carbon_ruby_sdk 0.1.13 → 0.1.15

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: f933446ddcb13014a55fd9b6f8744a083f495d54597cf898af90ca17caf3ac0b
-  data.tar.gz: b8d7e752e42e617d4faa0595a8cd41c6b0b3088c678542bc7a8f2be9abe506a3
+  metadata.gz: 3fd28e7eb718e7a438c849d5d9820cb60fc73d8cdacfe0ef13f30fb6e97c67b3
+  data.tar.gz: a0b09df8026bb03a9e9549a01b09929a7c1b98be45579b399e54811ac21363f1
 SHA512:
-  metadata.gz: 9d76f26f825fdc1560d7a2b4e103b4b29da0b81bad1131d17583779cb6f5254e717732cb29838b5d80ce2dfebd62dd4ca4b59bb41a1d005f8f88f7f9c237dc47
-  data.tar.gz: 1130a78f7270d44827e133cc2a8b6deffb10040b27091e02ec5e54bae72969d0ab63b6716d704244983ad4979351f1c2e5ca0f84005f1e4eae8e7c57c2f87348
+  metadata.gz: 8a75f3914ec5e430b9c1ea7fd75367b34ff96bc516c434146c36c049c275747f852d1c7a60d93dd32acb4321806ce80e5436f873b3fb21df43f8d7292da92ed0
+  data.tar.gz: 723037b41f5609aef0d2af2e410bf6cce42065ec5c5c8125e59bbc89d26a0bbd3b26af6c090ee4bfa41094c96aa563ad704fdea588e29c3110a3acdebe99c8d4

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    carbon_ruby_sdk (0.1.12)
+    carbon_ruby_sdk (0.1.14)
       faraday (>= 1.0.1, < 3.0)
       faraday-multipart (~> 1.0, >= 1.0.4)
 

data/README.md

@@ -6,7 +6,7 @@
 
 Connect external data to LLMs, no matter the source.
 
-[![npm](https://img.shields.io/badge/gem-v0.1.13-blue)](https://rubygems.org/gems/carbon_ruby_sdk/versions/0.1.13)
+[![npm](https://img.shields.io/badge/gem-v0.1.15-blue)](https://rubygems.org/gems/carbon_ruby_sdk/versions/0.1.15)
 
 </div>
 
@@ -80,7 +80,7 @@ Connect external data to LLMs, no matter the source.
 Add to Gemfile:
 
 ```ruby
-gem 'carbon_ruby_sdk', '~> 0.1.13'
+gem 'carbon_ruby_sdk', '~> 0.1.15'
 ```
 
 ## Getting Started<a id="getting-started"></a>
@@ -1381,6 +1381,8 @@ result = carbon.integrations.get_oauth_url(
   data_source_id: 1,
   connecting_new_account: false,
   request_id: "string_example",
+  use_ocr: false,
+  parse_pdf_tables_with_ocr: false,
 )
 p result
 ```
@@ -1424,6 +1426,10 @@ sync URL for an existing data source based on type and ID.
 This request id will be added to all files that get synced using the generated
 OAuth URL
 
+##### use_ocr: `Boolean`<a id="use_ocr-boolean"></a>
+Enable OCR for files that support it. Supported formats: pdf
+
+##### parse_pdf_tables_with_ocr: `Boolean`<a id="parse_pdf_tables_with_ocr-boolean"></a>
 #### 🔄 Return<a id="🔄-return"></a>
 
 [OuthURLResponse](./lib/carbon_ruby_sdk/models/outh_url_response.rb)
@@ -1641,6 +1647,8 @@ result = carbon.integrations.sync_confluence(
   max_items_per_chunk: 1,
   set_page_as_boundary: false,
   request_id: "string_example",
+  use_ocr: false,
+  parse_pdf_tables_with_ocr: false,
 )
 p result
 ```
@@ -1661,6 +1669,8 @@ Number of objects per chunk. For csv, tsv, xlsx, and json files only.
 
 ##### set_page_as_boundary: `Boolean`<a id="set_page_as_boundary-boolean"></a>
 ##### request_id: `String`<a id="request_id-string"></a>
+##### use_ocr: `Boolean`<a id="use_ocr-boolean"></a>
+##### parse_pdf_tables_with_ocr: `Boolean`<a id="parse_pdf_tables_with_ocr-boolean"></a>
 #### 🔄 Return<a id="🔄-return"></a>
 
 [GenericSuccessResponse](./lib/carbon_ruby_sdk/models/generic_success_response.rb)
@@ -1728,6 +1738,8 @@ result = carbon.integrations.sync_files(
   max_items_per_chunk: 1,
   set_page_as_boundary: false,
   request_id: "string_example",
+  use_ocr: false,
+  parse_pdf_tables_with_ocr: false,
 )
 p result
 ```
@@ -1748,6 +1760,8 @@ Number of objects per chunk. For csv, tsv, xlsx, and json files only.
 
 ##### set_page_as_boundary: `Boolean`<a id="set_page_as_boundary-boolean"></a>
 ##### request_id: `String`<a id="request_id-string"></a>
+##### use_ocr: `Boolean`<a id="use_ocr-boolean"></a>
+##### parse_pdf_tables_with_ocr: `Boolean`<a id="parse_pdf_tables_with_ocr-boolean"></a>
 #### 🔄 Return<a id="🔄-return"></a>
 
 [GenericSuccessResponse](./lib/carbon_ruby_sdk/models/generic_success_response.rb)
@@ -2081,6 +2095,8 @@ result = carbon.integrations.sync_s3_files(
   set_page_as_boundary: false,
   data_source_id: 1,
   request_id: "string_example",
+  use_ocr: false,
+  parse_pdf_tables_with_ocr: false,
 )
 p result
 ```
@@ -2101,6 +2117,8 @@ Number of objects per chunk. For csv, tsv, xlsx, and json files only.
 ##### set_page_as_boundary: `Boolean`<a id="set_page_as_boundary-boolean"></a>
 ##### data_source_id: `Integer`<a id="data_source_id-integer"></a>
 ##### request_id: `String`<a id="request_id-string"></a>
+##### use_ocr: `Boolean`<a id="use_ocr-boolean"></a>
+##### parse_pdf_tables_with_ocr: `Boolean`<a id="parse_pdf_tables_with_ocr-boolean"></a>
 #### 🔄 Return<a id="🔄-return"></a>
 
