context.dev 1.32.0 → 1.35.0

data/lib/context_dev/models/web_web_scrape_md_response.rb

@@ -10,6 +10,12 @@ module ContextDev
       #   @return [String]
       required :markdown, String
 
+      # @!attribute metadata
+      #   Metadata extracted from the scraped page HTML.
+      #
+      #   @return [ContextDev::Models::WebWebScrapeMdResponse::Metadata]
+      required :metadata, -> { ContextDev::Models::WebWebScrapeMdResponse::Metadata }
+
       # @!attribute success
       #   Indicates success
       #
@@ -29,18 +35,267 @@ module ContextDev
       #   @return [ContextDev::Models::WebWebScrapeMdResponse::KeyMetadata, nil]
       optional :key_metadata, -> { ContextDev::Models::WebWebScrapeMdResponse::KeyMetadata }
 
-      # @!method initialize(markdown:, success:, url:, key_metadata: nil)
+      # @!method initialize(markdown:, metadata:, success:, url:, key_metadata: nil)
       #   Some parameter documentations has been truncated, see
       #   {ContextDev::Models::WebWebScrapeMdResponse} for more details.
       #
       #   @param markdown [String] Page content converted to GitHub Flavored Markdown
       #
+      #   @param metadata [ContextDev::Models::WebWebScrapeMdResponse::Metadata] Metadata extracted from the scraped page HTML.
+      #
       #   @param success [Boolean, ContextDev::Models::WebWebScrapeMdResponse::Success] Indicates success
       #
       #   @param url [String] The URL that was scraped
       #
       #   @param key_metadata [ContextDev::Models::WebWebScrapeMdResponse::KeyMetadata] Metadata about the API key used for the request. Included in every response when
 
