context.dev 2.17.0 → 2.18.0

data/rbi/context_dev/models/web_web_scrape_html_params.rbi

@@ -58,6 +58,42 @@ module ContextDev
       sig { returns(T.nilable(T::Array[String])) }
       attr_accessor :exclude_selectors
 
+      # Optional CSS extraction rules applied to the returned HTML after selector and
+      # main-content filtering. Use selector strings ("h1", "a@href") or objects with
+      # selector, type (item or list), and output (text, html, @attribute, or nested
+      # rules). Text whitespace is normalized; html includes the matched element;
+      # attributes are returned as written. Missing items are null and missing lists are
+      # empty. CSS only; XPath is not supported. Maximum: 100 fields across 5 levels.
+      # Send a JSON-encoded string in the extractRules query parameter.
+      sig do
+        returns(
+          T.nilable(
+            T::Hash[
+              Symbol,
+              T.any(
+                String,
+                ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1
+              )
+            ]
+          )
+        )
+      end
+      attr_reader :extract_rules
+
+      sig do
+        params(
+          extract_rules:
+            T::Hash[
+              Symbol,
+              T.any(
+                String,
+                ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::OrHash
+              )
+            ]
+        ).void
+      end
+      attr_writer :extract_rules
+
       # Optional outbound HTTP headers forwarded only to the target URL, sent as
       # deep-object query params such as headers[X-Custom]=value. When provided, caching
       # is bypassed: the result is neither read from nor written to cache.
@@ -142,9 +178,10 @@ module ContextDev
       attr_accessor :wait_for_ms
 
       # Set to enabled to bypass shared caches and omit request and response content
-      # from retained usage logs. Requires zero data retention to be enabled for your
-      # organization (contact support@context.dev), otherwise the request fails with
-      # ZDR_NOT_ENABLED. Successful ZDR responses include X-Context-ZDR: true.
+      # from retained usage logs. Asset uploads are skipped, so hosted image URLs are
+      # omitted. Requires zero data retention to be enabled for your organization
+      # (contact support@context.dev), otherwise the request fails with ZDR_NOT_ENABLED.
+      # Successful ZDR responses include X-Context-ZDR: true.
       sig do
         returns(T.nilable(ContextDev::WebWebScrapeHTMLParams::Zdr::OrSymbol))
       end
@@ -170,6 +207,14 @@ module ContextDev
             ),
           country: ContextDev::WebWebScrapeHTMLParams::Country::OrSymbol,
           exclude_selectors: T.nilable(T::Array[String]),
+          extract_rules:
+            T::Hash[
+              Symbol,
+              T.any(
+                String,
+                ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::OrHash
+              )
+            ],
           headers: T::Hash[Symbol, String],
           include_frames: T::Boolean,
           include_selectors: T.nilable(T::Array[String]),
@@ -198,6 +243,14 @@ module ContextDev
         # Exclusion takes precedence: an element matching both is removed. Examples:
         # "nav", "footer", ".ad-banner", "[aria-hidden=true]".
         exclude_selectors: nil,
+        # Optional CSS extraction rules applied to the returned HTML after selector and
+        # main-content filtering. Use selector strings ("h1", "a@href") or objects with
+        # selector, type (item or list), and output (text, html, @attribute, or nested
+        # rules). Text whitespace is normalized; html includes the matched element;
+        # attributes are returned as written. Missing items are null and missing lists are
+        # empty. CSS only; XPath is not supported. Maximum: 100 fields across 5 levels.
+        # Send a JSON-encoded string in the extractRules query parameter.
+        extract_rules: nil,
         # Optional outbound HTTP headers forwarded only to the target URL, sent as
         # deep-object query params such as headers[X-Custom]=value. When provided, caching
         # is bypassed: the result is neither read from nor written to cache.
@@ -235,9 +288,10 @@ module ContextDev
         # TIMEOUT_TOO_SHORT_FOR_WAIT.
         wait_for_ms: nil,
         # Set to enabled to bypass shared caches and omit request and response content
-        # from retained usage logs. Requires zero data retention to be enabled for your
-        # organization (contact support@context.dev), otherwise the request fails with
-        # ZDR_NOT_ENABLED. Successful ZDR responses include X-Context-ZDR: true.
+        # from retained usage logs. Asset uploads are skipped, so hosted image URLs are
+        # omitted. Requires zero data retention to be enabled for your organization
+        # (contact support@context.dev), otherwise the request fails with ZDR_NOT_ENABLED.
+        # Successful ZDR responses include X-Context-ZDR: true.
         zdr: nil,
         request_options: {}
       )
@@ -259,6 +313,14 @@ module ContextDev
               ),
             country: ContextDev::WebWebScrapeHTMLParams::Country::OrSymbol,
             exclude_selectors: T.nilable(T::Array[String]),
+            extract_rules:
+              T::Hash[
+                Symbol,
+                T.any(
+                  String,
+                  ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1
+                )
+              ],
             headers: T::Hash[Symbol, String],
             include_frames: T::Boolean,
             include_selectors: T.nilable(T::Array[String]),
@@ -991,6 +1053,416 @@ module ContextDev
         end
       end
 
