Merge branch 'develop' into 'fb-leap-2036/summary'

Workflow run: https://github.com/HumanSignal/label-studio/actions/runs/14956470496

Author: hlomzik

.github/CODEOWNERS

@@ -2,6 +2,7 @@
 deploy/           @HumanSignal/devops
 Dockerfile*       @HumanSignal/devops
 docker-compose*   @HumanSignal/devops
-docs/.npmignore   @hlomzik @Gondragos @nicholasrq
-docs/package.json @hlomzik @Gondragos @nicholasrq
-docs/yarn.lock    @hlomzik @Gondragos @nicholasrq
+docs/.npmignore   @hlomzik @Gondragos @nick-skriabin
+docs/package.json @hlomzik @Gondragos @nick-skriabin
+docs/yarn.lock    @hlomzik @Gondragos @nick-skriabin
+web/libs/editor   @hlomzik @Gondragos @nick-skriabin

label_studio/io_storages/README.md

@@ -184,4 +184,29 @@ The Storage Proxy API behavior can be configured using the following environment
 | `RESOLVER_PROXY_MAX_RANGE_SIZE` | Maximum size in bytes for a single range request | 7*1024*1024 |
 | `RESOLVER_PROXY_CACHE_TIMEOUT` | Cache TTL in seconds for proxy responses | 3600 |
 
-These optimizations ensure that the Proxy API remains responsive and resource-efficient, even when handling large files or many concurrent requests.
+These optimizations ensure that the Proxy API remains responsive and resource-efficient, even when handling large files or many concurrent requests.
+
+## Multiple Storages and URL Resolving
+
+There are use cases where multiple storages can/must be used in a single project. This can cause some confusion as to which storage gets used when. Here are some common cases and how to set up mutliple storages properly.
+
+### Case 1 - Tasks Referencing Other Buckets
+* bucket-A containing JSON tasks
+* bucket-B containing images/text/other data
+* Tasks synced from bucket-A have references to data in bucket-B
+
+##### How To Setup
+* Add storage 1 for bucket-A
+* Add storage 2 for bucket-B (might be same or different credentials than bucket-A)
+* Sync storage 1
+* All references to data in bucket-B will be resolved using storage 2 automatically
+
+### Case 2 - Buckets with Different Credentials
+* bucket-A accessible by credentials 1
+* bucket-B accessible by credentials 2
+
+##### How To Setup
+* Add storage 1 for bucket-A with credentials 1
+* Add storage 2 for bucket-B with credentials 2
+* Sync both storages
+* The appropriate storage will be used to resolve urls/generate presigned URLs

label_studio/io_storages/base_models.py

@@ -27,7 +27,7 @@
 from django.utils import timezone
 from django.utils.translation import gettext_lazy as _
 from django_rq import job
-from io_storages.utils import get_uri_via_regex
+from io_storages.utils import get_uri_via_regex, parse_bucket_uri
 from rq.job import Job
 from tasks.models import Annotation, Task
 from tasks.serializers import AnnotationSerializer, PredictionSerializer
@@ -255,8 +255,19 @@ def can_resolve_scheme(self, url: Union[str, None]) -> bool:
             return False
         # TODO: Search for occurrences inside string, e.g. for cases like "gs://bucket/file.pdf" or "<embed src='gs://bucket/file.pdf'/>"
         _, prefix = get_uri_via_regex(url, prefixes=(self.url_scheme,))
-        if prefix == self.url_scheme:
-            return True
+        bucket_uri = parse_bucket_uri(url, self)
+
+        # If there is a prefix and the bucket matches the storage's bucket/container/path
+        if prefix == self.url_scheme and bucket_uri:
+            # bucket is used for s3 and gcs
+            if hasattr(self, 'bucket') and bucket_uri.bucket == self.bucket:
+                return True
+            # container is used for azure blob
+            if hasattr(self, 'container') and bucket_uri.bucket == self.container:
+                return True
+            # path is used for redis
+            if hasattr(self, 'path') and bucket_uri.bucket == self.path:
+                return True
         # if not found any occurrences - this Storage can't resolve url
         return False
 