+      # @see ContextDev::Models::WebWebScrapeMdResponse#metadata
+      class Metadata < ContextDev::Internal::Type::BaseModel
+        # @!attribute final_url
+        #   Final URL scraped after redirects or scraper fallback, when known. Falls back to
+        #   sourceUrl when unavailable.
+        #
+        #   @return [String]
+        required :final_url, String, api_name: :finalUrl
+
+        # @!attribute source_url
+        #   Original URL requested by the caller.
+        #
+        #   @return [String]
+        required :source_url, String, api_name: :sourceUrl
+
+        # @!attribute additional_meta
+        #   Additional non-social meta tags not promoted to top-level metadata fields.
+        #
+        #   @return [Hash{Symbol=>String, Array<String>}, nil]
+        optional :additional_meta,
+                 -> { ContextDev::Internal::Type::HashOf[union: ContextDev::Models::WebWebScrapeMdResponse::Metadata::AdditionalMeta] },
+                 api_name: :additionalMeta
+
+        # @!attribute alternates
+        #   Resolved alternate links from link rel=alternate tags.
+        #
+        #   @return [Array<ContextDev::Models::WebWebScrapeMdResponse::Metadata::Alternate>, nil]
+        optional :alternates,
+                 -> { ContextDev::Internal::Type::ArrayOf[ContextDev::Models::WebWebScrapeMdResponse::Metadata::Alternate] }
+
+        # @!attribute author
+        #   Author metadata, when present.
+        #
+        #   @return [String, nil]
+        optional :author, String
+
+        # @!attribute canonical_url
+        #   Resolved canonical URL, when present.
+        #
+        #   @return [String, nil]
+        optional :canonical_url, String, api_name: :canonicalUrl
+
+        # @!attribute description
+        #   Best description extracted from standard, Open Graph, or Twitter metadata.
+        #
+        #   @return [String, nil]
+        optional :description, String
+
+        # @!attribute favicon
+        #   Resolved favicon URL, when present.
+        #
+        #   @return [String, nil]
+        optional :favicon, String
+
+        # @!attribute image
+        #   Primary resolved preview image from Open Graph, Twitter, or image metadata.
+        #
+        #   @return [String, nil]
+        optional :image, String
+
+        # @!attribute json_ld
+        #   JSON-LD structured data blocks parsed from the page.
+        #
+        #   @return [Array<Hash{Symbol=>Object}>, nil]
+        optional :json_ld,
+                 ContextDev::Internal::Type::ArrayOf[ContextDev::Internal::Type::HashOf[ContextDev::Internal::Type::Unknown]],
+                 api_name: :jsonLd
+
+        # @!attribute keywords
+        #   Keywords extracted from the page's keywords meta tag.
+        #
+        #   @return [Array<String>, nil]
+        optional :keywords, ContextDev::Internal::Type::ArrayOf[String]
+
+        # @!attribute language
+        #   Language extracted from html lang or language meta tags.
+        #
+        #   @return [String, nil]
+        optional :language, String
+
+        # @!attribute modified_time
+        #   Modified timestamp/date from page metadata, when present.
+        #
+        #   @return [String, nil]
+        optional :modified_time, String, api_name: :modifiedTime
+
+        # @!attribute open_graph
+        #   Open Graph metadata with the og: prefix removed and keys camel-cased.
+        #
+        #   @return [Hash{Symbol=>String, Array<String>}, nil]
+        optional :open_graph,
+                 -> { ContextDev::Internal::Type::HashOf[union: ContextDev::Models::WebWebScrapeMdResponse::Metadata::OpenGraph] },
+                 api_name: :openGraph
+
+        # @!attribute published_time
+        #   Published timestamp/date from page metadata, when present.
+        #
+        #   @return [String, nil]
+        optional :published_time, String, api_name: :publishedTime
+
+        # @!attribute robots
+        #   Robots meta directive, when present.
+        #
+        #   @return [String, nil]
+        optional :robots, String
+
+        # @!attribute site_name
+        #   Site or application name from page metadata.
+        #
+        #   @return [String, nil]
+        optional :site_name, String, api_name: :siteName
+
+        # @!attribute title
+        #   Best title extracted from the page.
+        #
+        #   @return [String, nil]
+        optional :title, String
+
+        # @!attribute twitter
+        #   Twitter card metadata with the twitter: prefix removed and keys camel-cased.
+        #
+        #   @return [Hash{Symbol=>String, Array<String>}, nil]
+        optional :twitter,
+                 -> { ContextDev::Internal::Type::HashOf[union: ContextDev::Models::WebWebScrapeMdResponse::Metadata::Twitter] }
+
+        # @!method initialize(final_url:, source_url:, additional_meta: nil, alternates: nil, author: nil, canonical_url: nil, description: nil, favicon: nil, image: nil, json_ld: nil, keywords: nil, language: nil, modified_time: nil, open_graph: nil, published_time: nil, robots: nil, site_name: nil, title: nil, twitter: nil)
+        #   Some parameter documentations has been truncated, see
+        #   {ContextDev::Models::WebWebScrapeMdResponse::Metadata} for more details.
+        #
+        #   Metadata extracted from the scraped page HTML.
+        #
+        #   @param final_url [String] Final URL scraped after redirects or scraper fallback, when known. Falls back to
+        #
+        #   @param source_url [String] Original URL requested by the caller.
+        #
+        #   @param additional_meta [Hash{Symbol=>String, Array<String>}] Additional non-social meta tags not promoted to top-level metadata fields.
+        #
+        #   @param alternates [Array<ContextDev::Models::WebWebScrapeMdResponse::Metadata::Alternate>] Resolved alternate links from link rel=alternate tags.
+        #
+        #   @param author [String] Author metadata, when present.
+        #
+        #   @param canonical_url [String] Resolved canonical URL, when present.
+        #
+        #   @param description [String] Best description extracted from standard, Open Graph, or Twitter metadata.
+        #
+        #   @param favicon [String] Resolved favicon URL, when present.
+        #
+        #   @param image [String] Primary resolved preview image from Open Graph, Twitter, or image metadata.
+        #
+        #   @param json_ld [Array<Hash{Symbol=>Object}>] JSON-LD structured data blocks parsed from the page.
+        #
+        #   @param keywords [Array<String>] Keywords extracted from the page's keywords meta tag.
+        #
+        #   @param language [String] Language extracted from html lang or language meta tags.
+        #
+        #   @param modified_time [String] Modified timestamp/date from page metadata, when present.
+        #
+        #   @param open_graph [Hash{Symbol=>String, Array<String>}] Open Graph metadata with the og: prefix removed and keys camel-cased.
+        #
+        #   @param published_time [String] Published timestamp/date from page metadata, when present.
+        #
+        #   @param robots [String] Robots meta directive, when present.
+        #
+        #   @param site_name [String] Site or application name from page metadata.
+        #
+        #   @param title [String] Best title extracted from the page.
+        #
+        #   @param twitter [Hash{Symbol=>String, Array<String>}] Twitter card metadata with the twitter: prefix removed and keys camel-cased.
+
+        module AdditionalMeta
+          extend ContextDev::Internal::Type::Union
+
+          variant String
+
+          variant -> { ContextDev::Models::WebWebScrapeMdResponse::Metadata::AdditionalMeta::StringArray }
+
+          # @!method self.variants
+          #   @return [Array(String, Array<String>)]
+
+          # @type [ContextDev::Internal::Type::Converter]
+          StringArray = ContextDev::Internal::Type::ArrayOf[String]
+        end
+
+        class Alternate < ContextDev::Internal::Type::BaseModel
+          # @!attribute href
+          #   Resolved alternate URL.
+          #
+          #   @return [String]
+          required :href, String
+
+          # @!attribute hreflang
+          #   Language or locale for the alternate URL, when present.
+          #
+          #   @return [String, nil]
+          optional :hreflang, String
+
+          # @!attribute title
+          #   Alternate resource title, when present.
+          #
+          #   @return [String, nil]
+          optional :title, String
+
+          # @!attribute type
+          #   Alternate resource MIME type, when present.
+          #
+          #   @return [String, nil]
+          optional :type, String
+
+          # @!method initialize(href:, hreflang: nil, title: nil, type: nil)
+          #   @param href [String] Resolved alternate URL.
+          #
+          #   @param hreflang [String] Language or locale for the alternate URL, when present.
+          #
+          #   @param title [String] Alternate resource title, when present.
+          #
+          #   @param type [String] Alternate resource MIME type, when present.
+        end
+
+        module OpenGraph
+          extend ContextDev::Internal::Type::Union
+
+          variant String
+
+          variant -> { ContextDev::Models::WebWebScrapeMdResponse::Metadata::OpenGraph::StringArray }
+
+          # @!method self.variants
+          #   @return [Array(String, Array<String>)]
+
+          # @type [ContextDev::Internal::Type::Converter]
+          StringArray = ContextDev::Internal::Type::ArrayOf[String]
+        end
+
+        module Twitter
+          extend ContextDev::Internal::Type::Union
+
+          variant String
+
+          variant -> { ContextDev::Models::WebWebScrapeMdResponse::Metadata::Twitter::StringArray }
+
+          # @!method self.variants
+          #   @return [Array(String, Array<String>)]
+
+          # @type [ContextDev::Internal::Type::Converter]
+          StringArray = ContextDev::Internal::Type::ArrayOf[String]
+        end
+      end
+
       # Indicates success
       #
       # @see ContextDev::Models::WebWebScrapeMdResponse#success

data/lib/context_dev/resources/web.rb

@@ -163,7 +163,7 @@ module ContextDev
       #
       # Capture a screenshot of a website.
       #
-      # @overload screenshot(direct_url: nil, domain: nil, full_screenshot: nil, handle_cookie_popup: nil, max_age_ms: nil, page: nil, timeout_ms: nil, viewport: nil, wait_for_ms: nil, request_options: {})
+      # @overload screenshot(direct_url: nil, domain: nil, full_screenshot: nil, handle_cookie_popup: nil, max_age_ms: nil, page: nil, scroll_offset: nil, timeout_ms: nil, viewport: nil, wait_for_ms: nil, request_options: {})
       #
       # @param direct_url [String] A specific URL to screenshot directly, bypassing domain resolution (e.g., 'https
       #
@@ -177,6 +177,8 @@ module ContextDev
       #
       # @param page [Symbol, ContextDev::Models::WebScreenshotParams::Page] Optional parameter to specify which page type to screenshot. If provided, the sy
       #
+      # @param scroll_offset [Integer] Optional vertical scroll offset in pixels for capturing a long page in viewport-
+      #
       # @param timeout_ms [Integer] Optional timeout in milliseconds for the request. If the request takes longer th
       #
       # @param viewport [ContextDev::Models::WebScreenshotParams::Viewport] Optional browser viewport dimensions for the screenshot. Defaults to 1920x1080.
@@ -199,6 +201,7 @@ module ContextDev
             full_screenshot: "fullScreenshot",
             handle_cookie_popup: "handleCookiePopup",
             max_age_ms: "maxAgeMs",
+            scroll_offset: "scrollOffset",
             timeout_ms: "timeoutMS",
             wait_for_ms: "waitForMs"
           ),

data/lib/context_dev/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module ContextDev
-  VERSION = "1.32.0"
+  VERSION = "1.35.0"
 end

data/rbi/context_dev/models/web_screenshot_params.rbi

@@ -89,6 +89,18 @@ module ContextDev
       sig { params(page: ContextDev::WebScreenshotParams::Page::OrSymbol).void }
       attr_writer :page
 
+      # Optional vertical scroll offset in pixels for capturing a long page in
+      # viewport-sized chunks. When provided, the full page is captured once and the
+      # returned image is the viewport-sized slice that begins at this Y offset (e.g.
+      # request scrollOffset=0, then 1080, then 2160 to walk a 1920x1080 landing page
+      # top to bottom). The final slice may be shorter than the viewport height. Takes
+      # precedence over fullScreenshot. Max: 100000.
+      sig { returns(T.nilable(Integer)) }
+      attr_reader :scroll_offset
+
+      sig { params(scroll_offset: Integer).void }
+      attr_writer :scroll_offset
+
       # Optional timeout in milliseconds for the request. If the request takes longer
       # than this value, it will be aborted with a 408 status code. Maximum allowed
       # value is 300000ms (5 minutes).
@@ -126,6 +138,7 @@ module ContextDev
             ContextDev::WebScreenshotParams::HandleCookiePopup::OrSymbol,
           max_age_ms: Integer,
           page: ContextDev::WebScreenshotParams::Page::OrSymbol,
+          scroll_offset: Integer,
           timeout_ms: Integer,
           viewport: ContextDev::WebScreenshotParams::Viewport::OrHash,
           wait_for_ms: Integer,
@@ -159,6 +172,13 @@ module ContextDev
         # provided, screenshots the main domain landing page. Only applicable when using
         # 'domain', not 'directUrl'.
         page: nil,
+        # Optional vertical scroll offset in pixels for capturing a long page in
+        # viewport-sized chunks. When provided, the full page is captured once and the
+        # returned image is the viewport-sized slice that begins at this Y offset (e.g.
+        # request scrollOffset=0, then 1080, then 2160 to walk a 1920x1080 landing page
+        # top to bottom). The final slice may be shorter than the viewport height. Takes
+        # precedence over fullScreenshot. Max: 100000.
+        scroll_offset: nil,
         # Optional timeout in milliseconds for the request. If the request takes longer
         # than this value, it will be aborted with a 408 status code. Maximum allowed
         # value is 300000ms (5 minutes).
@@ -184,6 +204,7 @@ module ContextDev
               ContextDev::WebScreenshotParams::HandleCookiePopup::OrSymbol,
             max_age_ms: Integer,
             page: ContextDev::WebScreenshotParams::Page::OrSymbol,
+            scroll_offset: Integer,
             timeout_ms: Integer,
             viewport: ContextDev::WebScreenshotParams::Viewport,
             wait_for_ms: Integer,