 [GenericSuccessResponse](./lib/carbon_ruby_sdk/models/generic_success_response.rb)

data/lib/carbon_ruby_sdk/api/integrations_api.rb

@@ -543,9 +543,11 @@ module Carbon
     # @param data_source_id [Integer] Used to specify a data source to sync from if you have multiple connected. It can be skipped if you only have one data source of that type connected or are connecting a new account.
     # @param connecting_new_account [Boolean] Used to connect a new data source. If not specified, we will attempt to create a sync URL for an existing data source based on type and ID.
     # @param request_id [String] This request id will be added to all files that get synced using the generated OAuth URL
+    # @param use_ocr [Boolean] Enable OCR for files that support it. Supported formats: pdf
+    # @param parse_pdf_tables_with_ocr [Boolean] 
     # @param body [OAuthURLRequest] 
     # @param [Hash] extra additional parameters to pass along through :header_params, :query_params, or parameter name
-    def get_oauth_url(service:, tags: SENTINEL, scope: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', zendesk_subdomain: SENTINEL, microsoft_tenant: SENTINEL, sharepoint_site_name: SENTINEL, confluence_subdomain: SENTINEL, generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, salesforce_domain: SENTINEL, sync_files_on_connection: true, set_page_as_boundary: false, data_source_id: SENTINEL, connecting_new_account: false, request_id: SENTINEL, extra: {})
+    def get_oauth_url(service:, tags: SENTINEL, scope: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', zendesk_subdomain: SENTINEL, microsoft_tenant: SENTINEL, sharepoint_site_name: SENTINEL, confluence_subdomain: SENTINEL, generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, salesforce_domain: SENTINEL, sync_files_on_connection: true, set_page_as_boundary: false, data_source_id: SENTINEL, connecting_new_account: false, request_id: SENTINEL, use_ocr: false, parse_pdf_tables_with_ocr: false, extra: {})
       _body = {}
       _body[:tags] = tags if tags != SENTINEL
       _body[:scope] = scope if scope != SENTINEL
@@ -567,6 +569,8 @@ module Carbon
       _body[:data_source_id] = data_source_id if data_source_id != SENTINEL
       _body[:connecting_new_account] = connecting_new_account if connecting_new_account != SENTINEL
       _body[:request_id] = request_id if request_id != SENTINEL
+      _body[:use_ocr] = use_ocr if use_ocr != SENTINEL
+      _body[:parse_pdf_tables_with_ocr] = parse_pdf_tables_with_ocr if parse_pdf_tables_with_ocr != SENTINEL
       o_auth_url_request = _body
       api_response = get_oauth_url_with_http_info_impl(o_auth_url_request, extra)
       api_response.data
@@ -599,9 +603,11 @@ module Carbon
     # @param data_source_id [Integer] Used to specify a data source to sync from if you have multiple connected. It can be skipped if you only have one data source of that type connected or are connecting a new account.
     # @param connecting_new_account [Boolean] Used to connect a new data source. If not specified, we will attempt to create a sync URL for an existing data source based on type and ID.
     # @param request_id [String] This request id will be added to all files that get synced using the generated OAuth URL
+    # @param use_ocr [Boolean] Enable OCR for files that support it. Supported formats: pdf
+    # @param parse_pdf_tables_with_ocr [Boolean] 
     # @param body [OAuthURLRequest] 
     # @param [Hash] extra additional parameters to pass along through :header_params, :query_params, or parameter name
-    def get_oauth_url_with_http_info(service:, tags: SENTINEL, scope: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', zendesk_subdomain: SENTINEL, microsoft_tenant: SENTINEL, sharepoint_site_name: SENTINEL, confluence_subdomain: SENTINEL, generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, salesforce_domain: SENTINEL, sync_files_on_connection: true, set_page_as_boundary: false, data_source_id: SENTINEL, connecting_new_account: false, request_id: SENTINEL, extra: {})
+    def get_oauth_url_with_http_info(service:, tags: SENTINEL, scope: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', zendesk_subdomain: SENTINEL, microsoft_tenant: SENTINEL, sharepoint_site_name: SENTINEL, confluence_subdomain: SENTINEL, generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, salesforce_domain: SENTINEL, sync_files_on_connection: true, set_page_as_boundary: false, data_source_id: SENTINEL, connecting_new_account: false, request_id: SENTINEL, use_ocr: false, parse_pdf_tables_with_ocr: false, extra: {})
       _body = {}
       _body[:tags] = tags if tags != SENTINEL
       _body[:scope] = scope if scope != SENTINEL
@@ -623,6 +629,8 @@ module Carbon
       _body[:data_source_id] = data_source_id if data_source_id != SENTINEL
       _body[:connecting_new_account] = connecting_new_account if connecting_new_account != SENTINEL
       _body[:request_id] = request_id if request_id != SENTINEL
+      _body[:use_ocr] = use_ocr if use_ocr != SENTINEL
+      _body[:parse_pdf_tables_with_ocr] = parse_pdf_tables_with_ocr if parse_pdf_tables_with_ocr != SENTINEL
       o_auth_url_request = _body
       get_oauth_url_with_http_info_impl(o_auth_url_request, extra)
     end
@@ -1277,9 +1285,11 @@ module Carbon
     # @param max_items_per_chunk [Integer] Number of objects per chunk. For csv, tsv, xlsx, and json files only.
     # @param set_page_as_boundary [Boolean] 
     # @param request_id [String] 
+    # @param use_ocr [Boolean] 
+    # @param parse_pdf_tables_with_ocr [Boolean] 
     # @param body [SyncFilesRequest] 
     # @param [Hash] extra additional parameters to pass along through :header_params, :query_params, or parameter name
-    def sync_confluence(data_source_id:, ids:, tags: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, set_page_as_boundary: false, request_id: SENTINEL, extra: {})
+    def sync_confluence(data_source_id:, ids:, tags: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, set_page_as_boundary: false, request_id: SENTINEL, use_ocr: false, parse_pdf_tables_with_ocr: false, extra: {})
       _body = {}
       _body[:tags] = tags if tags != SENTINEL
       _body[:data_source_id] = data_source_id if data_source_id != SENTINEL
@@ -1293,6 +1303,8 @@ module Carbon
       _body[:max_items_per_chunk] = max_items_per_chunk if max_items_per_chunk != SENTINEL
       _body[:set_page_as_boundary] = set_page_as_boundary if set_page_as_boundary != SENTINEL
       _body[:request_id] = request_id if request_id != SENTINEL
+      _body[:use_ocr] = use_ocr if use_ocr != SENTINEL
+      _body[:parse_pdf_tables_with_ocr] = parse_pdf_tables_with_ocr if parse_pdf_tables_with_ocr != SENTINEL
       sync_files_request = _body
       api_response = sync_confluence_with_http_info_impl(sync_files_request, extra)
       api_response.data
@@ -1317,9 +1329,11 @@ module Carbon
     # @param max_items_per_chunk [Integer] Number of objects per chunk. For csv, tsv, xlsx, and json files only.
     # @param set_page_as_boundary [Boolean] 
     # @param request_id [String] 
+    # @param use_ocr [Boolean] 
+    # @param parse_pdf_tables_with_ocr [Boolean] 
     # @param body [SyncFilesRequest] 
     # @param [Hash] extra additional parameters to pass along through :header_params, :query_params, or parameter name
-    def sync_confluence_with_http_info(data_source_id:, ids:, tags: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, set_page_as_boundary: false, request_id: SENTINEL, extra: {})
+    def sync_confluence_with_http_info(data_source_id:, ids:, tags: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, set_page_as_boundary: false, request_id: SENTINEL, use_ocr: false, parse_pdf_tables_with_ocr: false, extra: {})
       _body = {}
       _body[:tags] = tags if tags != SENTINEL
       _body[:data_source_id] = data_source_id if data_source_id != SENTINEL
@@ -1333,6 +1347,8 @@ module Carbon
       _body[:max_items_per_chunk] = max_items_per_chunk if max_items_per_chunk != SENTINEL
       _body[:set_page_as_boundary] = set_page_as_boundary if set_page_as_boundary != SENTINEL
       _body[:request_id] = request_id if request_id != SENTINEL
+      _body[:use_ocr] = use_ocr if use_ocr != SENTINEL
+      _body[:parse_pdf_tables_with_ocr] = parse_pdf_tables_with_ocr if parse_pdf_tables_with_ocr != SENTINEL
       sync_files_request = _body
       sync_confluence_with_http_info_impl(sync_files_request, extra)
     end
@@ -1517,9 +1533,11 @@ module Carbon
     # @param max_items_per_chunk [Integer] Number of objects per chunk. For csv, tsv, xlsx, and json files only.
     # @param set_page_as_boundary [Boolean] 
     # @param request_id [String] 
+    # @param use_ocr [Boolean] 
+    # @param parse_pdf_tables_with_ocr [Boolean] 
     # @param body [SyncFilesRequest] 
     # @param [Hash] extra additional parameters to pass along through :header_params, :query_params, or parameter name
-    def sync_files(data_source_id:, ids:, tags: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, set_page_as_boundary: false, request_id: SENTINEL, extra: {})
+    def sync_files(data_source_id:, ids:, tags: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, set_page_as_boundary: false, request_id: SENTINEL, use_ocr: false, parse_pdf_tables_with_ocr: false, extra: {})
       _body = {}
       _body[:tags] = tags if tags != SENTINEL
       _body[:data_source_id] = data_source_id if data_source_id != SENTINEL
@@ -1533,6 +1551,8 @@ module Carbon
       _body[:max_items_per_chunk] = max_items_per_chunk if max_items_per_chunk != SENTINEL
       _body[:set_page_as_boundary] = set_page_as_boundary if set_page_as_boundary != SENTINEL
       _body[:request_id] = request_id if request_id != SENTINEL
+      _body[:use_ocr] = use_ocr if use_ocr != SENTINEL
+      _body[:parse_pdf_tables_with_ocr] = parse_pdf_tables_with_ocr if parse_pdf_tables_with_ocr != SENTINEL
       sync_files_request = _body
       api_response = sync_files_with_http_info_impl(sync_files_request, extra)
       api_response.data
@@ -1557,9 +1577,11 @@ module Carbon
     # @param max_items_per_chunk [Integer] Number of objects per chunk. For csv, tsv, xlsx, and json files only.
     # @param set_page_as_boundary [Boolean] 
     # @param request_id [String] 
+    # @param use_ocr [Boolean] 
+    # @param parse_pdf_tables_with_ocr [Boolean] 
     # @param body [SyncFilesRequest] 
     # @param [Hash] extra additional parameters to pass along through :header_params, :query_params, or parameter name
-    def sync_files_with_http_info(data_source_id:, ids:, tags: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, set_page_as_boundary: false, request_id: SENTINEL, extra: {})
+    def sync_files_with_http_info(data_source_id:, ids:, tags: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, set_page_as_boundary: false, request_id: SENTINEL, use_ocr: false, parse_pdf_tables_with_ocr: false, extra: {})
       _body = {}
       _body[:tags] = tags if tags != SENTINEL
       _body[:data_source_id] = data_source_id if data_source_id != SENTINEL
@@ -1573,6 +1595,8 @@ module Carbon
       _body[:max_items_per_chunk] = max_items_per_chunk if max_items_per_chunk != SENTINEL
       _body[:set_page_as_boundary] = set_page_as_boundary if set_page_as_boundary != SENTINEL
       _body[:request_id] = request_id if request_id != SENTINEL
+      _body[:use_ocr] = use_ocr if use_ocr != SENTINEL
+      _body[:parse_pdf_tables_with_ocr] = parse_pdf_tables_with_ocr if parse_pdf_tables_with_ocr != SENTINEL
       sync_files_request = _body
       sync_files_with_http_info_impl(sync_files_request, extra)
     end
@@ -2412,9 +2436,11 @@ module Carbon
     # @param set_page_as_boundary [Boolean] 
     # @param data_source_id [Integer] 
     # @param request_id [String] 
+    # @param use_ocr [Boolean] 
+    # @param parse_pdf_tables_with_ocr [Boolean] 
     # @param body [S3FileSyncInput] 
     # @param [Hash] extra additional parameters to pass along through :header_params, :query_params, or parameter name
-    def sync_s3_files(ids:, tags: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, set_page_as_boundary: false, data_source_id: SENTINEL, request_id: SENTINEL, extra: {})
+    def sync_s3_files(ids:, tags: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, set_page_as_boundary: false, data_source_id: SENTINEL, request_id: SENTINEL, use_ocr: false, parse_pdf_tables_with_ocr: false, extra: {})
       _body = {}
       _body[:tags] = tags if tags != SENTINEL
       _body[:ids] = ids if ids != SENTINEL
@@ -2428,6 +2454,8 @@ module Carbon
       _body[:set_page_as_boundary] = set_page_as_boundary if set_page_as_boundary != SENTINEL
       _body[:data_source_id] = data_source_id if data_source_id != SENTINEL
       _body[:request_id] = request_id if request_id != SENTINEL
+      _body[:use_ocr] = use_ocr if use_ocr != SENTINEL
+      _body[:parse_pdf_tables_with_ocr] = parse_pdf_tables_with_ocr if parse_pdf_tables_with_ocr != SENTINEL
       s3_file_sync_input = _body
       api_response = sync_s3_files_with_http_info_impl(s3_file_sync_input, extra)
       api_response.data
@@ -2451,9 +2479,11 @@ module Carbon
     # @param set_page_as_boundary [Boolean] 
     # @param data_source_id [Integer] 
     # @param request_id [String] 
+    # @param use_ocr [Boolean] 
+    # @param parse_pdf_tables_with_ocr [Boolean] 
     # @param body [S3FileSyncInput] 
     # @param [Hash] extra additional parameters to pass along through :header_params, :query_params, or parameter name
-    def sync_s3_files_with_http_info(ids:, tags: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, set_page_as_boundary: false, data_source_id: SENTINEL, request_id: SENTINEL, extra: {})
+    def sync_s3_files_with_http_info(ids:, tags: SENTINEL, chunk_size: 1500, chunk_overlap: 20, skip_embedding_generation: false, embedding_model: 'OPENAI', generate_sparse_vectors: false, prepend_filename_to_chunks: false, max_items_per_chunk: SENTINEL, set_page_as_boundary: false, data_source_id: SENTINEL, request_id: SENTINEL, use_ocr: false, parse_pdf_tables_with_ocr: false, extra: {})
       _body = {}
       _body[:tags] = tags if tags != SENTINEL
       _body[:ids] = ids if ids != SENTINEL
@@ -2467,6 +2497,8 @@ module Carbon
       _body[:set_page_as_boundary] = set_page_as_boundary if set_page_as_boundary != SENTINEL
       _body[:data_source_id] = data_source_id if data_source_id != SENTINEL
       _body[:request_id] = request_id if request_id != SENTINEL
+      _body[:use_ocr] = use_ocr if use_ocr != SENTINEL
+      _body[:parse_pdf_tables_with_ocr] = parse_pdf_tables_with_ocr if parse_pdf_tables_with_ocr != SENTINEL
       s3_file_sync_input = _body
       sync_s3_files_with_http_info_impl(s3_file_sync_input, extra)
     end

data/lib/carbon_ruby_sdk/models/document_response.rb

@@ -33,6 +33,8 @@ module Carbon
 
     attr_accessor :content_metadata
 
+    attr_accessor :chunk_index
+
     # Attribute mapping from ruby-style variable name to JSON key.
     def self.attribute_map
       {
@@ -46,7 +48,8 @@ module Carbon
         :'vector' => :'vector',
         :'score' => :'score',
         :'rank' => :'rank',
-        :'content_metadata' => :'content_metadata'
+        :'content_metadata' => :'content_metadata',
+        :'chunk_index' => :'chunk_index'
       }
     end
 
@@ -68,7 +71,8 @@ module Carbon
         :'vector' => :'Array<Float>',
         :'score' => :'Float',
         :'rank' => :'RankProperty',
-        :'content_metadata' => :'Object'
+        :'content_metadata' => :'Object',
+        :'chunk_index' => :'Integer'
       }
     end
 
@@ -83,7 +87,8 @@ module Carbon
         :'vector',
         :'score',
         :'rank',
-        :'content_metadata'
+        :'content_metadata',
+        :'chunk_index'
       ])
     end
 
@@ -149,6 +154,10 @@ module Carbon
       if attributes.key?(:'content_metadata')
         self.content_metadata = attributes[:'content_metadata']
       end
+
+      if attributes.key?(:'chunk_index')
+        self.chunk_index = attributes[:'chunk_index']
+      end
     end
 
     # Show invalid properties with the reasons. Usually used together with valid?
@@ -189,7 +198,8 @@ module Carbon
           vector == o.vector &&
           score == o.score &&
           rank == o.rank &&
-          content_metadata == o.content_metadata
+          content_metadata == o.content_metadata &&
+          chunk_index == o.chunk_index
     end
 
     # @see the `==` method
@@ -201,7 +211,7 @@ module Carbon
     # Calculates hash code according to all attributes.
     # @return [Integer] Hash code
     def hash
-      [tags, content, file_id, source, source_url, source_type, presigned_url, vector, score, rank, content_metadata].hash
+      [tags, content, file_id, source, source_url, source_type, presigned_url, vector, score, rank, content_metadata, chunk_index].hash
     end
 
     # Builds the object from hash

data/lib/carbon_ruby_sdk/models/o_auth_url_request.rb

@@ -56,6 +56,11 @@ module Carbon
     # This request id will be added to all files that get synced using the generated OAuth URL
     attr_accessor :request_id
 
+    # Enable OCR for files that support it. Supported formats: pdf
+    attr_accessor :use_ocr
+
+    attr_accessor :parse_pdf_tables_with_ocr
+
     # Attribute mapping from ruby-style variable name to JSON key.
     def self.attribute_map
       {
@@ -78,7 +83,9 @@ module Carbon
         :'set_page_as_boundary' => :'set_page_as_boundary',
         :'data_source_id' => :'data_source_id',
         :'connecting_new_account' => :'connecting_new_account',
-        :'request_id' => :'request_id'
+        :'request_id' => :'request_id',
+        :'use_ocr' => :'use_ocr',
+        :'parse_pdf_tables_with_ocr' => :'parse_pdf_tables_with_ocr'
       }
     end
 
@@ -109,7 +116,9 @@ module Carbon
         :'set_page_as_boundary' => :'Boolean',
         :'data_source_id' => :'Integer',
         :'connecting_new_account' => :'Boolean',
-        :'request_id' => :'String'
+        :'request_id' => :'String',
+        :'use_ocr' => :'Boolean',
+        :'parse_pdf_tables_with_ocr' => :'Boolean'
       }
     end
 
@@ -133,7 +142,9 @@ module Carbon
         :'sync_files_on_connection',
         :'data_source_id',
         :'connecting_new_account',
-        :'request_id'
+        :'request_id',
+        :'use_ocr',
+        :'parse_pdf_tables_with_ocr'
       ])
     end
 
@@ -249,6 +260,18 @@ module Carbon
       if attributes.key?(:'request_id')
         self.request_id = attributes[:'request_id']
       end
+
+      if attributes.key?(:'use_ocr')
+        self.use_ocr = attributes[:'use_ocr']
+      else
+        self.use_ocr = false
+      end
+
+      if attributes.key?(:'parse_pdf_tables_with_ocr')
+        self.parse_pdf_tables_with_ocr = attributes[:'parse_pdf_tables_with_ocr']
+      else
+        self.parse_pdf_tables_with_ocr = false
+      end
     end
 
     # Show invalid properties with the reasons. Usually used together with valid?
@@ -293,7 +316,9 @@ module Carbon
           set_page_as_boundary == o.set_page_as_boundary &&
           data_source_id == o.data_source_id &&
           connecting_new_account == o.connecting_new_account &&
-          request_id == o.request_id
+          request_id == o.request_id &&
+          use_ocr == o.use_ocr &&
+          parse_pdf_tables_with_ocr == o.parse_pdf_tables_with_ocr
     end
 
     # @see the `==` method
@@ -305,7 +330,7 @@ module Carbon
     # Calculates hash code according to all attributes.
     # @return [Integer] Hash code
     def hash
-      [tags, scope, service, chunk_size, chunk_overlap, skip_embedding_generation, embedding_model, zendesk_subdomain, microsoft_tenant, sharepoint_site_name, confluence_subdomain, generate_sparse_vectors, prepend_filename_to_chunks, max_items_per_chunk, salesforce_domain, sync_files_on_connection, set_page_as_boundary, data_source_id, connecting_new_account, request_id].hash
+      [tags, scope, service, chunk_size, chunk_overlap, skip_embedding_generation, embedding_model, zendesk_subdomain, microsoft_tenant, sharepoint_site_name, confluence_subdomain, generate_sparse_vectors, prepend_filename_to_chunks, max_items_per_chunk, salesforce_domain, sync_files_on_connection, set_page_as_boundary, data_source_id, connecting_new_account, request_id, use_ocr, parse_pdf_tables_with_ocr].hash
     end
 
     # Builds the object from hash

data/lib/carbon_ruby_sdk/models/organization_user_data_source_api.rb

@@ -35,6 +35,8 @@ module Carbon
 
     attr_accessor :last_sync_action
 
+    attr_accessor :enable_auto_sync
+
     attr_accessor :created_at
 
     attr_accessor :updated_at
@@ -54,6 +56,7 @@ module Carbon
         :'revoked_access' => :'revoked_access',
         :'last_synced_at' => :'last_synced_at',
         :'last_sync_action' => :'last_sync_action',
+        :'enable_auto_sync' => :'enable_auto_sync',
         :'created_at' => :'created_at',
         :'updated_at' => :'updated_at'
       }
@@ -79,6 +82,7 @@ module Carbon
         :'revoked_access' => :'Boolean',
         :'last_synced_at' => :'Time',
         :'last_sync_action' => :'DataSourceLastSyncActions',
+        :'enable_auto_sync' => :'Boolean',
         :'created_at' => :'Time',
         :'updated_at' => :'Time'
       }
@@ -90,6 +94,7 @@ module Carbon
         :'data_source_external_id',
         :'token',
         :'source_items_synced_at',
+        :'enable_auto_sync',
       ])
     end
 
@@ -156,6 +161,10 @@ module Carbon
         self.last_sync_action = attributes[:'last_sync_action']
       end
 
+      if attributes.key?(:'enable_auto_sync')
+        self.enable_auto_sync = attributes[:'enable_auto_sync']
+      end
+
       if attributes.key?(:'created_at')
         self.created_at = attributes[:'created_at']
       end
@@ -250,6 +259,7 @@ module Carbon
           revoked_access == o.revoked_access &&
           last_synced_at == o.last_synced_at &&
           last_sync_action == o.last_sync_action &&
+          enable_auto_sync == o.enable_auto_sync &&
           created_at == o.created_at &&
           updated_at == o.updated_at
     end
@@ -263,7 +273,7 @@ module Carbon
     # Calculates hash code according to all attributes.
     # @return [Integer] Hash code
     def hash
-      [id, data_source_external_id, data_source_type, token, sync_status, source_items_synced_at, organization_user_id, organization_id, organization_supplied_user_id, revoked_access, last_synced_at, last_sync_action, created_at, updated_at].hash
+      [id, data_source_external_id, data_source_type, token, sync_status, source_items_synced_at, organization_user_id, organization_id, organization_supplied_user_id, revoked_access, last_synced_at, last_sync_action, enable_auto_sync, created_at, updated_at].hash
     end
 
     # Builds the object from hash

data/lib/carbon_ruby_sdk/models/s3_file_sync_input.rb

@@ -36,6 +36,10 @@ module Carbon
 
     attr_accessor :request_id
 
+    attr_accessor :use_ocr
+
+    attr_accessor :parse_pdf_tables_with_ocr
+
     # Attribute mapping from ruby-style variable name to JSON key.
     def self.attribute_map
       {
@@ -50,7 +54,9 @@ module Carbon
         :'max_items_per_chunk' => :'max_items_per_chunk',
         :'set_page_as_boundary' => :'set_page_as_boundary',
         :'data_source_id' => :'data_source_id',
-        :'request_id' => :'request_id'
+        :'request_id' => :'request_id',
+        :'use_ocr' => :'use_ocr',
+        :'parse_pdf_tables_with_ocr' => :'parse_pdf_tables_with_ocr'
       }
     end
 
@@ -73,7 +79,9 @@ module Carbon
         :'max_items_per_chunk' => :'Integer',
         :'set_page_as_boundary' => :'Boolean',
         :'data_source_id' => :'Integer',
-        :'request_id' => :'String'
+        :'request_id' => :'String',
+        :'use_ocr' => :'Boolean',
+        :'parse_pdf_tables_with_ocr' => :'Boolean'
       }
     end
 
@@ -88,7 +96,9 @@ module Carbon
         :'prepend_filename_to_chunks',
         :'max_items_per_chunk',
         :'data_source_id',
-        :'request_id'
+        :'request_id',
+        :'use_ocr',
+        :'parse_pdf_tables_with_ocr'
       ])
     end
 
@@ -170,6 +180,18 @@ module Carbon
       if attributes.key?(:'request_id')
         self.request_id = attributes[:'request_id']
       end
+
+      if attributes.key?(:'use_ocr')
+        self.use_ocr = attributes[:'use_ocr']
+      else
+        self.use_ocr = false
+      end
+
+      if attributes.key?(:'parse_pdf_tables_with_ocr')
+        self.parse_pdf_tables_with_ocr = attributes[:'parse_pdf_tables_with_ocr']
+      else
+        self.parse_pdf_tables_with_ocr = false
+      end
     end
 
     # Show invalid properties with the reasons. Usually used together with valid?
@@ -206,7 +228,9 @@ module Carbon
           max_items_per_chunk == o.max_items_per_chunk &&
           set_page_as_boundary == o.set_page_as_boundary &&
           data_source_id == o.data_source_id &&
-          request_id == o.request_id
+          request_id == o.request_id &&
+          use_ocr == o.use_ocr &&
+          parse_pdf_tables_with_ocr == o.parse_pdf_tables_with_ocr
     end
 
     # @see the `==` method
@@ -218,7 +242,7 @@ module Carbon
     # Calculates hash code according to all attributes.
     # @return [Integer] Hash code
     def hash
-      [tags, ids, chunk_size, chunk_overlap, skip_embedding_generation, embedding_model, generate_sparse_vectors, prepend_filename_to_chunks, max_items_per_chunk, set_page_as_boundary, data_source_id, request_id].hash
+      [tags, ids, chunk_size, chunk_overlap, skip_embedding_generation, embedding_model, generate_sparse_vectors, prepend_filename_to_chunks, max_items_per_chunk, set_page_as_boundary, data_source_id, request_id, use_ocr, parse_pdf_tables_with_ocr].hash
     end
 
     # Builds the object from hash

data/lib/carbon_ruby_sdk/models/sync_files_request.rb

@@ -36,6 +36,10 @@ module Carbon
 
     attr_accessor :request_id
 
+    attr_accessor :use_ocr
+
+    attr_accessor :parse_pdf_tables_with_ocr
+
     # Attribute mapping from ruby-style variable name to JSON key.
     def self.attribute_map
       {
@@ -50,7 +54,9 @@ module Carbon
         :'prepend_filename_to_chunks' => :'prepend_filename_to_chunks',
         :'max_items_per_chunk' => :'max_items_per_chunk',
         :'set_page_as_boundary' => :'set_page_as_boundary',
-        :'request_id' => :'request_id'
+        :'request_id' => :'request_id',
+        :'use_ocr' => :'use_ocr',
+        :'parse_pdf_tables_with_ocr' => :'parse_pdf_tables_with_ocr'
       }
     end
 
@@ -73,7 +79,9 @@ module Carbon
         :'prepend_filename_to_chunks' => :'Boolean',
         :'max_items_per_chunk' => :'Integer',
         :'set_page_as_boundary' => :'Boolean',
-        :'request_id' => :'String'
+        :'request_id' => :'String',
+        :'use_ocr' => :'Boolean',
+        :'parse_pdf_tables_with_ocr' => :'Boolean'
       }
     end
 
@@ -88,7 +96,9 @@ module Carbon
         :'generate_sparse_vectors',
         :'prepend_filename_to_chunks',
         :'max_items_per_chunk',
-        :'request_id'
+        :'request_id',
+        :'use_ocr',
+        :'parse_pdf_tables_with_ocr'
       ])
     end
 
@@ -168,6 +178,18 @@ module Carbon
       if attributes.key?(:'request_id')
         self.request_id = attributes[:'request_id']
       end
+
+      if attributes.key?(:'use_ocr')
+        self.use_ocr = attributes[:'use_ocr']
+      else
+        self.use_ocr = false
+      end
+
+      if attributes.key?(:'parse_pdf_tables_with_ocr')
+        self.parse_pdf_tables_with_ocr = attributes[:'parse_pdf_tables_with_ocr']
+      else
+        self.parse_pdf_tables_with_ocr = false
+      end
     end
 
     # Show invalid properties with the reasons. Usually used together with valid?
@@ -209,7 +231,9 @@ module Carbon
           prepend_filename_to_chunks == o.prepend_filename_to_chunks &&
           max_items_per_chunk == o.max_items_per_chunk &&
           set_page_as_boundary == o.set_page_as_boundary &&
-          request_id == o.request_id
+          request_id == o.request_id &&
+          use_ocr == o.use_ocr &&
+          parse_pdf_tables_with_ocr == o.parse_pdf_tables_with_ocr
     end
 
     # @see the `==` method
@@ -221,7 +245,7 @@ module Carbon
     # Calculates hash code according to all attributes.
     # @return [Integer] Hash code
     def hash
-      [tags, data_source_id, ids, chunk_size, chunk_overlap, skip_embedding_generation, embedding_model, generate_sparse_vectors, prepend_filename_to_chunks, max_items_per_chunk, set_page_as_boundary, request_id].hash
+      [tags, data_source_id, ids, chunk_size, chunk_overlap, skip_embedding_generation, embedding_model, generate_sparse_vectors, prepend_filename_to_chunks, max_items_per_chunk, set_page_as_boundary, request_id, use_ocr, parse_pdf_tables_with_ocr].hash
     end
 
     # Builds the object from hash

data/lib/carbon_ruby_sdk/models/user_file.rb

@@ -45,6 +45,8 @@ module Carbon
 
     attr_accessor :chunk_properties
 
+    attr_accessor :ocr_properties
+
     attr_accessor :name
 
     attr_accessor :parent_id
@@ -89,6 +91,7 @@ module Carbon
         :'chunk_size' => :'chunk_size',
         :'chunk_overlap' => :'chunk_overlap',
         :'chunk_properties' => :'chunk_properties',
+        :'ocr_properties' => :'ocr_properties',
         :'name' => :'name',
         :'parent_id' => :'parent_id',
         :'enable_auto_sync' => :'enable_auto_sync',
@@ -129,6 +132,7 @@ module Carbon
         :'chunk_size' => :'Integer',
         :'chunk_overlap' => :'Integer',
         :'chunk_properties' => :'ChunkPropertiesNullable',
+        :'ocr_properties' => :'Object',
         :'name' => :'String',
         :'parent_id' => :'Integer',
         :'enable_auto_sync' => :'Boolean',
@@ -255,6 +259,10 @@ module Carbon
         self.chunk_properties = attributes[:'chunk_properties']
       end
 
+      if attributes.key?(:'ocr_properties')
+        self.ocr_properties = attributes[:'ocr_properties']
+      end
+
       if attributes.key?(:'name')
         self.name = attributes[:'name']
       end
@@ -332,6 +340,10 @@ module Carbon
         invalid_properties.push('invalid value for "sync_status", sync_status cannot be nil.')
       end
 
+      if @ocr_properties.nil?
+        invalid_properties.push('invalid value for "ocr_properties", ocr_properties cannot be nil.')
+      end
+
       if @skip_embedding_generation.nil?
         invalid_properties.push('invalid value for "skip_embedding_generation", skip_embedding_generation cannot be nil.')
       end
@@ -356,6 +368,7 @@ module Carbon
       return false if @organization_supplied_user_id.nil?
       return false if @external_file_id.nil?
       return false if @sync_status.nil?
+      return false if @ocr_properties.nil?
       return false if @skip_embedding_generation.nil?
       return false if @created_at.nil?
       return false if @updated_at.nil?
@@ -384,6 +397,7 @@ module Carbon
           chunk_size == o.chunk_size &&
           chunk_overlap == o.chunk_overlap &&
           chunk_properties == o.chunk_properties &&
+          ocr_properties == o.ocr_properties &&
           name == o.name &&
           parent_id == o.parent_id &&
           enable_auto_sync == o.enable_auto_sync &&
@@ -407,7 +421,7 @@ module Carbon
     # Calculates hash code according to all attributes.
     # @return [Integer] Hash code
     def hash
-      [tags, id, source, organization_id, organization_supplied_user_id, organization_user_data_source_id, external_file_id, external_url, sync_status, sync_error_message, last_sync, file_statistics, file_metadata, embedding_properties, chunk_size, chunk_overlap, chunk_properties, name, parent_id, enable_auto_sync, presigned_url, parsed_text_url, additional_presigned_urls, skip_embedding_generation, source_created_at, generate_sparse_vectors, request_id, created_at, updated_at].hash
+      [tags, id, source, organization_id, organization_supplied_user_id, organization_user_data_source_id, external_file_id, external_url, sync_status, sync_error_message, last_sync, file_statistics, file_metadata, embedding_properties, chunk_size, chunk_overlap, chunk_properties, ocr_properties, name, parent_id, enable_auto_sync, presigned_url, parsed_text_url, additional_presigned_urls, skip_embedding_generation, source_created_at, generate_sparse_vectors, request_id, created_at, updated_at].hash
     end
 
     # Builds the object from hash

data/lib/carbon_ruby_sdk/version.rb

@@ -7,5 +7,5 @@ The version of the OpenAPI document: 1.0.0
 =end
 
 module Carbon
-  VERSION = '0.1.13'
+  VERSION = '0.1.15'
 end

data/spec/models/document_response_spec.rb

@@ -85,4 +85,10 @@ describe Carbon::DocumentResponse do
     end
   end
 
+  describe 'test attribute "chunk_index"' do
+    it 'should work' do
+      # assertion here. ref: https://www.relishapp.com/rspec/rspec-expectations/docs/built-in-matchers
+    end
+  end
+
 end

data/spec/models/o_auth_url_request_spec.rb

@@ -139,4 +139,16 @@ describe Carbon::OAuthURLRequest do
     end
   end
 
+  describe 'test attribute "use_ocr"' do
+    it 'should work' do
+      # assertion here. ref: https://www.relishapp.com/rspec/rspec-expectations/docs/built-in-matchers
+    end
+  end
+
+  describe 'test attribute "parse_pdf_tables_with_ocr"' do
+    it 'should work' do
+      # assertion here. ref: https://www.relishapp.com/rspec/rspec-expectations/docs/built-in-matchers
+    end
+  end
+
 end

data/spec/models/organization_user_data_source_api_spec.rb

@@ -91,6 +91,12 @@ describe Carbon::OrganizationUserDataSourceAPI do
     end
   end
 
+  describe 'test attribute "enable_auto_sync"' do
+    it 'should work' do
+      # assertion here. ref: https://www.relishapp.com/rspec/rspec-expectations/docs/built-in-matchers
+    end
+  end
+
   describe 'test attribute "created_at"' do
     it 'should work' do
       # assertion here. ref: https://www.relishapp.com/rspec/rspec-expectations/docs/built-in-matchers

data/spec/models/s3_file_sync_input_spec.rb

@@ -91,4 +91,16 @@ describe Carbon::S3FileSyncInput do
     end
   end
 
+  describe 'test attribute "use_ocr"' do
+    it 'should work' do
+      # assertion here. ref: https://www.relishapp.com/rspec/rspec-expectations/docs/built-in-matchers
+    end
+  end
+
+  describe 'test attribute "parse_pdf_tables_with_ocr"' do
+    it 'should work' do
+      # assertion here. ref: https://www.relishapp.com/rspec/rspec-expectations/docs/built-in-matchers
+    end
+  end
+
 end

data/spec/models/sync_files_request_spec.rb

@@ -91,4 +91,16 @@ describe Carbon::SyncFilesRequest do
     end
   end
 
+  describe 'test attribute "use_ocr"' do
+    it 'should work' do
+      # assertion here. ref: https://www.relishapp.com/rspec/rspec-expectations/docs/built-in-matchers
+    end
+  end
+
+  describe 'test attribute "parse_pdf_tables_with_ocr"' do
+    it 'should work' do
+      # assertion here. ref: https://www.relishapp.com/rspec/rspec-expectations/docs/built-in-matchers
+    end
+  end
+
 end

data/spec/models/user_file_spec.rb

@@ -121,6 +121,12 @@ describe Carbon::UserFile do
     end
   end
 
+  describe 'test attribute "ocr_properties"' do
+    it 'should work' do
+      # assertion here. ref: https://www.relishapp.com/rspec/rspec-expectations/docs/built-in-matchers
+    end
+  end
+
   describe 'test attribute "name"' do
     it 'should work' do
       # assertion here. ref: https://www.relishapp.com/rspec/rspec-expectations/docs/built-in-matchers

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: carbon_ruby_sdk
 version: !ruby/object:Gem::Version
-  version: 0.1.13
+  version: 0.1.15
 platform: ruby
 authors:
 - Konfig
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-04-02 00:00:00.000000000 Z
+date: 2024-04-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: faraday