label_studio/io_storages/functions.py

@@ -54,6 +54,5 @@ def get_storage_by_url(url: Union[str, List, Dict], storage_objects: Iterable[Im
         for storage_object in storage_objects:
             if storage_object.can_resolve_url(url):
                 # note: only first found storage_object will be used for link resolving
-                # probably we need to use more advanced can_resolve_url mechanics
-                # that takes into account not only prefixes, but bucket path too
+                # can_resolve_url now checks both the scheme and the bucket to ensure the correct storage is used
                 return storage_object

label_studio/io_storages/s3/serializers.py

@@ -68,6 +68,8 @@ def validate(self, data):
         except TypeError as e:
             logger.info(f'It seems access keys are incorrect: {e}', exc_info=True)
             raise ValidationError('It seems access keys are incorrect')
+        except KeyError:
+            raise ValidationError(f'{storage.url_scheme}://{storage.bucket}/{storage.prefix} not found.')
         return data
 
 

label_studio/tasks/models.py

@@ -420,12 +420,10 @@ def prepare_filename(filename):
     def resolve_storage_uri(self, url) -> Optional[Mapping[str, Any]]:
         from io_storages.functions import get_storage_by_url
 
-        storage = self.storage
-        project = self.project
-
-        if not storage:
-            storage_objects = project.get_all_import_storage_objects
-            storage = get_storage_by_url(url, storage_objects)
+        # Instead of using self.storage, we check all storage objects for the project to
+        # support imported tasks that point to another bucket
+        storage_objects = self.project.get_all_import_storage_objects
+        storage = get_storage_by_url(url, storage_objects)
 
         if storage:
             return {
@@ -468,10 +466,9 @@ def resolve_uri(self, task_data, project):
 
                 # project storage
                 # TODO: to resolve nested lists and dicts we should improve get_storage_by_url(),
-                # TODO: problem with current approach: it can be used only the first storage that get_storage_by_url
-                # TODO: returns. However, maybe the second storage will resolve uris properly.
-                # TODO: resolve_uri() already supports them
-                storage = self.storage or get_storage_by_url(task_data[field], storage_objects)
+                # Now always using get_storage_by_url to ensure the storage with the correct bucket is used
+                # As a last fallback we can use self.storage which is the storage the Task was imported from
+                storage = get_storage_by_url(task_data[field], storage_objects) or self.storage
                 if storage:
                     try:
                         resolved_uri = storage.resolve_uri(task_data[field], self)

label_studio/tests/export.tavern.yml

@@ -139,13 +139,15 @@ stages:
         tags:
         - image segmentation
         - object detection
+        - keypoints
         name: COCO
       - title: COCO with Images
         description: COCO format with images downloaded.
         link: https://labelstud.io/guide/export.html#COCO
         tags:
         - image segmentation
         - object detection
+        - keypoints
         name: COCO_WITH_IMAGES
       - title: YOLO
         description: Popular TXT format is created for each image file. Each txt file contains
@@ -155,13 +157,15 @@ stages:
         tags:
         - image segmentation
         - object detection
+        - keypoints
         name: YOLO
       - title: YOLO with Images
         description: YOLO format with images downloaded.
         link: https://labelstud.io/guide/export.html#YOLO
         tags:
         - image segmentation
         - object detection
+        - keypoints
         name: YOLO_WITH_IMAGES
       - title: YOLOv8 OBB
         description: Popular TXT format is created for each image file. Each txt file contains

poetry.lock

@@ -73,7 +73,7 @@ dependencies = [
     "djangorestframework-simplejwt[crypto] (>=5.4.0,<6.0.0)",
     "tldextract (>=5.1.3)",
     ## HumanSignal repo dependencies :start
-    "label-studio-sdk @ https://github.com/HumanSignal/label-studio-sdk/archive/1ef88e2f8afe50a738fc1b03385cb1f3d6b2dd9f.zip",
+    "label-studio-sdk @ https://github.com/HumanSignal/label-studio-sdk/archive/c4ed0b0240c4a5e9f372393f84fdea01abe33141.zip",
     ## HumanSignal repo dependencies :end
 ]
 

pyproject.toml

@@ -8,28 +8,47 @@ import { DataManager } from "../DataManager/DataManager";
 import { Labeling } from "../Label/Label";
 import "./App.scss";
 
-class ErrorBoundary extends React.Component {
-  state = {
+interface ErrorBoundaryProps {
+  children: React.ReactNode;
+}
+
+interface ErrorBoundaryState {
+  error: Error | null;
+}
+
+class ErrorBoundary extends React.Component<ErrorBoundaryProps, ErrorBoundaryState> {
+  state: ErrorBoundaryState = {
     error: null,
   };
 
-  componentDidCatch(error) {
+  componentDidCatch(error: Error): void {
     this.setState({ error });
   }
 
-  render() {
-    return this.state.error ? <div className="error">{this.state.error}</div> : this.props.children;
+  render(): React.ReactNode {
+    return this.state.error ? <div className="error">{this.state.error.toString()}</div> : this.props.children;
   }
 }
 
+interface AppComponentProps {
+  app: {
+    SDK: {
+      mode: string;
+    };
+    crashed: boolean;
+    loading: boolean;
+    isLabeling: boolean;
+  };
+}
+
 /**
  * Main Component
- * @param {{app: import("../../stores/AppStore").AppStore} param0
  */
-const AppComponent = ({ app }) => {
+const AppComponent: React.FC<AppComponentProps> = ({ app }) => {
   const rootCN = cn("root");
   const rootClassName = rootCN.mod({ mode: app.SDK.mode }).toString();
   const crashCN = cn("crash");
+
   return (
     <ErrorBoundary>
       <Provider store={app}>

web/libs/datamanager/src/components/App/App.jsx renamed to web/libs/datamanager/src/components/App/App.tsx

@@ -19,27 +19,46 @@
   }
 
   &__action {
-    padding: 0;
-    margin: 0;
     margin: 2px;
     width: 28px;
     height: 28px;
     display: flex;
     align-items: center;
     justify-content: center;
     border-radius: 100%;
-    color: var(--sand_700);
     background-color: var(--color-neutral-surface);
+    padding: var(--spacing-tightest);
+    transition: all 150ms ease-out;
 
     &_type {
-      &_accept svg {
-        width: 15px;
-        height: 10px;
+      &_accept {
+        border: 1px solid var(--color-positive-border);
+        background-color: var(--color-positive-surface);
+        color: var(--color-positive-surface-content);
+
+        &:hover {
+          background-color: var(--color-positive-surface-hover);
+        }
+
+        & svg {
+          width: 20px;
+          height: 20px;
+        }
       }
 
-      &_reject svg {
-        width: 12.5px;
-        height: 12.5px;
+      &_reject {
+        border: 1px solid var(--color-negative-border);
+        background-color: var(--color-negative-surface);
+        color: var(--color-negative-surface-content);
+
+        &:hover {
+          background-color: var(--color-negative-surface-hover);
+        }
+
+        & svg {
+          width: 20px;
+          height: 20px;
+        }
       }
     }
   }

web/libs/editor/src/components/AnnotationTab/AutoAcceptToggle.scss

@@ -132,6 +132,18 @@
     display: grid;
     grid-template-columns: auto 1fr;
     grid-column-gap: 10px;
+    color: var(--color-neutral-content);
+
+    & .ant-typography {
+      color: var(--color-neutral-content);
+
+      & mark {
+        color: var(--color-neutral-content);
+        background-color: var(--color-primary-background);
+        padding: var(--spacing-tightest) var(--spacing-tighter);
+        border-radius: var(--corner-radius-smaller);
+      }
+    }
   }
 
   &__value {