+      module ExtractRule
+        extend ContextDev::Internal::Type::Union
+
+        Variants =
+          T.type_alias do
+            T.any(
+              String,
+              ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1
+            )
+          end
+
+        class UnionMember1 < ContextDev::Internal::Type::BaseModel
+          OrHash =
+            T.type_alias do
+              T.any(
+                ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1,
+                ContextDev::Internal::AnyHash
+              )
+            end
+
+          sig { returns(String) }
+          attr_accessor :selector
+
+          sig do
+            returns(
+              T.nilable(
+                T.any(
+                  ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::OrSymbol,
+                  String,
+                  T::Hash[
+                    Symbol,
+                    T.any(
+                      String,
+                      ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1
+                    )
+                  ]
+                )
+              )
+            )
+          end
+          attr_reader :output
+
+          sig do
+            params(
+              output:
+                T.any(
+                  ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::OrSymbol,
+                  String,
+                  T::Hash[
+                    Symbol,
+                    T.any(
+                      String,
+                      ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::OrHash
+                    )
+                  ]
+                )
+            ).void
+          end
+          attr_writer :output
+
+          sig do
+            returns(
+              T.nilable(
+                ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Type::OrSymbol
+              )
+            )
+          end
+          attr_reader :type
+
+          sig do
+            params(
+              type:
+                ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Type::OrSymbol
+            ).void
+          end
+          attr_writer :type
+
+          sig do
+            params(
+              selector: String,
+              output:
+                T.any(
+                  ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::OrSymbol,
+                  String,
+                  T::Hash[
+                    Symbol,
+                    T.any(
+                      String,
+                      ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::OrHash
+                    )
+                  ]
+                ),
+              type:
+                ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Type::OrSymbol
+            ).returns(T.attached_class)
+          end
+          def self.new(selector:, output: nil, type: nil)
+          end
+
+          sig do
+            override.returns(
+              {
+                selector: String,
+                output:
+                  T.any(
+                    ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::OrSymbol,
+                    String,
+                    T::Hash[
+                      Symbol,
+                      T.any(
+                        String,
+                        ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1
+                      )
+                    ]
+                  ),
+                type:
+                  ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Type::OrSymbol
+              }
+            )
+          end
+          def to_hash
+          end
+
+          module Output
+            extend ContextDev::Internal::Type::Union
+
+            Variants =
+              T.type_alias do
+                T.any(
+                  ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::TaggedSymbol,
+                  String,
+                  T::Hash[
+                    Symbol,
+                    ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::Variants
+                  ]
+                )
+              end
+
+            module HTMLExtractionRule
+              extend ContextDev::Internal::Type::Union
+
+              Variants =
+                T.type_alias do
+                  T.any(
+                    String,
+                    ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1
+                  )
+                end
+
+              class UnionMember1 < ContextDev::Internal::Type::BaseModel
+                OrHash =
+                  T.type_alias do
+                    T.any(
+                      ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1,
+                      ContextDev::Internal::AnyHash
+                    )
+                  end
+
+                sig { returns(String) }
+                attr_accessor :selector
+
+                sig do
+                  returns(
+                    T.nilable(
+                      T.any(
+                        ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output::OrSymbol,
+                        String,
+                        T.anything
+                      )
+                    )
+                  )
+                end
+                attr_reader :output
+
+                sig do
+                  params(
+                    output:
+                      T.any(
+                        ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output::OrSymbol,
+                        String,
+                        T.anything
+                      )
+                  ).void
+                end
+                attr_writer :output
+
+                sig do
+                  returns(
+                    T.nilable(
+                      ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Type::OrSymbol
+                    )
+                  )
+                end
+                attr_reader :type
+
+                sig do
+                  params(
+                    type:
+                      ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Type::OrSymbol
+                  ).void
+                end
+                attr_writer :type
+
+                sig do
+                  params(
+                    selector: String,
+                    output:
+                      T.any(
+                        ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output::OrSymbol,
+                        String,
+                        T.anything
+                      ),
+                    type:
+                      ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Type::OrSymbol
+                  ).returns(T.attached_class)
+                end
+                def self.new(selector:, output: nil, type: nil)
+                end
+
+                sig do
+                  override.returns(
+                    {
+                      selector: String,
+                      output:
+                        T.any(
+                          ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output::OrSymbol,
+                          String,
+                          T.anything
+                        ),
+                      type:
+                        ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Type::OrSymbol
+                    }
+                  )
+                end
+                def to_hash
+                end
+
+                module Output
+                  extend ContextDev::Internal::Type::Union
+
+                  Variants =
+                    T.type_alias do
+                      T.any(
+                        ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output::TaggedSymbol,
+                        String,
+                        T.anything
+                      )
+                    end
+
+                  sig do
+                    override.returns(
+                      T::Array[
+                        ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output::Variants
+                      ]
+                    )
+                  end
+                  def self.variants
+                  end
+
+                  TaggedSymbol =
+                    T.type_alias do
+                      T.all(
+                        Symbol,
+                        ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output
+                      )
+                    end
+                  OrSymbol = T.type_alias { T.any(Symbol, String) }
+
+                  TEXT =
+                    T.let(
+                      :text,
+                      ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output::TaggedSymbol
+                    )
+                  HTML =
+                    T.let(
+                      :html,
+                      ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Output::TaggedSymbol
+                    )
+                end
+
+                module Type
+                  extend ContextDev::Internal::Type::Enum
+
+                  TaggedSymbol =
+                    T.type_alias do
+                      T.all(
+                        Symbol,
+                        ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Type
+                      )
+                    end
+                  OrSymbol = T.type_alias { T.any(Symbol, String) }
+
+                  ITEM =
+                    T.let(
+                      :item,
+                      ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Type::TaggedSymbol
