feat:OpenAPI updates: nullable/oneOf, think/truncate/shift, remote metadata

src/libs/Ollama/Generated/Ollama..JsonSerializerContext.g.cs

@@ -13,12 +13,20 @@ namespace Ollama
         DefaultIgnoreCondition = global::System.Text.Json.Serialization.JsonIgnoreCondition.WhenWritingNull,
         Converters = new global::System.Type[] 
         { 
+            typeof(global::Ollama.JsonConverters.GenerateCompletionRequestFormatEnumJsonConverter),
+            typeof(global::Ollama.JsonConverters.GenerateCompletionRequestFormatEnumNullableJsonConverter),
+            typeof(global::Ollama.JsonConverters.GenerateCompletionRequestThinkJsonConverter),
+            typeof(global::Ollama.JsonConverters.GenerateCompletionRequestThinkNullableJsonConverter),
             typeof(global::Ollama.JsonConverters.ResponseFormatEnumJsonConverter),
             typeof(global::Ollama.JsonConverters.ResponseFormatEnumNullableJsonConverter),
             typeof(global::Ollama.JsonConverters.MessageRoleJsonConverter),
             typeof(global::Ollama.JsonConverters.MessageRoleNullableJsonConverter),
+            typeof(global::Ollama.JsonConverters.GenerateChatCompletionRequestFormatEnumJsonConverter),
+            typeof(global::Ollama.JsonConverters.GenerateChatCompletionRequestFormatEnumNullableJsonConverter),
             typeof(global::Ollama.JsonConverters.ToolTypeJsonConverter),
             typeof(global::Ollama.JsonConverters.ToolTypeNullableJsonConverter),
+            typeof(global::Ollama.JsonConverters.GenerateChatCompletionRequestThinkJsonConverter),
+            typeof(global::Ollama.JsonConverters.GenerateChatCompletionRequestThinkNullableJsonConverter),
             typeof(global::Ollama.JsonConverters.DoneReasonEnumJsonConverter),
             typeof(global::Ollama.JsonConverters.DoneReasonEnumNullableJsonConverter),
             typeof(global::Ollama.JsonConverters.CreateModelStatusEnumJsonConverter),
@@ -33,6 +41,10 @@ namespace Ollama
             typeof(global::Ollama.JsonConverters.DoneReasonJsonConverter),
             typeof(global::Ollama.JsonConverters.CreateModelStatusJsonConverter),
             typeof(global::Ollama.JsonConverters.PullModelStatusJsonConverter),
+            typeof(global::Ollama.JsonConverters.OneOfJsonConverter<global::Ollama.GenerateCompletionRequestFormatEnum?, object>),
+            typeof(global::Ollama.JsonConverters.OneOfJsonConverter<bool?, global::Ollama.GenerateCompletionRequestThink?>),
+            typeof(global::Ollama.JsonConverters.OneOfJsonConverter<global::Ollama.GenerateChatCompletionRequestFormatEnum?, object>),
+            typeof(global::Ollama.JsonConverters.OneOfJsonConverter<bool?, global::Ollama.GenerateChatCompletionRequestThink?>),
             typeof(global::Ollama.JsonConverters.AnyOfJsonConverter<string, global::Ollama.PushModelResponseStatus?>),
             typeof(global::Ollama.JsonConverters.UnixTimestampJsonConverter),
         })]

