Inference.

app/api/api_v1/endpoints/openai.py

@@ -43,25 +43,6 @@ PIL.Image.MAX_IMAGE_PIXELS = None
 router = APIRouter()
 embeddings = OpenAIEmbeddings()
 
-def retry_with_backoff(retries = 5, backoff_in_seconds = 1):
-    def rwb(f):
-        def wrapper(*args, **kwargs):
-          x = 0
-          while True:
-            try:
-              return f(*args, **kwargs)
-            except:
-              if x == retries:
-                raise
-
-              sleep = (backoff_in_seconds * 2 ** x +
-                       random.uniform(0, 1))
-              time.sleep(sleep)
-              x += 1
-                
-        return wrapper
-    return rwb
-
 def convert_from_ls(result):
     if 'original_width' not in result or 'original_height' not in result:
         return None
@@ -86,14 +67,6 @@ def embed_documents(collection: CollectionStore):
         pre_delete_embeddings=True
     )
 
-@retry_with_backoff(retries=6)
-def export_tasks(task_id, tmp_archive):
-    label_studio_project.export_tasks("COCO", False, True, [task_id], tmp_archive)
-
-@retry_with_backoff(retries=6)
-def import_tasks(tmp_file):
-    return label_studio_project.import_tasks(tmp_file)
-
 def embed(db: Session, payload: schemas.LabelStudio):
 
     task_id = payload.task["id"]
@@ -112,7 +85,7 @@ def embed(db: Session, payload: schemas.LabelStudio):
 
     notebook_dir = "{}/notebooks".format(tmp_dir)
 
-    export_tasks(task_id, tmp_archive)
+    label_studio_project.export_tasks("COCO", False, True, [task_id], tmp_archive)
 
     with zipfile.ZipFile(tmp_archive, 'r') as zip_ref:
         zip_ref.extractall(tmp_dir)
@@ -207,7 +180,7 @@ async def create_collection(
     target_height = round(height - (height * 0.75))
     cover = cover.resize((target_width, target_height), PIL.Image.Resampling.LANCZOS)
     cover.save(tmp_file, optimize=True, quality=95)
-    tasks = import_tasks(tmp_file)
+    tasks = label_studio_project.import_tasks(tmp_file)
     os.remove(tmp_file)