Merge pull request #268 from HumanSignal/fb-LEAP-608/support-Repeater…

…-in-lsc

feat: LEAP-608: Support Repeater in JSON-MIN exports

label_studio_converter/converter.py

@@ -1,6 +1,9 @@
+from typing import Optional
+
 import os
 import io
 import math
+import re
 import logging
 import ujson as json
 import ijson
@@ -374,6 +377,31 @@ def iter_from_json_file(self, json_file):
                         if item is not None:
                             yield item
 
+    def _maybe_matching_tag_from_schema(self, from_name: str) -> Optional[str]:
+        """If the from name exactly matches an output tag from the schema, return that tag.
+
+        Otherwise, certain tags (like those from Repeater) contain
+        placeholders like {{idx}}. Such placeholders are mapped to a regex in self._schema.
+        For example, if "my_output_tag_{{idx}}" is a tag in the schema,
+        then the from_name "my_output_tag_0" should match it, and we should return "my_output_tag_{{idx}}".
+        """
+
+        for tag_name, tag_info in self._schema.items():
+            if tag_name == from_name:
+                return tag_name
+
+            if not tag_info.get('regex'):
+                continue
+
+            tag_name_pattern = tag_name
+            for variable, regex in tag_info['regex'].items():
+                tag_name_pattern = tag_name_pattern.replace(variable, regex)
+
+            if re.compile(tag_name_pattern).match(from_name):
+                return tag_name
+
+        return None
+
     def annotation_result_from_task(self, task):
         has_annotations = 'completions' in task or 'annotations' in task
         if not has_annotations:
@@ -412,9 +440,9 @@ def annotation_result_from_task(self, task):
 
             # get results only as output
             for r in result:
-                if 'from_name' in r and r['from_name'] in self._output_tags:
+                if 'from_name' in r and (tag_name := self._maybe_matching_tag_from_schema(r['from_name'])):
                     v = deepcopy(r['value'])
-                    v['type'] = self._schema[r['from_name']]['type']
+                    v['type'] = self._schema[tag_name]['type']
                     if 'original_width' in r:
                         v['original_width'] = r['original_width']
                     if 'original_height' in r:

tests/__init__.py

@@ -0,0 +1 @@
+"""All tests for label-studio-converter should be written using pytest"""

tests/data/test_export_json_min/data.json

@@ -0,0 +1,79 @@
+[
+    {
+       "id": 87,
+       "annotations": [
+          {
+             "id": 22,
+             "completed_by": 1,
+             "result": [
+                {
+                   "id": "qd1jIVAq1h",
+                   "type": "polygonlabels",
+                   "value": {
+                      "closed": true,
+                      "points": [
+                         [
+                            44.44444444444444,
+                            48.25396825396825
+                         ],
+                         [
+                            42.53968253968254,
+                            60.317460317460316
+                         ],
+                         [
+                            56.19047619047619,
+                            50.79365079365079
+                         ]
+                      ],
+                      "polygonlabels": [
+                         "Airplane"
+                      ]
+                   },
+                   "origin": "manual",
+                   "to_name": "image",
+                   "from_name": "label",
+                   "image_rotation": 0,
+                   "original_width": 128,
+                   "original_height": 128
+                }
+             ],
+             "was_cancelled": false,
+             "ground_truth": false,
+             "created_at": "2024-01-30T02:40:43.893565Z",
+             "updated_at": "2024-01-30T02:40:43.893578Z",
+             "draft_created_at": "2024-01-30T02:40:42.917892Z",
+             "lead_time": 8.65,
+             "prediction": {},
+             "result_count": 0,
+             "unique_id": "3bc9da85-967f-47a1-bfa8-410206e54cf9",
+             "import_id": null,
+             "last_action": null,
+             "task": 87,
+             "project": 29,
+             "updated_by": 1,
+             "parent_prediction": null,
+             "parent_annotation": null,
+             "last_created_by": null
+          }
+       ],
+       "file_upload": "example.png",
+       "drafts": [],
+       "predictions": [],
+       "data": {
+          "image": "/data/upload/29/example.png"
+       },
+       "meta": {},
+       "created_at": "2024-01-30T02:40:33.867975Z",
+       "updated_at": "2024-01-30T02:40:43.923985Z",
+       "inner_id": 1,
+       "total_annotations": 1,
+       "cancelled_annotations": 0,
+       "total_predictions": 0,
+       "comment_count": 0,
+       "unresolved_comment_count": 0,
+       "last_comment_updated_at": null,
+       "project": 29,
+       "updated_by": 1,
+       "comment_authors": []
+    }
+ ]