src/libs/Ollama/Generated/Ollama.ChatClient.GenerateChatCompletion.g.cs

@@ -123,7 +123,11 @@ partial void ProcessGenerateChatCompletionResponse(
         /// If `false` the response will be returned as a single response object, otherwise the response will be streamed as a series of objects.<br/>
         /// Default Value: true
         /// </param>
-        /// <param name="format"></param>
+        /// <param name="format">
+        /// The format to return a response in. Can be:<br/>
+        /// - "json" string to enable JSON mode<br/>
+        /// - JSON schema object for structured output validation
+        /// </param>
         /// <param name="keepAlive">
         /// How long (in minutes) to keep the model loaded in memory.<br/>
         /// - If set to a positive duration (e.g. 20), the model will stay loaded for the provided duration.<br/>
@@ -138,22 +142,30 @@ partial void ProcessGenerateChatCompletionResponse(
         /// Additional model parameters listed in the documentation for the Modelfile such as `temperature`.
         /// </param>
         /// <param name="think">
-        /// Think controls whether thinking/reasoning models will think before<br/>
-        /// responding. Needs to be a pointer so we can distinguish between false<br/>
-        /// (request that thinking _not_ be used) and unset (use the old behavior<br/>
-        /// before this option was introduced).
+        /// Controls whether thinking/reasoning models will think before responding.<br/>
+        /// Can be:<br/>
+        /// - boolean: true/false to enable/disable thinking<br/>
+        /// - string: "high", "medium", "low" to set thinking intensity level
+        /// </param>
+        /// <param name="truncate">
+        /// Truncates the end of the chat history if it exceeds the context length
+        /// </param>
+        /// <param name="shift">
+        /// Shifts the oldest messages out of the context window when the context limit is reached
         /// </param>
         /// <param name="cancellationToken">The token to cancel the operation with</param>
         /// <exception cref="global::System.InvalidOperationException"></exception>
         public async global::System.Collections.Generic.IAsyncEnumerable<global::Ollama.GenerateChatCompletionResponse> GenerateChatCompletionAsync(
             string model,
             global::System.Collections.Generic.IList<global::Ollama.Message> messages,
             bool? stream = default,
-            global::Ollama.ResponseFormat? format = default,
+            global::Ollama.OneOf<global::Ollama.GenerateChatCompletionRequestFormatEnum?, object>? format = default,
             int? keepAlive = default,
             global::System.Collections.Generic.IList<global::Ollama.Tool>? tools = default,
             global::Ollama.RequestOptions? options = default,
-            bool? think = default,
+            global::Ollama.OneOf<bool?, global::Ollama.GenerateChatCompletionRequestThink?>? think = default,
+            bool? truncate = default,
+            bool? shift = default,
             [global::System.Runtime.CompilerServices.EnumeratorCancellation] global::System.Threading.CancellationToken cancellationToken = default)
         {
             var __request = new global::Ollama.GenerateChatCompletionRequest
@@ -166,6 +178,8 @@ partial void ProcessGenerateChatCompletionResponse(
                 Tools = tools,
                 Options = options,
                 Think = think,
+                Truncate = truncate,
+                Shift = shift,
             };
 
             var __enumerable = GenerateChatCompletionAsync(

src/libs/Ollama/Generated/Ollama.CompletionsClient.GenerateCompletion.g.cs

@@ -137,10 +137,14 @@ partial void ProcessGenerateCompletionResponse(
         /// Default Value: true
         /// </param>
         /// <param name="raw">
-        /// If `true` no formatting will be applied to the prompt and no context will be returned. <br/>
+        /// If `true` no formatting will be applied to the prompt and no context will be returned.<br/>
         /// You may choose to use the `raw` parameter if you are specifying a full templated prompt in your request to the API, and are managing history yourself.
         /// </param>
-        /// <param name="format"></param>
+        /// <param name="format">
+        /// The format to return a response in. Can be:<br/>
+        /// - "json" string to enable JSON mode<br/>
+        /// - JSON schema object for structured output validation
+        /// </param>
         /// <param name="keepAlive">
         /// How long (in minutes) to keep the model loaded in memory.<br/>
         /// - If set to a positive duration (e.g. 20), the model will stay loaded for the provided duration.<br/>
@@ -155,10 +159,16 @@ partial void ProcessGenerateCompletionResponse(
         /// Additional model parameters listed in the documentation for the Modelfile such as `temperature`.
         /// </param>
         /// <param name="think">
-        /// Think controls whether thinking/reasoning models will think before<br/>
-        /// responding. Needs to be a pointer so we can distinguish between false<br/>
-        /// (request that thinking _not_ be used) and unset (use the old behavior<br/>
-        /// before this option was introduced).
+        /// Controls whether thinking/reasoning models will think before responding.<br/>
+        /// Can be:<br/>
+        /// - boolean: true/false to enable/disable thinking<br/>
+        /// - string: "high", "medium", "low" to set thinking intensity level
+        /// </param>
+        /// <param name="truncate">
+        /// Truncates the end of the prompt if it exceeds the context length
+        /// </param>
+        /// <param name="shift">
+        /// Shifts the oldest parts out of the context window when the context limit is reached
         /// </param>
         /// <param name="cancellationToken">The token to cancel the operation with</param>
         /// <exception cref="global::System.InvalidOperationException"></exception>
@@ -171,11 +181,13 @@ partial void ProcessGenerateCompletionResponse(
             global::System.Collections.Generic.IList<long>? context = default,
             bool? stream = default,
             bool? raw = default,
-            global::Ollama.ResponseFormat? format = default,
+            global::Ollama.OneOf<global::Ollama.GenerateCompletionRequestFormatEnum?, object>? format = default,
             int? keepAlive = default,
             global::System.Collections.Generic.IList<string>? images = default,
             global::Ollama.RequestOptions? options = default,
-            bool? think = default,
+            global::Ollama.OneOf<bool?, global::Ollama.GenerateCompletionRequestThink?>? think = default,
+            bool? truncate = default,
+            bool? shift = default,
             [global::System.Runtime.CompilerServices.EnumeratorCancellation] global::System.Threading.CancellationToken cancellationToken = default)
         {
             var __request = new global::Ollama.GenerateCompletionRequest
@@ -193,6 +205,8 @@ partial void ProcessGenerateCompletionResponse(
                 Images = images,
                 Options = options,
                 Think = think,
+                Truncate = truncate,
+                Shift = shift,
             };
 
