Feature - Default Category (#333)

* Update model methods to run background task when creating the endpoint and creating score endpoint for runpod to use * Handle error from Runpod, URL inference from yaml file * Update backend/app/api/endpoints/base/score.py Co-authored-by: Rafael Mosquera <[email protected]> * Rename method * rename endpoint * Print background tasks and correct typo * Correct typos and correct schemas * model uid * remove c * ad ta to metada (facepalm) * add update publish status to make change the publish state of the models * change order of background tasks * print instead of raising error for the second background task to run * add show category optional for Generative text * add functionality to remove user_input in artifacts for generative text --------- Co-authored-by: Rafael Mosquera <[email protected]>
mlcommons · Feb 5, 2025 · bd0598d · bd0598d
1 parent e54520d
commit bd0598d
Showing 1 changed file with 34 additions and 17 deletions.
diff --git a/...nts/CreateSamples/CreateSamples/AnnotationInterfaces/Contexts/EvaluateTextsGenerative.tsx b/...nts/CreateSamples/CreateSamples/AnnotationInterfaces/Contexts/EvaluateTextsGenerative.tsx
@@ -31,7 +31,7 @@ const EvaluateTextsGenerative: FC<
   const [signInConsent, setSignInConsent] = useState(true);
   const [showInput, setShowInput] = useState(false);
   const [showLoader, setShowLoader] = useState(false);
-  const [showCategory, setShowCategory] = useState(true);
+  const [showCategory, setShowCategory] = useState(false);
   const [finishConversation, setFinishConversation] = useState(false);
   const [disabledInput, setDisabledInput] = useState(false);
   const [disableTypeOfConversation, setDisableTypeOfConversation] =
@@ -47,15 +47,15 @@ const EvaluateTextsGenerative: FC<
   const [isCreatingTexts, setIsCreatingTexts] = useState(false);
   const [isTied, setIsTied] = useState(true);
   const [artifactsInput, setArtifactsInput] = useState<any>(
-    generative_context.artifacts,
+    generative_context.artifacts
   );
   const [prompt, setPrompt] = useState(
-    "Enter text here. Do not copy and paste",
+    "Enter text here. Do not copy and paste"
   );
   const { post, response } = useFetch();
   const { user } = useContext(UserContext);
   const { metadataExample, modelInputs, updateModelInputs } = useContext(
-    CreateInterfaceContext,
+    CreateInterfaceContext
   );
   const neccessaryFields = ["original_prompt", "category"];
 
@@ -87,7 +87,7 @@ const EvaluateTextsGenerative: FC<
         task_id: taskId,
         user_id: user.id,
         round_id: realRoundId,
-      },
+      }
     );
     if (response.ok) {
       if (redirectUrl) {
@@ -103,10 +103,6 @@ const EvaluateTextsGenerative: FC<
     }
   };
 
-  useEffect(() => {
-    checkIfUserReachedNecessaryExamples();
-  }, []);
-
   const generateTexts = async () => {
     if (neccessaryFields.every((item) => modelInputs.hasOwnProperty(item))) {
       setIsCreatingTexts(true);
@@ -129,7 +125,7 @@ const EvaluateTextsGenerative: FC<
           generatedTexts.map((text: any) => ({
             ...text,
             score: 50,
-          })),
+          }))
         );
         updateModelInputs({
           [field_names_for_the_model.generated_answers ?? "generated_answers"]:
@@ -179,18 +175,18 @@ const EvaluateTextsGenerative: FC<
 
   const checkIsTied = () => {
     const isTied = texts.every(
-      (text) => text.score === texts[0].score && texts.length > 1,
+      (text) => text.score === texts[0].score && texts.length > 1
     );
     setIsTied(isTied);
   };
 
   const checkNotAnswers = async (generatedTexts: any) => {
     // check if in some of the texts the provider name is None, in that case return True
     const notAnswers = generatedTexts.every(
-      (text: any) => text.provider === "None",
+      (text: any) => text.provider === "None"
     );
     const allTheAnswersAreEmpty = generatedTexts.every(
-      (text: any) => text.text === "\n",
+      (text: any) => text.text === "\n"
     );
     return notAnswers || allTheAnswersAreEmpty;
   };
@@ -199,13 +195,13 @@ const EvaluateTextsGenerative: FC<
     updateModelInputs({
       [field_names_for_the_model.best_answer ?? "best_answer"]: texts.reduce(
         (max: { score: number }, answer: { score: number }) =>
-          answer.score > max.score ? answer : max,
+          answer.score > max.score ? answer : max
       ),
     });
     setBestAnswer(
       texts.reduce((max: { score: number }, answer: { score: number }) =>
-        answer.score > max.score ? answer : max,
-      ),
+        answer.score > max.score ? answer : max
+      )
     );
 
     setChatHistory({
@@ -216,7 +212,7 @@ const EvaluateTextsGenerative: FC<
           id: "1",
           text: texts.reduce(
             (max: { score: number }, answer: { score: number }) =>
-              answer.score > max.score ? answer : max,
+              answer.score > max.score ? answer : max
           ).text,
         },
       ],
@@ -259,6 +255,27 @@ const EvaluateTextsGenerative: FC<
     setShowInput(true);
   };
 
+  useEffect(() => {
+    checkIfUserReachedNecessaryExamples();
+    if ("show_category" in artifactsInput) {
+      setShowCategory(artifactsInput?.show_category);
+      let category = "Unguided";
+      if ("default_category" in artifactsInput) {
+        category = artifactsInput?.default_category;
+      }
+      updateModelInputs({
+        category: category,
+      });
+      updateModelInputs({
+        initial_timestamp: Date.now(),
+      });
+      setShowInput(true);
+      setShowCategory(false);
+    } else {
+      setShowCategory(true);
+    }
+  }, []);
+
   useEffect(() => {
     updateModelInputs({
       [field_names_for_the_model.generated_answers ?? "generated_answers"]: