mindee 4.8.0 → 4.10.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6bbe7fd9450d6e87f07b3b96c636ef8715200370503eb97d65db46225b6ccf8e
-  data.tar.gz: 36b67f8ca703982d40048e4d7e9a3b20e4204c7a36386028b3bd11da79560a20
+  metadata.gz: a0ab40fd2712787376dff65bab0aecb06a9ecd0337aa5ffe8b0d2ed115fac2a6
+  data.tar.gz: 661847259c9047171d1fe198a20459afd5545b3e298b253252a8ad59f12ac3c0
 SHA512:
-  metadata.gz: '08dd6e3e40ea73f874647ab537f2539342aca107bb1d324c04fa8b8de1fa803f5893117bcaf0e840dd6210b55e6e7c163897bb6cba6a36389b44daeab6a886e2'
-  data.tar.gz: 6028ec9f0d38e9e7840c33c98a517d8db02aef34058b8ae517c560cdc0b88524a0e2ab4a472dfe4e1f12ef027fce5bd2153cebd628e2c66f29ac3c8733c1cec3
+  metadata.gz: 9d33a23521dcc763584f92f22ca11b02dbea63d016c4f39173a15fe865805f362801bdd7f205761030524e70b65ff4f73d90d742212a49f3f21b6cf8979c5556
+  data.tar.gz: 19dee30dd72aa0adc05f4754f8a62b7c3299bdca7de25fda9aed315c485a11aa4878d28f3f086d9a6fed81371a599fa5c9ce3b4d883e835edd9ced77dce9b4c2

data/CHANGELOG.md

@@ -1,5 +1,15 @@
 # Mindee Ruby API Library Changelog
 
+## v4.10.0 - 2025-12-19
+### Changes
+:sparkles: add support for dataschema parameter
+
+
+## v4.9.0 - 2025-12-02
+### Changes
+* :sparkles: add support for text_context parameter
+
+
 ## v4.8.0 - 2025-11-18
 ### Changes
 * :sparkles: add support for better errors

data/lib/mindee/http/mindee_api_v2.rb

@@ -111,8 +111,26 @@ module Mindee
         poll("#{@settings.base_url}/inferences/#{queue_id}")
       end
 
+      # Handle parameters for the enqueue form
+      # @param form_data [Array] Array of form fields
+      # @param params [Input::InferenceParameters] Inference options.
+      def enqueue_form_options(form_data, params)
+        # deal with optional features
+        form_data.push(['rag', params.rag.to_s]) unless params.rag.nil?
+        form_data.push(['raw_text', params.raw_text.to_s]) unless params.raw_text.nil?
+        form_data.push(['polygon', params.polygon.to_s]) unless params.polygon.nil?
+        form_data.push(['confidence', params.confidence.to_s]) unless params.confidence.nil?
+        form_data.push ['file_alias', params.file_alias] if params.file_alias
+        form_data.push ['text_context', params.text_context] if params.text_context
+        form_data.push ['data_schema', params.data_schema.to_s] if params.data_schema
+        unless params.webhook_ids.nil? || params.webhook_ids.empty?
+          form_data.push ['webhook_ids', params.webhook_ids.join(',')]
+        end
+        form_data
+      end
+
       # @param input_source [Mindee::Input::Source::LocalInputSource, Mindee::Input::Source::URLInputSource]
-      # @param params [Input::InferenceParameters] Parse options.
+      # @param params [Input::InferenceParameters] Inference options.
       # @return [Net::HTTPResponse, nil]
       def enqueue(input_source, params)
         uri = URI("#{@settings.base_url}/inferences/enqueue")
@@ -125,16 +143,9 @@ module Mindee
                     end
         form_data.push(['model_id', params.model_id])
 
-        # deal with optional features
-        form_data.push(['rag', params.rag.to_s]) unless params.rag.nil?
-        form_data.push(['raw_text', params.raw_text.to_s]) unless params.raw_text.nil?
-        form_data.push(['polygon', params.polygon.to_s]) unless params.polygon.nil?
-        form_data.push(['confidence', params.confidence.to_s]) unless params.confidence.nil?
+        # deal with other parameters
+        form_data = enqueue_form_options(form_data, params)
 
