliseconds. Defaults to 1 day (86400000 ms) when
|
|
79
|
-
# omitted. Max is 30 days (2592000000 ms). Set to 0 to always scrape fresh.
|
|
80
|
-
#
|
|
81
|
-
# @return [Integer, nil]
|
|
82
|
-
optional :max_age_ms, Integer, nil?: true
|
|
83
|
-
|
|
84
|
-
# @!attribute pdf
|
|
85
|
-
# PDF parsing controls. Use start/end to limit text extraction and embedded-image
|
|
86
|
-
# detection/OCR to an inclusive 1-based page range.
|
|
87
|
-
#
|
|
88
|
-
# @return [ContextDev::Models::WebWebScrapeHTMLParams::Pdf, nil]
|
|
89
|
-
optional :pdf, -> { ContextDev::WebWebScrapeHTMLParams::Pdf }
|
|
90
|
-
|
|
91
|
-
# @!attribute settle_animations
|
|
92
|
-
# When true, waits briefly for CSS and transition animations to settle before
|
|
93
|
-
# extracting HTML. Defaults to false. This adds a bit of latency in exchange for
|
|
94
|
-
# more stable output on animated pages.
|
|
95
|
-
#
|
|
96
|
-
# @return [Boolean, nil]
|
|
97
|
-
optional :settle_animations, ContextDev::Internal::Type::Boolean
|
|
98
|
-
|
|
99
|
-
# @!attribute tags
|
|
100
|
-
# Comma-separated tags for tracking request usage. Up to 20 tags, each 1-50
|
|
101
|
-
# characters.
|
|
102
|
-
#
|
|
103
|
-
# @return [Array<String>, nil]
|
|
104
|
-
optional :tags, ContextDev::Internal::Type::ArrayOf[String]
|
|
105
|
-
|
|
106
|
-
# @!attribute timeout_opts
|
|
107
|
-
# Optional request deadline and behavior on timeout. For GET requests, use
|
|
108
|
-
# timeoutOpts[milliseconds]=30000&timeoutOpts[behavior]=fail or a JSON-encoded
|
|
109
|
-
# timeoutOpts object.
|
|
110
|
-
#
|
|
111
|
-
# @return [ContextDev::Models::WebWebScrapeHTMLParams::TimeoutOpts, nil]
|
|
112
|
-
optional :timeout_opts, -> { ContextDev::WebWebScrapeHTMLParams::TimeoutOpts }
|
|
113
|
-
|
|
114
|
-
# @!attribute use_main_content_only
|
|
115
|
-
# When true, return only the page's main content in the HTML response, excluding
|
|
116
|
-
# headers, footers, sidebars, and navigation when detectable.
|
|
117
|
-
#
|
|
118
|
-
# @return [Boolean, nil]
|
|
119
|
-
optional :use_main_content_only, ContextDev::Internal::Type::Boolean
|
|
120
|
-
|
|
121
|
-
# @!attribute wait_for_ms
|
|
122
|
-
# Optional browser wait time in milliseconds after initial page load. Min: 0. Max:
|
|
123
|
-
# 30000 (30 seconds). When combined with timeoutOpts, timeoutOpts.milliseconds
|
|
124
|
-
# must be at least waitForMs + 10000 ms; a shorter deadline is rejected with 400
|
|
125
|
-
# TIMEOUT_TOO_SHORT_FOR_WAIT.
|
|
126
|
-
#
|
|
127
|
-
# @return [Integer, nil]
|
|
128
|
-
optional :wait_for_ms, Integer, nil?: true
|
|
129
|
-
|
|
130
|
-
# @!attribute zdr
|
|
131
|
-
# Set to enabled to bypass shared caches and omit request and response content
|
|
132
|
-
# from retained usage logs. Asset uploads are skipped, so hosted image URLs are
|
|
133
|
-
# omitted. Requires zero data retention to be enabled for your organization
|
|
134
|
-
# (contact support@context.dev), otherwise the request fails with ZDR_NOT_ENABLED.
|
|
135
|
-
# Successful ZDR responses include X-Context-ZDR: true.
|
|
136
|
-
#
|
|
137
|
-
# @return [Symbol, ContextDev::Models::WebWebScrapeHTMLParams::Zdr, nil]
|
|
138
|
-
optional :zdr, enum: -> { ContextDev::WebWebScrapeHTMLParams::Zdr }
|
|
139
|
-
|
|
140
|
-
# @!method initialize(url:, actions: nil, country: nil, exclude_selectors: nil, extract_rules: nil, headers: nil, include_frames: nil, include_selectors: nil, max_age_ms: nil, pdf: nil, settle_animations: nil, tags: nil, timeout_opts: nil, use_main_content_only: nil, wait_for_ms: nil, zdr: nil, request_options: {})
|
|
141
|
-
# Some parameter documentations has been truncated, see
|
|
142
|
-
# {ContextDev::Models::WebWebScrapeHTMLParams} for more details.
|
|
143
|
-
#
|
|
144
|
-
# @param url [String] Full URL to scrape (must include http:// or https:// protocol)
|
|
145
|
-
#
|
|
146
|
-
# @param actions [Array<ContextDev::Models::WebWebScrapeHTMLParams::Action::Wait, ContextDev::Models::WebWebScrapeHTMLParams::Action::Perform, ContextDev::Models::WebWebScrapeHTMLParams::Action::Scroll>, nil] Optional browser actions executed in array order after the page loads and before
|
|
147
|
-
#
|
|
148
|
-
# @param country [Symbol, ContextDev::Models::WebWebScrapeHTMLParams::Country] Fetch the target page through a residential proxy in this country (ISO 3166-1 al
|
|
149
|
-
#
|
|
150
|
-
# @param exclude_selectors [Array<String>, nil] CSS selectors to remove from the result. Applied after includeSelectors. Exclusi
|
|
151
|
-
#
|
|
152
|
-
# @param extract_rules [Hash{Symbol=>String, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1}] Optional CSS extraction rules applied to the returned HTML after selector and ma
|
|
153
|
-
#
|
|
154
|
-
# @param headers [Hash{Symbol=>String}] Optional outbound HTTP headers forwarded only to the target URL, sent as deep-ob
|
|
155
|
-
#
|
|
156
|
-
# @param include_frames [Boolean] When true, iframes are rendered inline into the returned HTML.
|
|
157
|
-
#
|
|
158
|
-
# @param include_selectors [Array<String>, nil] CSS selectors. When provided, only matching subtrees (and their descendants) are
|
|
159
|
-
#
|
|
160
|
-
# @param max_age_ms [Integer, nil] Return a cached result if a prior scrape for the same parameters exists and is y
|
|
161
|
-
#
|
|
162
|
-
# @param pdf [ContextDev::Models::WebWebScrapeHTMLParams::Pdf] PDF parsing controls. Use start/end to limit text extraction and embedded-image
|
|
163
|
-
#
|
|
164
|
-
# @param settle_animations [Boolean] When true, waits briefly for CSS and transition animations to settle before extr
|
|
165
|
-
#
|
|
166
|
-
# @param tags [Array<String>] Comma-separated tags for tracking request usage. Up to 20 tags, each 1-50 charac
|
|
167
|
-
#
|
|
168
|
-
# @param timeout_opts [ContextDev::Models::WebWebScrapeHTMLParams::TimeoutOpts] Optional request deadline and behavior on timeout. For GET requests, use timeout
|
|
169
|
-
#
|
|
170
|
-
# @param use_main_content_only [Boolean] When true, return only the page's main content in the HTML response, excluding h
|
|
171
|
-
#
|
|
172
|
-
# @param wait_for_ms [Integer, nil] Optional browser wait time in milliseconds after initial page load. Min: 0. Max:
|
|
173
|
-
#
|
|
174
|
-
# @param zdr [Symbol, ContextDev::Models::WebWebScrapeHTMLParams::Zdr] Set to enabled to bypass shared caches and omit request and response content fro
|
|
175
|
-
#
|
|
176
|
-
# @param request_options [ContextDev::RequestOptions, Hash{Symbol=>Object}]
|
|
177
|
-
|
|
178
|
-
# Browser action discriminated by `do`. Each variant exposes only its applicable
|
|
179
|
-
# fields.
|
|
180
|
-
module Action
|
|
181
|
-
extend ContextDev::Internal::Type::Union
|
|
182
|
-
|
|
183
|
-
discriminator :do
|
|
184
|
-
|
|
185
|
-
# Pause for a fixed number of milliseconds before continuing to the next action.
|
|
186
|
-
variant :wait, -> { ContextDev::WebWebScrapeHTMLParams::Action::Wait }
|
|
187
|
-
|
|
188
|
-
# Resolve and perform one natural-language browser action.
|
|
189
|
-
variant :perform, -> { ContextDev::WebWebScrapeHTMLParams::Action::Perform }
|
|
190
|
-
|
|
191
|
-
# Scroll the page or a selected scrollable container, waiting adaptively for content and dimensions to settle after each iteration.
|
|
192
|
-
variant :scroll, -> { ContextDev::WebWebScrapeHTMLParams::Action::Scroll }
|
|
193
|
-
|
|
194
|
-
class Wait < ContextDev::Internal::Type::BaseModel
|
|
195
|
-
# @!attribute do_
|
|
196
|
-
#
|
|
197
|
-
# @return [Symbol, :wait]
|
|
198
|
-
required :do_, const: :wait, api_name: :do
|
|
199
|
-
|
|
200
|
-
# @!attribute time_ms
|
|
201
|
-
#
|
|
202
|
-
# @return [Integer]
|
|
203
|
-
required :time_ms, Integer, api_name: :timeMs
|
|
204
|
-
|
|
205
|
-
# @!method initialize(time_ms:, do_: :wait)
|
|
206
|
-
# Pause for a fixed number of milliseconds before continuing to the next action.
|
|
207
|
-
#
|
|
208
|
-
# @param time_ms [Integer]
|
|
209
|
-
# @param do_ [Symbol, :wait]
|
|
210
|
-
end
|
|
211
|
-
|
|
212
|
-
class Perform < ContextDev::Internal::Type::BaseModel
|
|
213
|
-
# @!attribute action
|
|
214
|
-
#
|
|
215
|
-
# @return [String]
|
|
216
|
-
required :action, String
|
|
217
|
-
|
|
218
|
-
# @!attribute do_
|
|
219
|
-
#
|
|
220
|
-
# @return [Symbol, :perform]
|
|
221
|
-
required :do_, const: :perform, api_name: :do
|
|
222
|
-
|
|
223
|
-
# @!method initialize(action:, do_: :perform)
|
|
224
|
-
# Resolve and perform one natural-language browser action.
|
|
225
|
-
#
|
|
226
|
-
# @param action [String]
|
|
227
|
-
# @param do_ [Symbol, :perform]
|
|
228
|
-
end
|
|
229
|
-
|
|
230
|
-
class Scroll < ContextDev::Internal::Type::BaseModel
|
|
231
|
-
# @!attribute do_
|
|
232
|
-
#
|
|
233
|
-
# @return [Symbol, :scroll]
|
|
234
|
-
required :do_, const: :scroll, api_name: :do
|
|
235
|
-
|
|
236
|
-
# @!attribute amount
|
|
237
|
-
# Pixels per scroll, one visible viewport, or the current scroll boundary.
|
|
238
|
-
# Defaults to viewport.
|
|
239
|
-
#
|
|
240
|
-
# @return [Integer, Symbol, ContextDev::Models::WebWebScrapeHTMLParams::Action::Scroll::Amount, nil]
|
|
241
|
-
optional :amount, union: -> { ContextDev::WebWebScrapeHTMLParams::Action::Scroll::Amount }
|
|
242
|
-
|
|
243
|
-
# @!attribute container
|
|
244
|
-
# CSS selector for the first matching scroll container. Defaults to the page.
|
|
245
|
-
#
|
|
246
|
-
# @return [String, nil]
|
|
247
|
-
optional :container, String
|
|
248
|
-
|
|
249
|
-
# @!attribute direction
|
|
250
|
-
# Direction to scroll. Defaults to down.
|
|
251
|
-
#
|
|
252
|
-
# @return [Symbol, ContextDev::Models::WebWebScrapeHTMLParams::Action::Scroll::Direction, nil]
|
|
253
|
-
optional :direction, enum: -> { ContextDev::WebWebScrapeHTMLParams::Action::Scroll::Direction }
|
|
254
|
-
|
|
255
|
-
# @!attribute max_scrolls
|
|
256
|
-
# Maximum scroll iterations. Stops early when scrolling and scrollable extent stop
|
|
257
|
-
# changing. Defaults to 1.
|
|
258
|
-
#
|
|
259
|
-
# @return [Integer, nil]
|
|
260
|
-
optional :max_scrolls, Integer, api_name: :maxScrolls
|
|
261
|
-
|
|
262
|
-
# @!method initialize(amount: nil, container: nil, direction: nil, max_scrolls: nil, do_: :scroll)
|
|
263
|
-
# Some parameter documentations has been truncated, see
|
|
264
|
-
# {ContextDev::Models::WebWebScrapeHTMLParams::Action::Scroll} for more details.
|
|
265
|
-
#
|
|
266
|
-
# Scroll the page or a selected scrollable container, waiting adaptively for
|
|
267
|
-
# content and dimensions to settle after each iteration.
|
|
268
|
-
#
|
|
269
|
-
# @param amount [Integer, Symbol, ContextDev::Models::WebWebScrapeHTMLParams::Action::Scroll::Amount] Pixels per scroll, one visible viewport, or the current scroll boundary. Default
|
|
270
|
-
#
|
|
271
|
-
# @param container [String] CSS selector for the first matching scroll container. Defaults to the page.
|
|
272
|
-
#
|
|
273
|
-
# @param direction [Symbol, ContextDev::Models::WebWebScrapeHTMLParams::Action::Scroll::Direction] Direction to scroll. Defaults to down.
|
|
274
|
-
#
|
|
275
|
-
# @param max_scrolls [Integer] Maximum scroll iterations. Stops early when scrolling and scrollable extent stop
|
|
276
|
-
#
|
|
277
|
-
# @param do_ [Symbol, :scroll]
|
|
278
|
-
|
|
279
|
-
# Pixels per scroll, one visible viewport, or the current scroll boundary.
|
|
280
|
-
# Defaults to viewport.
|
|
281
|
-
#
|
|
282
|
-
# @see ContextDev::Models::WebWebScrapeHTMLParams::Action::Scroll#amount
|
|
283
|
-
module Amount
|
|
284
|
-
extend ContextDev::Internal::Type::Union
|
|
285
|
-
|
|
286
|
-
variant Integer
|
|
287
|
-
|
|
288
|
-
variant const: -> { ContextDev::Models::WebWebScrapeHTMLParams::Action::Scroll::Amount::VIEWPORT }
|
|
289
|
-
|
|
290
|
-
variant const: -> { ContextDev::Models::WebWebScrapeHTMLParams::Action::Scroll::Amount::MAX }
|
|
291
|
-
|
|
292
|
-
# @!method self.variants
|
|
293
|
-
# @return [Array(Integer, Symbol)]
|
|
294
|
-
|
|
295
|
-
define_sorbet_constant!(:Variants) do
|
|
296
|
-
T.type_alias { T.any(Integer, ContextDev::WebWebScrapeHTMLParams::Action::Scroll::Amount::TaggedSymbol) }
|
|
297
|
-
end
|
|
298
|
-
|
|
299
|
-
# @!group
|
|
300
|
-
|
|
301
|
-
VIEWPORT = :viewport
|
|
302
|
-
MAX = :max
|
|
303
|
-
|
|
304
|
-
# @!endgroup
|
|
305
|
-
end
|
|
306
|
-
|
|
307
|
-
# Direction to scroll. Defaults to down.
|
|
308
|
-
#
|
|
309
|
-
# @see ContextDev::Models::WebWebScrapeHTMLParams::Action::Scroll#direction
|
|
310
|
-
module Direction
|
|
311
|
-
extend ContextDev::Internal::Type::Enum
|
|
312
|
-
|
|
313
|
-
UP = :up
|
|
314
|
-
DOWN = :down
|
|
315
|
-
LEFT = :left
|
|
316
|
-
RIGHT = :right
|
|
317
|
-
|
|
318
|
-
# @!method self.values
|
|
319
|
-
# @return [Array<Symbol>]
|
|
320
|
-
end
|
|
321
|
-
end
|
|
322
|
-
|
|
323
|
-
# @!method self.variants
|
|
324
|
-
# @return [Array(ContextDev::Models::WebWebScrapeHTMLParams::Action::Wait, ContextDev::Models::WebWebScrapeHTMLParams::Action::Perform, ContextDev::Models::WebWebScrapeHTMLParams::Action::Scroll)]
|
|
325
|
-
end
|
|
326
|
-
|
|
327
|
-
# Fetch the target page through a residential proxy in this country (ISO 3166-1
|
|
328
|
-
# alpha-2).
|
|
329
|
-
module Country
|
|
330
|
-
extend ContextDev::Internal::Type::Enum
|
|
331
|
-
|
|
332
|
-
AD = :ad
|
|
333
|
-
AE = :ae
|
|
334
|
-
AF = :af
|
|
335
|
-
AG = :ag
|
|
336
|
-
AI = :ai
|
|
337
|
-
AL = :al
|
|
338
|
-
AM = :am
|
|
339
|
-
AO = :ao
|
|
340
|
-
AR = :ar
|
|
341
|
-
AT = :at
|
|
342
|
-
AU = :au
|
|
343
|
-
AW = :aw
|
|
344
|
-
AZ = :az
|
|
345
|
-
BA = :ba
|
|
346
|
-
BB = :bb
|
|
347
|
-
BD = :bd
|
|
348
|
-
BE = :be
|
|
349
|
-
BF = :bf
|
|
350
|
-
BG = :bg
|
|
351
|
-
BH = :bh
|
|
352
|
-
BI = :bi
|
|
353
|
-
BJ = :bj
|
|
354
|
-
BM = :bm
|
|
355
|
-
BN = :bn
|
|
356
|
-
BO = :bo
|
|
357
|
-
BQ = :bq
|
|
358
|
-
BR = :br
|
|
359
|
-
BS = :bs
|
|
360
|
-
BW = :bw
|
|
361
|
-
BY = :by
|
|
362
|
-
BZ = :bz
|
|
363
|
-
CA = :ca
|
|
364
|
-
CD = :cd
|
|
365
|
-
CF = :cf
|
|
366
|
-
CG = :cg
|
|
367
|
-
CH = :ch
|
|
368
|
-
CI = :ci
|
|
369
|
-
CL = :cl
|
|
370
|
-
CM = :cm
|
|
371
|
-
CN = :cn
|
|
372
|
-
CO = :co
|
|
373
|
-
CR = :cr
|
|
374
|
-
CV = :cv
|
|
375
|
-
CW = :cw
|
|
376
|
-
CY = :cy
|
|
377
|
-
CZ = :cz
|
|
378
|
-
DE = :de
|
|
379
|
-
DJ = :dj
|
|
380
|
-
DK = :dk
|
|
381
|
-
DM = :dm
|
|
382
|
-
DO = :do
|
|
383
|
-
DZ = :dz
|
|
384
|
-
EC = :ec
|
|
385
|
-
EE = :ee
|
|
386
|
-
EG = :eg
|
|
387
|
-
ES = :es
|
|
388
|
-
ET = :et
|
|
389
|
-
FI = :fi
|
|
390
|
-
FJ = :fj
|
|
391
|
-
FR = :fr
|
|
392
|
-
GA = :ga
|
|
393
|
-
GB = :gb
|
|
394
|
-
GD = :gd
|
|
395
|
-
GE = :ge
|
|
396
|
-
GF = :gf
|
|
397
|
-
GG = :gg
|
|
398
|
-
GH = :gh
|
|
399
|
-
GM = :gm
|
|
400
|
-
GN = :gn
|
|
401
|
-
GP = :gp
|
|
402
|
-
GQ = :gq
|
|
403
|
-
GR = :gr
|
|
404
|
-
GT = :gt
|
|
405
|
-
GU = :gu
|
|
406
|
-
GW = :gw
|
|
407
|
-
GY = :gy
|
|
408
|
-
HK = :hk
|
|
409
|
-
HN = :hn
|
|
410
|
-
HR = :hr
|
|
411
|
-
HT = :ht
|
|
412
|
-
HU = :hu
|
|
413
|
-
ID = :id
|
|
414
|
-
IE = :ie
|
|
415
|
-
IL = :il
|
|
416
|
-
IM = :im
|
|
417
|
-
IN = :in
|
|
418
|
-
IQ = :iq
|
|
419
|
-
IR = :ir
|
|
420
|
-
IS = :is
|
|
421
|
-
IT = :it
|
|
422
|
-
JE = :je
|
|
423
|
-
JM = :jm
|
|
424
|
-
JO = :jo
|
|
425
|
-
JP = :jp
|
|
426
|
-
KE = :ke
|
|
427
|
-
KG = :kg
|
|
428
|
-
KH = :kh
|
|
429
|
-
KN = :kn
|
|
430
|
-
KR = :kr
|
|
431
|
-
KW = :kw
|
|
432
|
-
KY = :ky
|
|
433
|
-
KZ = :kz
|
|
434
|
-
LA = :la
|
|
435
|
-
LB = :lb
|
|
436
|
-
LC = :lc
|
|
437
|
-
LK = :lk
|
|
438
|
-
LR = :lr
|
|
439
|
-
LS = :ls
|
|
440
|
-
LT = :lt
|
|
441
|
-
LU = :lu
|
|
442
|
-
LV = :lv
|
|
443
|
-
LY = :ly
|
|
444
|
-
MA = :ma
|
|
445
|
-
MC = :mc
|
|
446
|
-
MD = :md
|
|
447
|
-
ME = :me
|
|
448
|
-
MF = :mf
|
|
449
|
-
MG = :mg
|
|
450
|
-
MK = :mk
|
|
451
|
-
ML = :ml
|
|
452
|
-
MM = :mm
|
|
453
|
-
MN = :mn
|
|
454
|
-
MO = :mo
|
|
455
|
-
MQ = :mq
|
|
456
|
-
MR = :mr
|
|
457
|
-
MT = :mt
|
|
458
|
-
MU = :mu
|
|
459
|
-
MV = :mv
|
|
460
|
-
MW = :mw
|
|
461
|
-
MX = :mx
|
|
462
|
-
MY = :my
|
|
463
|
-
MZ = :mz
|
|
464
|
-
NA = :na
|
|
465
|
-
NC = :nc
|
|
466
|
-
NE = :ne
|
|
467
|
-
NG = :ng
|
|
468
|
-
NI = :ni
|
|
469
|
-
NL = :nl
|
|
470
|
-
NO = :no
|
|
471
|
-
NP = :np
|
|
472
|
-
NZ = :nz
|
|
473
|
-
OM = :om
|
|
474
|
-
PA = :pa
|
|
475
|
-
PE = :pe
|
|
476
|
-
PF = :pf
|
|
477
|
-
PG = :pg
|
|
478
|
-
PH = :ph
|
|
479
|
-
PK = :pk
|
|
480
|
-
PL = :pl
|
|
481
|
-
PR = :pr
|
|
482
|
-
PS = :ps
|
|
483
|
-
PT = :pt
|
|
484
|
-
PY = :py
|
|
485
|
-
QA = :qa
|
|
486
|
-
RE = :re
|
|
487
|
-
RO = :ro
|
|
488
|
-
RS = :rs
|
|
489
|
-
RU = :ru
|
|
490
|
-
RW = :rw
|
|
491
|
-
SA = :sa
|
|
492
|
-
SC = :sc
|
|
493
|
-
SD = :sd
|
|
494
|
-
SE = :se
|
|
495
|
-
SG = :sg
|
|
496
|
-
SI = :si
|
|
497
|
-
SK = :sk
|
|
498
|
-
SL = :sl
|
|
499
|
-
SM = :sm
|
|
500
|
-
SN = :sn
|
|
501
|
-
SO = :so
|
|
502
|
-
SR = :sr
|
|
503
|
-
SS = :ss
|
|
504
|
-
ST = :st
|
|
505
|
-
SV = :sv
|
|
506
|
-
SX = :sx
|
|
507
|
-
SY = :sy
|
|
508
|
-
SZ = :sz
|
|
509
|
-
TC = :tc
|
|
510
|
-
TD = :td
|
|
511
|
-
TG = :tg
|
|
512
|
-
TH = :th
|
|
513
|
-
TJ = :tj
|
|
514
|
-
TL = :tl
|
|
515
|
-
TM = :tm
|
|
516
|
-
TN = :tn
|
|
517
|
-
TR = :tr
|
|
518
|
-
TT = :tt
|
|
519
|
-
TW = :tw
|
|
520
|
-
TZ = :tz
|
|
521
|
-
UA = :ua
|
|
522
|
-
UG = :ug
|
|
523
|
-
US = :us
|
|
524
|
-
UY = :uy
|
|
525
|
-
UZ = :uz
|
|
526
|
-
VC = :vc
|
|
527
|
-
VE = :ve
|
|
528
|
-
VG = :vg
|
|
529
|
-
VI = :vi
|
|
530
|
-
VN = :vn
|
|
531
|
-
YE = :ye
|
|
532
|
-
YT = :yt
|
|
533
|
-
ZA = :za
|
|
534
|
-
ZM = :zm
|
|
535
|
-
ZW = :zw
|
|
536
|
-
|
|
537
|
-
# @!method self.values
|
|
538
|
-
# @return [Array<Symbol>]
|
|
539
|
-
end
|
|
540
|
-
|
|
541
|
-
module ExtractRule
|
|
542
|
-
extend ContextDev::Internal::Type::Union
|
|
543
|
-
|
|
544
|
-
variant String
|
|
545
|
-
|
|
546
|
-
variant -> { ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1 }
|
|
547
|
-
|
|
548
|
-
class UnionMember1 < ContextDev::Internal::Type::BaseModel
|
|
549
|
-
# @!attribute selector
|
|
550
|
-
#
|
|
551
|
-
# @return [String]
|
|
552
|
-
required :selector, String
|
|
553
|
-
|
|
554
|
-
# @!attribute output
|
|
555
|
-
#
|
|
556
|
-
# @return [Symbol, String, Hash{Symbol=>String, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1}, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output, nil]
|
|
557
|
-
optional :output, union: -> { ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output }
|
|
558
|
-
|
|
559
|
-
# @!attribute type
|
|
560
|
-
#
|
|
561
|
-
# @return [Symbol, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Type, nil]
|
|
562
|
-
optional :type, enum: -> { ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Type }
|
|
563
|
-
|
|
564
|
-
# @!method initialize(selector:, output: nil, type: nil)
|
|
565
|
-
# @param selector [String]
|
|
566
|
-
# @param output [Symbol, String, Hash{Symbol=>String, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1}, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output]
|
|
567
|
-
# @param type [Symbol, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Type]
|
|
568
|
-
|
|
569
|
-
# @see ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1#output
|
|
570
|
-
module Output
|
|
571
|
-
extend ContextDev::Internal::Type::Union
|
|
572
|
-
|
|
573
|
-
variant const: -> { ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::TEXT }
|
|
574
|
-
|
|
575
|
-
variant const: -> { ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTML }
|
|
576
|
-
|
|
577
|
-
variant String
|
|
578
|
-
|
|
579
|
-
variant -> { ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRuleMap }
|
|
580
|
-
|
|
581
|
-
module HTMLExtractionRule
|
|
582
|
-
extend ContextDev::Internal::Type::Union
|
|
583
|
-
|
|
584
|
-
variant String
|
|
585
|
-
|
|
586
|
-
variant -> { ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1 }
|
|
587
|
-
|
|
588
|
-
class UnionMember1 < ContextDev::Internal::Type::BaseModel
|
|
589
|
-
# @!attribute selector
|
|
590
|
-
#
|
|
591
|
-
# @return [String]
|
|
592
|
-
required :selector, String
|
|
593
|
-
|
|
594
|
-
# @!attribute output
|
|
595
|
-
#
|
|
596
|
-
# @return [Symbol, String, Object, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output, nil]
|
|
597
|
-
optional :output,
|
|
598
|
-
union: -> { ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output }
|
|
599
|
-
|
|
600
|
-
# @!attribute type
|
|
601
|
-
#
|
|
602
|
-
# @return [Symbol, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Type, nil]
|
|
603
|
-
optional :type,
|
|
604
|
-
enum: -> { ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Type }
|
|
605
|
-
|
|
606
|
-
# @!method initialize(selector:, output: nil, type: nil)
|
|
607
|
-
# @param selector [String]
|
|
608
|
-
# @param output [Symbol, String, Object, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output]
|
|
609
|
-
# @param type [Symbol, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Type]
|
|
610
|
-
|
|
611
|
-
# @see ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1#output
|
|
612
|
-
module Output
|
|
613
|
-
extend ContextDev::Internal::Type::Union
|
|
614
|
-
|
|
615
|
-
variant const: -> { ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output::TEXT }
|
|
616
|
-
|
|
617
|
-
variant const: -> { ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output::HTML }
|
|
618
|
-
|
|
619
|
-
variant String
|
|
620
|
-
|
|
621
|
-
variant ContextDev::Internal::Type::Unknown
|
|
622
|
-
|
|
623
|
-
# @!method self.variants
|
|
624
|
-
# @return [Array(Symbol, String, Object)]
|
|
625
|
-
|
|
626
|
-
define_sorbet_constant!(:Variants) do
|
|
627
|
-
T.type_alias do
|
|
628
|
-
T.any(
|
|
629
|
-
ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output::TaggedSymbol,
|
|
630
|
-
String,
|
|
631
|
-
T.anything
|
|
632
|
-
)
|
|
633
|
-
end
|
|
634
|
-
end
|
|
635
|
-
|
|
636
|
-
# @!group
|
|
637
|
-
|
|
638
|
-
TEXT = :text
|
|
639
|
-
HTML = :html
|
|
640
|
-
|
|
641
|
-
# @!endgroup
|
|
642
|
-
end
|
|
643
|
-
|
|
644
|
-
# @see ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1#type
|
|
645
|
-
module Type
|
|
646
|
-
extend ContextDev::Internal::Type::Enum
|
|
647
|
-
|
|
648
|
-
ITEM = :item
|
|
649
|
-
LIST = :list
|
|
650
|
-
|
|
651
|
-
# @!method self.values
|
|
652
|
-
# @return [Array<Symbol>]
|
|
653
|
-
end
|
|
654
|
-
end
|
|
655
|
-
|
|
656
|
-
# @!method self.variants
|
|
657
|
-
# @return [Array(String, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1)]
|
|
658
|
-
end
|
|
659
|
-
|
|
660
|
-
# @!method self.variants
|
|
661
|
-
# @return [Array(Symbol, String, Hash{Symbol=>String, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1})]
|
|
662
|
-
|
|
663
|
-
define_sorbet_constant!(:Variants) do
|
|
664
|
-
T.type_alias do
|
|
665
|
-
T.any(
|
|
666
|
-
ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::TaggedSymbol,
|
|
667
|
-
String,
|
|
668
|
-
T::Hash[Symbol, ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::Variants]
|
|
669
|
-
)
|
|
670
|
-
end
|
|
671
|
-
end
|
|
672
|
-
|
|
673
|
-
# @!group
|
|
674
|
-
|
|
675
|
-
TEXT = :text
|
|
676
|
-
HTML = :html
|
|
677
|
-
|
|
678
|
-
# @!endgroup
|
|
679
|
-
|
|
680
|
-
# @type [ContextDev::Internal::Type::Converter]
|
|
681
|
-
HTMLExtractionRuleMap =
|
|
682
|
-
ContextDev::Internal::Type::HashOf[union: -> {
|
|
683
|
-
ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule
|
|
684
|
-
}]
|
|
685
|
-
end
|
|
686
|
-
|
|
687
|
-
# @see ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1#type
|
|
688
|
-
module Type
|
|
689
|
-
extend ContextDev::Internal::Type::Enum
|
|
690
|
-
|
|
691
|
-
ITEM = :item
|
|
692
|
-
LIST = :list
|
|
693
|
-
|
|
694
|
-
# @!method self.values
|
|
695
|
-
# @return [Array<Symbol>]
|
|
696
|
-
end
|
|
697
|
-
end
|
|
698
|
-
|
|
699
|
-
# @!method self.variants
|
|
700
|
-
# @return [Array(String, ContextDev::Models::WebWebScrapeHTMLParams::ExtractRule::UnionMember1)]
|
|
701
|
-
end
|
|
702
|
-
|
|
703
|
-
class Pdf < ContextDev::Internal::Type::BaseModel
|
|
704
|
-
# @!attribute end_
|
|
705
|
-
# Last 1-based PDF page to parse. When omitted, parsing ends at the last page.
|
|
706
|
-
# Must be greater than or equal to start when both are provided.
|
|
707
|
-
#
|
|
708
|
-
# @return [Integer, nil]
|
|
709
|
-
optional :end_, Integer, api_name: :end
|
|
710
|
-
|
|
711
|
-
# @!attribute ocr
|
|
712
|
-
# When true, OCR the selected PDF pages that have no usable text layer (scans),
|
|
713
|
-
# replacing each recovered page's text with the OCR result while pages with a real
|
|
714
|
-
# text layer keep it. Billed at 1 credit per page OCR actually recovered, on top
|
|
715
|
-
# of the base request cost. When false, no OCR runs.
|
|
716
|
-
#
|
|
717
|
-
# @return [Boolean, nil]
|
|
718
|
-
optional :ocr, ContextDev::Internal::Type::Boolean
|
|
719
|
-
|
|
720
|
-
# @!attribute should_parse
|
|
721
|
-
# When true, PDF URLs are fetched and parsed. When false, PDF URLs are skipped and
|
|
722
|
-
# a 400 PDF_SKIPPED is returned.
|
|
723
|
-
#
|
|
724
|
-
# @return [Boolean, nil]
|
|
725
|
-
optional :should_parse, ContextDev::Internal::Type::Boolean, api_name: :shouldParse