             var __enumerable = GenerateCompletionAsync(

src/libs/Ollama/Generated/Ollama.IChatClient.GenerateChatCompletion.g.cs

@@ -31,7 +31,11 @@ public partial interface IChatClient
         /// If `false` the response will be returned as a single response object, otherwise the response will be streamed as a series of objects.<br/>
         /// Default Value: true
         /// </param>
-        /// <param name="format"></param>
+        /// <param name="format">
+        /// The format to return a response in. Can be:<br/>
+        /// - "json" string to enable JSON mode<br/>
+        /// - JSON schema object for structured output validation
+        /// </param>
         /// <param name="keepAlive">
         /// How long (in minutes) to keep the model loaded in memory.<br/>
         /// - If set to a positive duration (e.g. 20), the model will stay loaded for the provided duration.<br/>
@@ -46,22 +50,30 @@ public partial interface IChatClient
         /// Additional model parameters listed in the documentation for the Modelfile such as `temperature`.
         /// </param>
         /// <param name="think">
-        /// Think controls whether thinking/reasoning models will think before<br/>
-        /// responding. Needs to be a pointer so we can distinguish between false<br/>
-        /// (request that thinking _not_ be used) and unset (use the old behavior<br/>
-        /// before this option was introduced).
+        /// Controls whether thinking/reasoning models will think before responding.<br/>
+        /// Can be:<br/>
+        /// - boolean: true/false to enable/disable thinking<br/>
+        /// - string: "high", "medium", "low" to set thinking intensity level
+        /// </param>
+        /// <param name="truncate">
+        /// Truncates the end of the chat history if it exceeds the context length
+        /// </param>
+        /// <param name="shift">
+        /// Shifts the oldest messages out of the context window when the context limit is reached
         /// </param>
         /// <param name="cancellationToken">The token to cancel the operation with</param>
         /// <exception cref="global::System.InvalidOperationException"></exception>
         global::System.Collections.Generic.IAsyncEnumerable<global::Ollama.GenerateChatCompletionResponse> GenerateChatCompletionAsync(
             string model,
             global::System.Collections.Generic.IList<global::Ollama.Message> messages,
             bool? stream = default,
-            global::Ollama.ResponseFormat? format = default,
+            global::Ollama.OneOf<global::Ollama.GenerateChatCompletionRequestFormatEnum?, object>? format = default,
             int? keepAlive = default,
             global::System.Collections.Generic.IList<global::Ollama.Tool>? tools = default,
             global::Ollama.RequestOptions? options = default,
-            bool? think = default,
+            global::Ollama.OneOf<bool?, global::Ollama.GenerateChatCompletionRequestThink?>? think = default,
+            bool? truncate = default,
+            bool? shift = default,
             global::System.Threading.CancellationToken cancellationToken = default);
     }
 }

src/libs/Ollama/Generated/Ollama.ICompletionsClient.GenerateCompletion.g.cs

@@ -45,10 +45,14 @@ public partial interface ICompletionsClient
         /// Default Value: true
         /// </param>
         /// <param name="raw">
-        /// If `true` no formatting will be applied to the prompt and no context will be returned. <br/>
+        /// If `true` no formatting will be applied to the prompt and no context will be returned.<br/>
         /// You may choose to use the `raw` parameter if you are specifying a full templated prompt in your request to the API, and are managing history yourself.
         /// </param>
-        /// <param name="format"></param>
+        /// <param name="format">
+        /// The format to return a response in. Can be:<br/>
+        /// - "json" string to enable JSON mode<br/>
+        /// - JSON schema object for structured output validation
+        /// </param>
         /// <param name="keepAlive">
         /// How long (in minutes) to keep the model loaded in memory.<br/>
         /// - If set to a positive duration (e.g. 20), the model will stay loaded for the provided duration.<br/>
@@ -63,10 +67,16 @@ public partial interface ICompletionsClient
         /// Additional model parameters listed in the documentation for the Modelfile such as `temperature`.
         /// </param>
         /// <param name="think">
-        /// Think controls whether thinking/reasoning models will think before<br/>
-        /// responding. Needs to be a pointer so we can distinguish between false<br/>
-        /// (request that thinking _not_ be used) and unset (use the old behavior<br/>
-        /// before this option was introduced).
+        /// Controls whether thinking/reasoning models will think before responding.<br/>
+        /// Can be:<br/>
+        /// - boolean: true/false to enable/disable thinking<br/>
+        /// - string: "high", "medium", "low" to set thinking intensity level
+        /// </param>
+        /// <param name="truncate">
+        /// Truncates the end of the prompt if it exceeds the context length
+        /// </param>
+        /// <param name="shift">
+        /// Shifts the oldest parts out of the context window when the context limit is reached
         /// </param>
         /// <param name="cancellationToken">The token to cancel the operation with</param>
         /// <exception cref="global::System.InvalidOperationException"></exception>
@@ -79,11 +89,13 @@ public partial interface ICompletionsClient
             global::System.Collections.Generic.IList<long>? context = default,
             bool? stream = default,
             bool? raw = default,
-            global::Ollama.ResponseFormat? format = default,
+            global::Ollama.OneOf<global::Ollama.GenerateCompletionRequestFormatEnum?, object>? format = default,
             int? keepAlive = default,
             global::System.Collections.Generic.IList<string>? images = default,
             global::Ollama.RequestOptions? options = default,
-            bool? think = default,
+            global::Ollama.OneOf<bool?, global::Ollama.GenerateCompletionRequestThink?>? think = default,
+            bool? truncate = default,
+            bool? shift = default,
             global::System.Threading.CancellationToken cancellationToken = default);
     }
 }

src/libs/Ollama/Generated/Ollama.IModelsClient.CreateModel.g.cs

@@ -38,14 +38,38 @@ public partial interface IModelsClient
         /// If `false` the response will be returned as a single response object, otherwise the response will be streamed as a series of objects.<br/>
         /// Default Value: true
         /// </param>
+        /// <param name="from">
+        /// Name of the model or file to use as the source
+        /// </param>
+        /// <param name="files">
+        /// Map of files to include when creating the model
+        /// </param>
+        /// <param name="adapters">
+        /// Map of LoRA adapters to include when creating the model
+        /// </param>
+        /// <param name="template">
+        /// Template used when constructing a request to the model
+        /// </param>
+        /// <param name="system">
+        /// System prompt for the model
+        /// </param>
+        /// <param name="parameters">
+        /// Map of hyper-parameters which are applied to the model
+        /// </param>
         /// <param name="cancellationToken">The token to cancel the operation with</param>
         /// <exception cref="global::System.InvalidOperationException"></exception>
         global::System.Collections.Generic.IAsyncEnumerable<global::Ollama.CreateModelResponse> CreateModelAsync(
             string model,
-            string modelfile,
+            string? modelfile = default,
             string? path = default,
             string? quantize = default,
             bool? stream = default,
+            string? from = default,
+            global::System.Collections.Generic.Dictionary<string, string>? files = default,
+            global::System.Collections.Generic.Dictionary<string, string>? adapters = default,
+            string? template = default,
+            string? system = default,
+            object? parameters = default,
             global::System.Threading.CancellationToken cancellationToken = default);
     }
 }