-        form_data.push ['file_alias', params.file_alias] if params.file_alias
-        unless params.webhook_ids.nil? || params.webhook_ids.empty?
-          form_data.push ['webhook_ids', params.webhook_ids.join(',')]
-        end
         headers = {
           'Authorization' => @settings.api_key,
           'User-Agent' => @settings.user_agent,

data/lib/mindee/input/data_schema.rb

@@ -0,0 +1,126 @@
+# frozen_string_literal: true
+
+module Mindee
+  module Input
+    # Data Schema Field.
+    class DataSchemaField
+      # @return [String] Display name for the field, also impacts inference results.
+      attr_reader :title
+      # @return [String] Name of the field in the data schema.
+      attr_reader :name
+      # @return [Boolean] Whether this field can contain multiple values.
+      attr_reader :is_array
+      # @return [String] Data type of the field.
+      attr_reader :type
+      # @return [Array<String>, nil] Allowed values when type is `classification`. Leave empty for other types.
+      attr_reader :classification_values
+      # @return [Boolean, nil] Whether to remove duplicate values in the array.
+      # Only applicable if `is_array` is True.
+      attr_reader :unique_values
+      # @return [String, nil] Detailed description of what this field represents.
+      attr_reader :description
+      # @return [String, nil] Optional extraction guidelines.
+      attr_reader :guidelines
+      # @return [Array<Hash>, nil] Nested fields.
+      attr_reader :nested_fields
+
+      # @param field [Hash]
+      def initialize(field)
+        field.transform_keys!(&:to_sym)
+        @name = field[:name]
+        @title = field[:title]
+        @is_array = field[:is_array]
+        @type = field[:type]
+        @classification_values = field[:classification_values]
+        @unique_values = field[:unique_values]
+        @description = field[:description]
+        @guidelines = field[:guidelines]
+        @nested_fields = field[:nested_fields]
+      end
+
+      # @return [Hash]
+      def to_hash
+        out = {
+          name: @name,
+          title: @title,
+          is_array: @is_array,
+          type: @type,
+        } # @type var out: Hash[Symbol, untyped]
+        out[:classification_values] = @classification_values unless @classification_values.nil?
+        out[:unique_values] = @unique_values unless @unique_values.nil?
+        out[:description] = @description unless @description.nil?
+        out[:guidelines] = @guidelines unless @guidelines.nil?
+        out[:nested_fields] = @nested_fields unless @nested_fields.nil?
+        out
+      end
+
+      # @return [String]
+      def to_s
+        to_hash.to_json
+      end
+    end
+
+    # The structure to completely replace the data schema of the model.
+    class DataSchemaReplace
+      # @return [Array<DataSchemaField>] Subfields when type is `nested_object`. Leave empty for other types.
+      attr_reader :fields
+
+      # @param data_schema_replace [Hash]
+      def initialize(data_schema_replace)
+        data_schema_replace.transform_keys!(&:to_sym)
+        fields_list = data_schema_replace[:fields]
+        raise Mindee::Errors::MindeeError, 'Invalid Data Schema provided.' if fields_list.nil?
+        raise TypeError, 'Data Schema replacement fields cannot be empty.' if fields_list.empty?
+
+        @fields = fields_list.map { |field| DataSchemaField.new(field) }
+      end
+
+      # @return [Hash]
+      def to_hash
+        { fields: @fields.map(&:to_hash) }
+      end
+
+      # @return [String]
+      def to_s
+        to_hash.to_json
+      end
+    end
+
+    # Modify the Data Schema.
+    class DataSchema
+      # @return [Mindee::Input::DataSchemaReplace]
+      attr_reader :replace
+
+      # @param data_schema [Hash, String]
+      def initialize(data_schema)
+        case data_schema
+        when String
+          parsed = JSON.parse(data_schema.to_s, object_class: Hash)
+          parsed.transform_keys!(&:to_sym)
+          @replace = DataSchemaReplace.new(parsed[:replace])
+        when Hash
+          data_schema.transform_keys!(&:to_sym)
+          @replace = if data_schema[:replace].is_a?(DataSchemaReplace)
+                       data_schema[:replace]
+                     else
+                       DataSchemaReplace.new(data_schema[:replace])
+                     end
+        when DataSchema
+          @replace = data_schema.replace
+        else
+          raise TypeError, 'Invalid Data Schema provided.'
+        end
+      end
+
+      # @return [Hash]
+      def to_hash
+        { replace: @replace.to_hash }
+      end
+
+      # @return [String]
+      def to_s
+        to_hash.to_json
+      end
+    end
+  end
+end

data/lib/mindee/input/inference_parameters.rb

@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require_relative 'data_schema'
+
 module Mindee
   module Input
     # Parameters to set when sending a file for inference.
@@ -25,12 +27,19 @@ module Mindee
       # @return [String, nil] Optional alias for the file.
       attr_reader :file_alias
 
