feat: DIA-2214: Utilities to support PDF use cases (#452)

Co-authored-by: nik <[email protected]>
Co-authored-by: Copilot <[email protected]>

Author: 3 people

src/label_studio_sdk/label_interface/interface.py

@@ -8,9 +8,11 @@
 import json
 import jsonschema
 
+from functools import cached_property
 from typing import Dict, Optional, List, Tuple, Any, Callable, Union
 from pydantic import BaseModel
 
+
 # from typing import Dict, Optional, List, Tuple, Any
 from collections import defaultdict, OrderedDict
 from lxml import etree
@@ -517,6 +519,19 @@ def find_tags(
 
         return lst
 
+    @cached_property
+    def ner_tags(self):
+        return self.find_tags('controls', lambda t: t.tag.lower() in ('labels', 'hypertextlabels'))
+    
+    @cached_property
+    def image_tags(self):
+        return self.find_tags('objects', lambda t: t.tag.lower() == 'image')
+    
+    @cached_property
+    def pdf_tags(self):
+        return self.find_tags('objects', lambda t: t.tag.lower() == 'pdf')
+
+
     def load_task(self, task):
         """Loads a task and substitutes the value in each object tag
         with actual data from the task, returning a copy of the

src/label_studio_sdk/label_interface/object_tags.py

@@ -189,12 +189,17 @@ def _generate_example(self, examples, only_urls=False):
 
 
 class ImageTag(ObjectTag):
-    """ """
+    """Image tag"""
     tag: str = "Image"
 
     def _generate_example(self, examples, only_urls=False):
         """ """
         return examples.get("Image")
+    
+    @property
+    def is_image_list(self):
+        """Check if the tag is an image list, i.e. it has a valueList attribute that accepts list of images"""
+        return bool(self.attr.get("valueList")) if self.attr else False
 
 
 class TableTag(ObjectTag):