+                    )
+                  LIST =
+                    T.let(
+                      :list,
+                      ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Type::TaggedSymbol
+                    )
+
+                  sig do
+                    override.returns(
+                      T::Array[
+                        ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::UnionMember1::Type::TaggedSymbol
+                      ]
+                    )
+                  end
+                  def self.values
+                  end
+                end
+              end
+
+              sig do
+                override.returns(
+                  T::Array[
+                    ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule::Variants
+                  ]
+                )
+              end
+              def self.variants
+              end
+            end
+
+            sig do
+              override.returns(
+                T::Array[
+                  ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::Variants
+                ]
+              )
+            end
+            def self.variants
+            end
+
+            TaggedSymbol =
+              T.type_alias do
+                T.all(
+                  Symbol,
+                  ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output
+                )
+              end
+            OrSymbol = T.type_alias { T.any(Symbol, String) }
+
+            TEXT =
+              T.let(
+                :text,
+                ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::TaggedSymbol
+              )
+            HTML =
+              T.let(
+                :html,
+                ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::TaggedSymbol
+              )
+
+            HTMLExtractionRuleMap =
+              T.let(
+                ContextDev::Internal::Type::HashOf[
+                  union:
+                    ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Output::HTMLExtractionRule
+                ],
+                ContextDev::Internal::Type::Converter
+              )
+          end
+
+          module Type
+            extend ContextDev::Internal::Type::Enum
+
+            TaggedSymbol =
+              T.type_alias do
+                T.all(
+                  Symbol,
+                  ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Type
+                )
+              end
+            OrSymbol = T.type_alias { T.any(Symbol, String) }
+
+            ITEM =
+              T.let(
+                :item,
+                ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Type::TaggedSymbol
+              )
+            LIST =
+              T.let(
+                :list,
+                ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Type::TaggedSymbol
+              )
+
+            sig do
+              override.returns(
+                T::Array[
+                  ContextDev::WebWebScrapeHTMLParams::ExtractRule::UnionMember1::Type::TaggedSymbol
+                ]
+              )
+            end
+            def self.values
+            end
+          end
+        end
+
+        sig do
+          override.returns(
+            T::Array[ContextDev::WebWebScrapeHTMLParams::ExtractRule::Variants]
+          )
+        end
+        def self.variants
+        end
+      end
+
       class Pdf < ContextDev::Internal::Type::BaseModel
         OrHash =
           T.type_alias do
@@ -1091,7 +1563,7 @@ module ContextDev
         # credits. "return-partial" returns usable results collected so far; if none are
         # available, the request still fails without charging credits. Partial results are
         # not cached as complete results. "return-partial" requires milliseconds of at
-        # least 15000.
+        # least 5000.
         sig do
           returns(
             T.nilable(
@@ -1126,7 +1598,7 @@ module ContextDev
           # credits. "return-partial" returns usable results collected so far; if none are
           # available, the request still fails without charging credits. Partial results are
           # not cached as complete results. "return-partial" requires milliseconds of at
-          # least 15000.
+          # least 5000.
           behavior: nil
         )
         end
@@ -1147,7 +1619,7 @@ module ContextDev
         # credits. "return-partial" returns usable results collected so far; if none are
         # available, the request still fails without charging credits. Partial results are
         # not cached as complete results. "return-partial" requires milliseconds of at
-        # least 15000.
+        # least 5000.
         module Behavior
           extend ContextDev::Internal::Type::Enum
 