+      # @return [String, nil] Additional text context used by the model during inference.
+      #   Not recommended, for specific use only.
+      attr_reader :text_context
+
       # @return [Array<String>, nil] Optional list of Webhooks IDs to propagate the API response to.
       attr_reader :webhook_ids
 
       # @return [PollingOptions] Options for polling. Set only if having timeout issues.
       attr_reader :polling_options
 
+      # @return [DataSchemaField]
+      attr_reader :data_schema
+
       # @return [Boolean, nil] Whether to close the file after parsing.
       attr_reader :close_file
 
@@ -52,8 +61,10 @@ module Mindee
         confidence: nil,
         file_alias: nil,
         webhook_ids: nil,
+        text_context: nil,
         polling_options: nil,
-        close_file: true
+        close_file: true,
+        data_schema: nil
       )
         raise Errors::MindeeInputError, 'Model ID is required.' if model_id.empty? || model_id.nil?
 
@@ -64,8 +75,10 @@ module Mindee
         @confidence = confidence
         @file_alias = file_alias
         @webhook_ids = webhook_ids || []
+        @text_context = text_context
         @polling_options = get_clean_polling_options(polling_options)
         @close_file = close_file.nil? || close_file
+        @data_schema = DataSchema.new(data_schema) unless data_schema.nil?
         # rubocop:enable Metrics/ParameterLists
       end
 

data/lib/mindee/input.rb

@@ -1,5 +1,6 @@
 # frozen_string_literal: true
 
+require_relative 'input/data_schema'
 require_relative 'input/inference_parameters'
 require_relative 'input/polling_options'
 require_relative 'input/sources'

data/lib/mindee/parsing/v2/inference_active_options.rb

@@ -3,6 +3,23 @@
 module Mindee
   module Parsing
     module V2
+      # Data schema options activated during the inference.
+      class DataSchemaActiveOption
+        # @return [Boolean]
+        attr_reader :replace
+
+        # @param server_response [Hash]
+        def initialize(server_response)
+          @replace = server_response[:replace] || server_response['replace']
+        end
+
+        # String representation.
+        # @return [String]
+        def to_s
+          "Data Schema\n-----------\n:Replace: #{@replace ? 'True' : 'False'}"
+        end
+      end
+
       # Options which were activated during the inference.
       class InferenceActiveOptions
         # @return [Boolean] Whether the Raw Text feature was activated.
@@ -13,6 +30,10 @@ module Mindee
         attr_reader :confidence
         # @return [Boolean] Whether the Retrieval-Augmented Generation feature was activated.
         attr_reader :rag
+        # @return [Boolean] Whether the text context feature was activated.
+        attr_reader :text_context
+        # @return [DataSchemaActiveOption]
+        attr_reader :data_schema
 
         # @param server_response [Hash] Raw JSON parsed into a Hash.
         def initialize(server_response)
@@ -20,6 +41,8 @@ module Mindee
           @polygon = server_response['polygon']
           @confidence = server_response['confidence']
           @rag = server_response['rag']
+          @text_context = server_response['text_context']
+          @data_schema = DataSchemaActiveOption.new(server_response['data_schema'])
         end
 
         # String representation.
@@ -32,6 +55,8 @@ module Mindee
             ":Polygon: #{@polygon ? 'True' : 'False'}",
             ":Confidence: #{@confidence ? 'True' : 'False'}",
             ":RAG: #{@rag ? 'True' : 'False'}",
+            ":Text Context: #{@text_context ? 'True' : 'False'}\n",
+            @data_schema.to_s,
             '',
           ]
           parts.join("\n")

data/lib/mindee/version.rb

@@ -3,7 +3,7 @@
 # Mindee
 module Mindee
   # Current version.
-  VERSION = '4.8.0'
+  VERSION = '4.10.0'
 
   # Finds and return the current platform.
   # @return [Symbol, Hash[String | Symbol, Regexp], Nil?]

data/sig/mindee/http/mindee_api_v2.rbs

@@ -13,6 +13,10 @@ module Mindee
       def inference_job_req_get: (String) -> Net::HTTPResponse
       def inference_result_req_get: (String) -> Net::HTTPResponse
       def enqueue: (Input::Source::LocalInputSource | Input::Source::URLInputSource, Input::InferenceParameters) -> Net::HTTPResponse?
+
+      private
+
+      def enqueue_form_options: (Array[untyped], Input::InferenceParameters) -> Array[untyped]
     end
   end
 end

data/sig/mindee/input/data_schema.rbs

@@ -0,0 +1,34 @@
+module Mindee
+  module Input
+    class DataSchemaField
+      attr_reader title: String
+      attr_reader name: String
+      attr_reader is_array: bool
+      attr_reader type: String
+      attr_reader classification_values: String|nil
+      attr_reader unique_values: bool|nil
+      attr_reader description: String|nil
+      attr_reader guidelines: String|nil
+      attr_reader nested_fields: Array[Hash[String|Symbol, untyped]]|nil
+
+      def initialize: (Hash[Symbol, untyped]) -> void
+      def to_hash: () -> Hash[Symbol, untyped]
+      def to_string: () -> String
+    end
+
+    class DataSchemaReplace
+      attr_reader fields: Array[DataSchemaField]
+      def initialize: (Hash[Symbol, untyped]) -> void
+      def to_hash: () -> Hash[Symbol, untyped]
+      def to_string: () -> String
+    end
+
+    class DataSchema
+      attr_reader replace: DataSchemaReplace
+
+      def initialize: (Hash[String|Symbol, untyped]|String|DataSchema) -> void
+      def to_hash: () -> Hash[Symbol, untyped]
+      def to_s: -> String
+    end
+  end
+end

data/sig/mindee/input/inference_parameters.rbs

@@ -10,7 +10,9 @@ module Mindee
       attr_reader polygon: bool?
       attr_reader rag: bool?
       attr_reader raw_text: bool?
+      attr_reader text_context: String?
       attr_reader webhook_ids: Array[String]?
+      attr_reader data_schema: DataSchema?
 
       def initialize: (
           String,
@@ -19,9 +21,11 @@ module Mindee
           ?polygon: bool?,
           ?confidence: bool?,
           ?file_alias: String?,
+          ?text_context: String?,
           ?webhook_ids: Array[String]?,
           ?polling_options: Hash[Symbol | String, untyped] | PollingOptions?,
-          ?close_file: bool?
+          ?close_file: bool?,
+          ?data_schema: DataSchema|String|Hash[Symbol | String, untyped]?
         ) -> void
 
       def self.from_hash: (params: Hash[String | Symbol, untyped]) -> InferenceParameters

data/sig/mindee/input/local_response.rbs

@@ -3,7 +3,7 @@ module Mindee
   module Input
     class LocalResponse
       def file: -> StringIO
-      def initialize: (File | IO | StringIO | String | Pathname) -> void
+      def initialize: (File | IO | StringIO | String | Pathname | Tempfile) -> void
       def as_hash: -> Hash[String | Symbol, untyped]
       def self.process_secret_key: (String) -> String
       def get_hmac_signature: (String) -> String

data/sig/mindee/parsing/v2/inference_active_options.rbs

@@ -1,13 +1,22 @@
 module Mindee
   module Parsing
     module V2
+      class DataSchemaActiveOption
+        attr_reader replace: bool
+
+        def initialize: (Hash[Symbol |string, untyped]) -> void
+        def to_s: () -> String
+      end
       class InferenceActiveOptions
         attr_reader confidence: bool
         attr_reader polygon: bool
         attr_reader rag: bool
         attr_reader raw_text: bool
+        attr_reader text_context: bool
+        attr_reader data_schema: DataSchemaActiveOption
 
         def initialize: (Hash[String | Symbol, untyped]) -> void
+        def to_s: () -> String
       end
     end
   end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: mindee
 version: !ruby/object:Gem::Version
-  version: 4.8.0
+  version: 4.10.0
 platform: ruby
 authors:
 - Mindee, SA
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-11-18 00:00:00.000000000 Z
+date: 2025-12-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: base64
@@ -293,6 +293,7 @@ files:
 - lib/mindee/image/image_extractor.rb
 - lib/mindee/image/image_utils.rb
 - lib/mindee/input.rb
+- lib/mindee/input/data_schema.rb
 - lib/mindee/input/inference_parameters.rb
 - lib/mindee/input/local_response.rb
 - lib/mindee/input/polling_options.rb
@@ -574,6 +575,7 @@ files:
 - sig/mindee/image/image_compressor.rbs
 - sig/mindee/image/image_extractor.rbs
 - sig/mindee/image/image_utils.rbs
+- sig/mindee/input/data_schema.rbs
 - sig/mindee/input/inference_parameters.rbs
 - sig/mindee/input/local_response.rbs
 - sig/mindee/input/polling_options.rbs