|
|
726
|
-
|
|
727
|
-
# @!attribute start
|
|
728
|
-
# First 1-based PDF page to parse. When omitted, parsing starts at the first page.
|
|
729
|
-
#
|
|
730
|
-
# @return [Integer, nil]
|
|
731
|
-
optional :start, Integer
|
|
732
|
-
|
|
733
|
-
# @!method initialize(end_: nil, ocr: nil, should_parse: nil, start: nil)
|
|
734
|
-
# Some parameter documentations has been truncated, see
|
|
735
|
-
# {ContextDev::Models::WebWebScrapeHTMLParams::Pdf} for more details.
|
|
736
|
-
#
|
|
737
|
-
# PDF parsing controls. Use start/end to limit text extraction and embedded-image
|
|
738
|
-
# detection/OCR to an inclusive 1-based page range.
|
|
739
|
-
#
|
|
740
|
-
# @param end_ [Integer] Last 1-based PDF page to parse. When omitted, parsing ends at the last page. Mus
|
|
741
|
-
#
|
|
742
|
-
# @param ocr [Boolean] When true, OCR the selected PDF pages that have no usable text layer (scans), re
|
|
743
|
-
#
|
|
744
|
-
# @param should_parse [Boolean] When true, PDF URLs are fetched and parsed. When false, PDF URLs are skipped and
|
|
745
|
-
#
|
|
746
|
-
# @param start [Integer] First 1-based PDF page to parse. When omitted, parsing starts at the first page.
|
|
747
|
-
end
|
|
748
|
-
|
|
749
|
-
class TimeoutOpts < ContextDev::Internal::Type::BaseModel
|
|
750
|
-
# @!attribute milliseconds
|
|
751
|
-
# Request deadline in milliseconds. Maximum: 300000 (5 minutes).
|
|
752
|
-
#
|
|
753
|
-
# @return [Integer]
|
|
754
|
-
required :milliseconds, Integer
|
|
755
|
-
|
|
756
|
-
# @!attribute behavior
|
|
757
|
-
# What to do at the deadline. "fail" returns 408 REQUEST_TIMEOUT without charging
|
|
758
|
-
# credits. "return-partial" returns usable results collected so far; if none are
|
|
759
|
-
# available, the request still fails without charging credits. Partial results are
|
|
760
|
-
# not cached as complete results. "return-partial" requires milliseconds of at
|
|
761
|
-
# least 5000.
|
|
762
|
-
#
|
|
763
|
-
# @return [Symbol, ContextDev::Models::WebWebScrapeHTMLParams::TimeoutOpts::Behavior, nil]
|
|
764
|
-
optional :behavior, enum: -> { ContextDev::WebWebScrapeHTMLParams::TimeoutOpts::Behavior }
|
|
765
|
-
|
|
766
|
-
# @!method initialize(milliseconds:, behavior: nil)
|
|
767
|
-
# Some parameter documentations has been truncated, see
|
|
768
|
-
# {ContextDev::Models::WebWebScrapeHTMLParams::TimeoutOpts} for more details.
|
|
769
|
-
#
|
|
770
|
-
# Optional request deadline and behavior on timeout. For GET requests, use
|
|
771
|
-
# timeoutOpts[milliseconds]=30000&timeoutOpts[behavior]=fail or a JSON-encoded
|
|
772
|
-
# timeoutOpts object.
|
|
773
|
-
#
|
|
774
|
-
# @param milliseconds [Integer] Request deadline in milliseconds. Maximum: 300000 (5 minutes).
|
|
775
|
-
#
|
|
776
|
-
# @param behavior [Symbol, ContextDev::Models::WebWebScrapeHTMLParams::TimeoutOpts::Behavior] What to do at the deadline. "fail" returns 408 REQUEST_TIMEOUT without charging
|
|
777
|
-
|
|
778
|
-
# What to do at the deadline. "fail" returns 408 REQUEST_TIMEOUT without charging
|
|
779
|
-
# credits. "return-partial" returns usable results collected so far; if none are
|
|
780
|
-
# available, the request still fails without charging credits. Partial results are
|
|
781
|
-
# not cached as complete results. "return-partial" requires milliseconds of at
|
|
782
|
-
# least 5000.
|
|
783
|
-
#
|
|
784
|
-
# @see ContextDev::Models::WebWebScrapeHTMLParams::TimeoutOpts#behavior
|
|
785
|
-
module Behavior
|
|
786
|
-
extend ContextDev::Internal::Type::Enum
|
|
787
|
-
|
|
788
|
-
FAIL = :fail
|
|
789
|
-
RETURN_PARTIAL = :"return-partial"
|
|
790
|
-
|
|
791
|
-
# @!method self.values
|
|
792
|
-
# @return [Array<Symbol>]
|
|
793
|
-
end
|
|
794
|
-
end
|
|
795
|
-
|
|
796
|
-
# Set to enabled to bypass shared caches and omit request and response content
|
|
797
|
-
# from retained usage logs. Asset uploads are skipped, so hosted image URLs are
|
|
798
|
-
# omitted. Requires zero data retention to be enabled for your organization
|
|
799
|
-
# (contact support@context.dev), otherwise the request fails with ZDR_NOT_ENABLED.
|
|
800
|
-
# Successful ZDR responses include X-Context-ZDR: true.
|
|
801
|
-
module Zdr
|
|
802
|
-
extend ContextDev::Internal::Type::Enum
|
|
803
|
-
|
|
804
|
-
ENABLED = :enabled
|
|
805
|
-
DISABLED = :disabled
|
|
806
|
-
|
|
807
|
-
# @!method self.values
|
|
808
|
-
# @return [Array<Symbol>]
|
|
809
|
-
end
|
|
810
|
-
end
|
|
811
|
-
end
|
|
812
|
-
end
|