feat(swagger): update swagger (#9824)

Signed-off-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: mudler <2420543+mudler@users.noreply.github.com>

Author: localai-bot

swagger/docs.go

@@ -5347,6 +5347,14 @@ const docTemplate = `{
                 "stream": {
                     "type": "boolean"
                 },
+                "stream_options": {
+                    "description": "StreamOptions opts into OpenAI streaming extensions, e.g. include_usage.",
+                    "allOf": [
+                        {
+                            "$ref": "#/definitions/schema.StreamOptions"
+                        }
+                    ]
+                },
                 "temperature": {
                     "type": "number"
                 },
@@ -5412,7 +5420,12 @@ const docTemplate = `{
                     "type": "string"
                 },
                 "usage": {
-                    "$ref": "#/definitions/schema.OpenAIUsage"
+                    "description": "Usage is intentionally a pointer with omitempty: per the OpenAI\nchat-completion streaming spec, intermediate chunks must not carry\na ` + "`" + `usage` + "`" + ` field. Marshalling a value-typed usage would emit\n` + "`" + `\"usage\":{\"prompt_tokens\":0,...}` + "`" + ` on every chunk and break\nOpenAI-SDK consumers that filter on a truthy ` + "`" + `result.usage` + "`" + `\n(continuedev/continue, Kilo Code, Roo Code, etc.).",
+                    "allOf": [
+                        {
+                            "$ref": "#/definitions/schema.OpenAIUsage"
+                        }
+                    ]
                 }
             }
         },
@@ -5578,6 +5591,14 @@ const docTemplate = `{
                 }
             }
         },
+        "schema.StreamOptions": {
+            "type": "object",
+            "properties": {
+                "include_usage": {
+                    "type": "boolean"
+                }
+            }
+        },
         "schema.SysInfoModel": {
             "type": "object",
             "properties": {

swagger/swagger.json

@@ -5344,6 +5344,14 @@
                 "stream": {
                     "type": "boolean"
                 },
+                "stream_options": {
+                    "description": "StreamOptions opts into OpenAI streaming extensions, e.g. include_usage.",
+                    "allOf": [
+                        {
+                            "$ref": "#/definitions/schema.StreamOptions"
+                        }
+                    ]
+                },
                 "temperature": {
                     "type": "number"
                 },
@@ -5409,7 +5417,12 @@
                     "type": "string"
                 },
                 "usage": {
-                    "$ref": "#/definitions/schema.OpenAIUsage"
+                    "description": "Usage is intentionally a pointer with omitempty: per the OpenAI\nchat-completion streaming spec, intermediate chunks must not carry\na `usage` field. Marshalling a value-typed usage would emit\n`\"usage\":{\"prompt_tokens\":0,...}` on every chunk and break\nOpenAI-SDK consumers that filter on a truthy `result.usage`\n(continuedev/continue, Kilo Code, Roo Code, etc.).",
+                    "allOf": [
+                        {
+                            "$ref": "#/definitions/schema.OpenAIUsage"
+                        }
+                    ]
                 }
             }
         },
@@ -5575,6 +5588,14 @@
                 }
             }
         },
+        "schema.StreamOptions": {
+            "type": "object",
+            "properties": {
+                "include_usage": {
+                    "type": "boolean"
+                }
+            }
+        },
         "schema.SysInfoModel": {
             "type": "object",
             "properties": {

swagger/swagger.yaml

@@ -1650,6 +1650,10 @@ definitions:
       stop: {}
       stream:
         type: boolean
+      stream_options:
+        allOf:
+        - $ref: '#/definitions/schema.StreamOptions'
+        description: StreamOptions opts into OpenAI streaming extensions, e.g. include_usage.
       temperature:
         type: number
       tfz:
@@ -1698,7 +1702,15 @@ definitions:
       object:
         type: string
       usage:
-        $ref: '#/definitions/schema.OpenAIUsage'
+        allOf:
+        - $ref: '#/definitions/schema.OpenAIUsage'
+        description: |-
+          Usage is intentionally a pointer with omitempty: per the OpenAI
+          chat-completion streaming spec, intermediate chunks must not carry
+          a `usage` field. Marshalling a value-typed usage would emit
+          `"usage":{"prompt_tokens":0,...}` on every chunk and break
+          OpenAI-SDK consumers that filter on a truthy `result.usage`
+          (continuedev/continue, Kilo Code, Roo Code, etc.).
     type: object
   schema.OpenAIUsage:
     properties:
@@ -1813,6 +1825,11 @@ definitions:
           $ref: '#/definitions/schema.NodeData'
         type: array
     type: object
+  schema.StreamOptions:
+    properties:
+      include_usage:
+        type: boolean
+    type: object
   schema.SysInfoModel:
     properties:
       id: