Fixing missing object field for regular completions. (#2175)

* Fixing missing `object` field for regular completions.

* Fixing docs by re-adding missing `Prompt`.

Author: Narsil

docs/openapi.json

@@ -946,6 +946,38 @@
           }
         }
       },
+      "Chunk": {
+        "type": "object",
+        "required": [
+          "id",
+          "created",
+          "choices",
+          "model",
+          "system_fingerprint"
+        ],
+        "properties": {
+          "choices": {
+            "type": "array",
+            "items": {
+              "$ref": "#/components/schemas/CompletionComplete"
+            }
+          },
+          "created": {
+            "type": "integer",
+            "format": "int64",
+            "minimum": 0
+          },
+          "id": {
+            "type": "string"
+          },
+          "model": {
+            "type": "string"
+          },
+          "system_fingerprint": {
+            "type": "string"
+          }
+        }
+      },
       "CompatGenerateRequest": {
         "type": "object",
         "required": [
@@ -965,6 +997,55 @@
           }
         }
       },
+      "Completion": {
+        "oneOf": [
+          {
+            "allOf": [
+              {
+                "$ref": "#/components/schemas/Chunk"
+              },
+              {
+                "type": "object",
+                "required": [
+                  "object"
+                ],
+                "properties": {
+                  "object": {
+                    "type": "string",
+                    "enum": [
+                      "text_completion"
+                    ]
+                  }
+                }
+              }
+            ]
+          },
+          {
+            "allOf": [
+              {
+                "$ref": "#/components/schemas/CompletionFinal"
+              },
+              {
+                "type": "object",
+                "required": [
+                  "object"
+                ],
+                "properties": {
+                  "object": {
+                    "type": "string",
+                    "enum": [
+                      "text_completion"
+                    ]
+                  }
+                }
+              }
+            ]
+          }
+        ],
+        "discriminator": {
+          "propertyName": "object"
+        }
+      },
       "CompletionComplete": {
         "type": "object",
         "required": [
@@ -994,14 +1075,15 @@
           }
         }
       },
-      "CompletionCompleteChunk": {
+      "CompletionFinal": {
         "type": "object",
         "required": [
           "id",
           "created",
-          "choices",
           "model",
-          "system_fingerprint"
+          "system_fingerprint",
+          "choices",
+          "usage"
         ],
         "properties": {
           "choices": {
@@ -1013,16 +1095,21 @@
           "created": {
             "type": "integer",
             "format": "int64",
+            "example": "1706270835",
             "minimum": 0
           },
           "id": {
             "type": "string"
           },
           "model": {
-            "type": "string"
+            "type": "string",
+            "example": "mistralai/Mistral-7B-Instruct-v0.2"
           },
           "system_fingerprint": {
             "type": "string"
+          },
+          "usage": {
+            "$ref": "#/components/schemas/Usage"
           }
         }
       },
@@ -1647,6 +1734,12 @@
           }
         }
       },
+      "Prompt": {
+        "type": "array",
+        "items": {
+          "type": "string"
+        }
+      },
       "SimpleToken": {
         "type": "object",
         "required": [

router/src/lib.rs

@@ -433,8 +433,17 @@ pub struct CompletionRequest {
     pub stop: Option<Vec<String>>,
 }
 
+#[derive(Clone, Serialize, ToSchema)]
+#[serde(tag = "object")]
+enum Completion {
+    #[serde(rename = "text_completion")]
+    Chunk(Chunk),
+    #[serde(rename = "text_completion")]
+    Final(CompletionFinal),
+}
+
 #[derive(Clone, Deserialize, Serialize, ToSchema, Default)]
-pub(crate) struct Completion {
+pub(crate) struct CompletionFinal {
     pub id: String,
     #[schema(example = "1706270835")]
     pub created: u64,
@@ -453,6 +462,15 @@ pub(crate) struct CompletionComplete {
     pub finish_reason: String,
 }
 
+#[derive(Clone, Deserialize, Serialize, ToSchema)]
+pub(crate) struct Chunk {
+    pub id: String,
+    pub created: u64,
+    pub choices: Vec<CompletionComplete>,
+    pub model: String,
+    pub system_fingerprint: String,
+}
+
 #[derive(Clone, Deserialize, Serialize, ToSchema)]
 pub(crate) struct ChatCompletion {
     pub id: String,
@@ -614,15 +632,6 @@ impl ChatCompletion {
         }
     }
 }
-#[derive(Clone, Deserialize, Serialize, ToSchema)]
-pub(crate) struct CompletionCompleteChunk {
-    pub id: String,
-    pub created: u64,
-    pub choices: Vec<CompletionComplete>,
-    pub model: String,
-    pub system_fingerprint: String,
-}
-
 #[derive(Clone, Serialize, ToSchema)]
 pub(crate) struct ChatCompletionChunk {
     pub id: String,

router/src/server.rs

@@ -19,8 +19,8 @@ use crate::{
 use crate::{
     ChatCompletion, ChatCompletionChoice, ChatCompletionChunk, ChatCompletionComplete,
     ChatCompletionDelta, ChatCompletionLogprob, ChatCompletionLogprobs, ChatCompletionTopLogprob,
-    ChatRequest, CompatGenerateRequest, Completion, CompletionComplete, CompletionCompleteChunk,
-    CompletionRequest, CompletionType, DeltaToolCall, Function, Tool, VertexRequest,
+    ChatRequest, Chunk, CompatGenerateRequest, Completion, CompletionComplete, CompletionFinal,
+    CompletionRequest, CompletionType, DeltaToolCall, Function, Prompt, Tool, VertexRequest,
     VertexResponse,
 };
 use crate::{FunctionDefinition, HubPreprocessorConfig, ToolCall, ToolType};
@@ -706,7 +706,7 @@ async fn completions(
                         .as_secs();
 
                     event
-                        .json_data(CompletionCompleteChunk {
+                        .json_data(Completion::Chunk(Chunk {
                             id: "".to_string(),
                             created: current_time,
 
@@ -719,7 +719,7 @@ async fn completions(
 
                             model: model_id.clone(),
                             system_fingerprint: system_fingerprint.clone(),
-                        })
+                        }))
                         .unwrap_or_else(|_e| Event::default())
                 };
 
@@ -932,7 +932,7 @@ async fn completions(
             .collect::<Result<Vec<_>, _>>()
             .map_err(|(status, Json(err))| (status, Json(err)))?;
 
-        let response = Completion {
+        let response = Completion::Final(CompletionFinal {
             id: "".to_string(),
             created: current_time,
             model: info.model_id.clone(),
@@ -947,7 +947,7 @@ async fn completions(
                 completion_tokens,
                 total_tokens,
             },
-        };
+        });
 
         // headers similar to `generate` but aggregated
         let mut headers = HeaderMap::new();
@@ -1466,7 +1466,10 @@ pub async fn run(
     ChatCompletion,
     CompletionRequest,
     CompletionComplete,
-    CompletionCompleteChunk,
+    Chunk,
+    Completion,
+    CompletionFinal,
+    Prompt,
     GenerateParameters,
     PrefillToken,
     Token,