tests/data/test_export_json_min/data_repeater.json

@@ -0,0 +1,101 @@
+[
+    {
+       "id": 86,
+       "annotations": [
+          {
+             "id": 21,
+             "completed_by": 1,
+             "result": [
+                {
+                   "id": "FmoGRgYXtw",
+                   "type": "rectanglelabels",
+                   "value": {
+                      "x": 19.466292134831463,
+                      "y": 53.70786516853933,
+                      "width": 20.35112359550562,
+                      "height": 20.224719101123597,
+                      "rotation": 0,
+                      "rectanglelabels": [
+                         "Document Title"
+                      ]
+                   },
+                   "origin": "manual",
+                   "to_name": "page_0",
+                   "from_name": "labels_0",
+                   "image_rotation": 0,
+                   "original_width": 1280,
+                   "original_height": 720
+                },
+                {
+                   "id": "FmoGRgYXtw",
+                   "type": "taxonomy",
+                   "value": {
+                      "x": 19.466292134831463,
+                      "y": 53.70786516853933,
+                      "width": 20.35112359550562,
+                      "height": 20.224719101123597,
+                      "rotation": 0,
+                      "taxonomy": [
+                         [
+                            "Archaea"
+                         ]
+                      ]
+                   },
+                   "origin": "manual",
+                   "to_name": "page_0",
+                   "from_name": "categories_0",
+                   "image_rotation": 0,
+                   "original_width": 1280,
+                   "original_height": 720
+                }
+             ],
+             "was_cancelled": false,
+             "ground_truth": false,
+             "created_at": "2024-01-25T23:52:15.182319Z",
+             "updated_at": "2024-01-25T23:52:15.182333Z",
+             "draft_created_at": "2024-01-25T23:52:13.732002Z",
+             "lead_time": 13.042,
+             "prediction": {},
+             "result_count": 0,
+             "unique_id": "5061fcba-4030-45b3-8079-71b7a134c254",
+             "import_id": null,
+             "last_action": null,
+             "task": 86,
+             "project": 28,
+             "updated_by": 1,
+             "parent_prediction": null,
+             "parent_annotation": null,
+             "last_created_by": null
+          }
+       ],
+       "file_upload": "a5ef7e93-task_copy.json",
+       "drafts": [],
+       "predictions": [],
+       "data": {
+          "images": [
+             {
+                "url": "https://example.org/example1.jpeg"
+             },
+             {
+                "url": "https://example.org/example2.jpeg"
+             },
+             {
+                "url": "https://example.org/example3.jpeg"
+             }
+          ]
+       },
+       "meta": {},
+       "created_at": "2024-01-25T23:52:00.852852Z",
+       "updated_at": "2024-01-25T23:52:15.207596Z",
+       "inner_id": 1,
+       "total_annotations": 1,
+       "cancelled_annotations": 0,
+       "total_predictions": 0,
+       "comment_count": 0,
+       "unresolved_comment_count": 0,
+       "last_comment_updated_at": null,
+       "project": 28,
+       "updated_by": 1,
+       "comment_authors": []
+    }
+ ]

tests/data/test_export_json_min/label_config.xml

@@ -0,0 +1,13 @@
+<View>
+
+  <Header value="Select label and click the image to start"/>
+  <Image name="image" value="$image" zoom="true"/>
+
+  <PolygonLabels name="label" toName="image"
+                 strokeWidth="3" pointSize="small"
+                 opacity="0.9">
+    <Label value="Airplane" background="red"/>
+    <Label value="Car" background="blue"/>
+  </PolygonLabels>
+
+</View>

tests/data/test_export_json_min/label_config_repeater.json

@@ -0,0 +1,72 @@
+{
+  "labels_{{idx}}": {
+     "type": "RectangleLabels",
+     "regex": {
+        "{{idx}}": ".*"
+     },
+     "inputs": [
+        {
+           "type": "Image",
+           "value": "images[{{idx}}].url"
+        }
+     ],
+     "labels": [
+        "Document Title",
+        "Document Date"
+     ],
+     "to_name": [
+        "page_{{idx}}"
+     ],
+     "labels_attrs": {
+        "Document Date": {
+           "value": "Document Date"
+        },
+        "Document Title": {
+           "value": "Document Title"
+        }
+     }
+  },
+  "categories_{{idx}}": {
+     "type": "Taxonomy",
+     "regex": {
+        "{{idx}}": ".*"
+     },
+     "inputs": [
+        {
+           "type": "Image",
+           "value": "images[{{idx}}].url"
+        }
+     ],
+     "labels": [
+        "Archaea",
+        "Bacteria",
+        "Eukarya",
+        "Human",
+        "Oppossum",
+        "Extraterrestrial"
+     ],
+     "to_name": [
+        "page_{{idx}}"
+     ],
+     "labels_attrs": {
+        "Human": {
+           "value": "Human"
+        },
+        "Archaea": {
+           "value": "Archaea"
+        },
+        "Eukarya": {
+           "value": "Eukarya"
+        },
+        "Bacteria": {
+           "value": "Bacteria"
+        },
+        "Oppossum": {
+           "value": "Oppossum"
+        },
+        "Extraterrestrial": {
+           "value": "Extraterrestrial"
+        }
+     }
+  }
+}

tests/test_brush.py

@@ -1,9 +1,9 @@
 """
 Test for the brush.py module
 """
-import unittest
 import urllib3
 import json
+import os
 
 from label_studio_converter.brush import encode_rle, image2annotation
 
@@ -13,7 +13,7 @@ def test_image2annotation():
     Import from png to LS annotation with RLE values
     """
     annotation = image2annotation(
-        'tests/test.png',
+        os.path.abspath(os.path.dirname(__file__)) + '/data/test_brush/test.png',
         label_name='Airplane',
         from_name='tag',
         to_name='image',
@@ -28,7 +28,7 @@ def test_image2annotation():
     }
 
     """ You can import this `task.json` to the Label Studio project with this labeling config:
-    
+
     <View>
       <Image name="image" value="$image" zoom="true"/>
       <BrushLabels name="tag" toName="image">