@@ -1184,9 +1656,10 @@ module ContextDev
       end
 
       # Set to enabled to bypass shared caches and omit request and response content
-      # from retained usage logs. Requires zero data retention to be enabled for your
-      # organization (contact support@context.dev), otherwise the request fails with
-      # ZDR_NOT_ENABLED. Successful ZDR responses include X-Context-ZDR: true.
+      # from retained usage logs. Asset uploads are skipped, so hosted image URLs are
+      # omitted. Requires zero data retention to be enabled for your organization
+      # (contact support@context.dev), otherwise the request fails with ZDR_NOT_ENABLED.
+      # Successful ZDR responses include X-Context-ZDR: true.
       module Zdr
         extend ContextDev::Internal::Type::Enum
 

data/rbi/context_dev/models/web_web_scrape_html_response.rbi

@@ -115,6 +115,37 @@ module ContextDev
       sig { params(actions_html_stale: T::Boolean).void }
       attr_writer :actions_html_stale
 
+      # Present only when extractRules is supplied. Keys match the requested fields.
+      # Values are normalized text, raw attribute strings, outer HTML, nested objects,
+      # or lists. Missing items are null; lists with no matches are empty. Rules run on
+      # the returned HTML after filtering.
+      sig do
+        returns(
+          T.nilable(
+            T::Hash[
+              Symbol,
+              T.nilable(
+                ContextDev::Models::WebWebScrapeHTMLResponse::Extracted::Variants
+              )
+            ]
+          )
+        )
+      end
+      attr_reader :extracted
+
+      sig do
+        params(
+          extracted:
+            T::Hash[
+              Symbol,
+              T.nilable(
+                ContextDev::Models::WebWebScrapeHTMLResponse::Extracted::Variants
+              )
+            ]
+        ).void
+      end
+      attr_writer :extracted
+
       # Credit usage, included whenever a valid API key is provided.
       sig do
         returns(
@@ -150,6 +181,13 @@ module ContextDev
               ContextDev::Models::WebWebScrapeHTMLResponse::ActionsApplied::OrHash
             ],
           actions_html_stale: T::Boolean,
+          extracted:
+            T::Hash[
+              Symbol,
+              T.nilable(
+                ContextDev::Models::WebWebScrapeHTMLResponse::Extracted::Variants
+              )
+            ],
           key_metadata:
             ContextDev::Models::WebWebScrapeHTMLResponse::KeyMetadata::OrHash
         ).returns(T.attached_class)
@@ -188,6 +226,11 @@ module ContextDev
         # True when an action was applied but the returned content could not be refreshed
         # afterward.
         actions_html_stale: nil,
+        # Present only when extractRules is supplied. Keys match the requested fields.
+        # Values are normalized text, raw attribute strings, outer HTML, nested objects,
+        # or lists. Missing items are null; lists with no matches are empty. Rules run on
+        # the returned HTML after filtering.
+        extracted: nil,
         # Credit usage, included whenever a valid API key is provided.
         key_metadata: nil
       )
@@ -213,6 +256,13 @@ module ContextDev
                 ContextDev::Models::WebWebScrapeHTMLResponse::ActionsApplied
               ],
             actions_html_stale: T::Boolean,
+            extracted:
+              T::Hash[
+                Symbol,
+                T.nilable(
+                  ContextDev::Models::WebWebScrapeHTMLResponse::Extracted::Variants
+                )
+              ],
             key_metadata:
               ContextDev::Models::WebWebScrapeHTMLResponse::KeyMetadata
           }
@@ -1140,6 +1190,59 @@ module ContextDev
         end
       end
 
+      module Extracted
+        extend ContextDev::Internal::Type::Union
+
+        Variants =
+          T.type_alias do
+            T.any(
+              String,
+              T.anything,
+              T::Array[
+                T.nilable(
+                  ContextDev::Models::WebWebScrapeHTMLResponse::Extracted::UnionMember2::Variants
+                )
+              ]
+            )
+          end
+
+        module UnionMember2
+          extend ContextDev::Internal::Type::Union
+
+          Variants = T.type_alias { T.any(String, T.anything) }
+
+          sig do
+            override.returns(
+              T::Array[
+                ContextDev::Models::WebWebScrapeHTMLResponse::Extracted::UnionMember2::Variants
+              ]
+            )
+          end
+          def self.variants
+          end
+        end
+
+        sig do
+          override.returns(
+            T::Array[
+              ContextDev::Models::WebWebScrapeHTMLResponse::Extracted::Variants
+            ]
+          )
+        end
+        def self.variants
+        end
+
+        UnionMember2Array =
+          T.let(
+            ContextDev::Internal::Type::ArrayOf[
+              union:
+                ContextDev::Models::WebWebScrapeHTMLResponse::Extracted::UnionMember2,
+              nil?: true
+            ],
+            ContextDev::Internal::Type::Converter
+          )
+      end
+
       class KeyMetadata < ContextDev::Internal::Type::BaseModel
         OrHash =
           T.type_alias do