Reorders labelstudio import (#76)

* Reorders labelstudio import

* Adds hex colors and colors for selections

* Allows outlier slices to be exported

* Websocket integration outlier slice

Co-authored-by: JWittmeyer <jens.wittmeyer@onetask.ai>

Author: JWittmeyer

controller/transfer/export_parser.py

@@ -3,6 +3,7 @@
 
 from typing import Any, Dict, List, Optional, Tuple, Union
 from submodules.model.business_objects import attribute, general, project, data_slice
+from .labelstudio import export_parser as ls_export_parser
 
 import pandas as pd
 import numpy as np
@@ -31,7 +32,6 @@ def parse(
             if str(col).endswith("__created_by"):
                 df.drop(col, axis="columns", inplace=True)
     elif export_format == enums.RecordExportFormats.LABEL_STUDIO.value:
-        from .labelstudio import export_parser as ls_export_parser
 
         df = ls_export_parser.parse_dataframe_data(project_id, df)
     else:
@@ -61,7 +61,10 @@ def infer_file_name(
     row_option = export_options.get("rows")
     if row_option.get("type") == enums.RecordExportAmountTypes.SLICE.value:
         slice_item = data_slice.get(project_id, row_option.get("id"))
-        amount_type_addition = slice_item.name
+        if slice_item.slice_type == enums.SliceTypes.STATIC_OUTLIER.value:
+            amount_type_addition = "outlier"
+        else:
+            amount_type_addition = slice_item.name
     else:
         amount_type_addition = row_option.get("type")
 

controller/transfer/labelstudio/template_generator.py

@@ -4,6 +4,27 @@
 
 INDENT_SIZE = "    "
 
+# tested a bit and text-color-400 seems to be the best case -> can be adjusted if necessary
+COLOR_LOOKUP = {
+    "rose": "#fb7185",
+    "pink": "#f472b6",
+    "fuchsia": "#e879f9",
+    "purple": "#c084fc",
+    "violet": "#a78bfa",
+    "indigo": "#818cf8",
+    "blue": "#60a5fa",
+    "sky": "#38bdf8",
+    "cyan": "#22d3ee",
+    "teal": "#2dd4bf",
+    "emerald": "#34d399",
+    "green": "#4ade80",
+    "lime": "#a3e635",
+    "yellow": "#facc15",
+    "amber": "#fbbf24",
+    "orange": "#fb923c",
+    "red": "#f87171",
+}
+
 
 def generate_template(
     project_id: str, labeling_task_ids: List[str], attribute_ids: List[str]
@@ -48,7 +69,7 @@ def __generate_extraction_task(
         f'{INDENT_SIZE}<Labels name="{task_data["name"]}" toName="{task_data["attribute_name"]}">'
     )
     values += [
-        f'{INDENT_SIZE*2}<Label value="{label["name"]}" background="{label["color"]}"/>'
+        f'{INDENT_SIZE*2}<Label value="{label["name"]}" background="{COLOR_LOOKUP[label["color"]]}" kern-color="{label["color"]}"/>'
         for label in task_data["labels"]
     ]
     values.append(f"{INDENT_SIZE}</Labels>")
@@ -59,9 +80,9 @@ def __generate_extraction_task(
 #   <View style="box-shadow: 2px 2px 5px #999; padding: 20px; margin-top: 2em; border-radius: 5px;">
 #     <Header value="Choose text sentiment"/>
 #     <Choices name="sentiment" toName="text" choice="single" showInLine="true">
-#       <Choice value="Positive"/>
-#       <Choice value="Negative"/>
-#       <Choice value="Neutral"/>
+#       <Choice value="Positive" style="background:#blue"/>
+#       <Choice value="Negative" style="background:#blue"/>
+#       <Choice value="Neutral" style="background:#blue"/>
 #     </Choices>
 #   </View>
 def __generate_classification_task(
@@ -84,7 +105,7 @@ def __generate_classification_task(
         f'{INDENT_SIZE*2}<Choices name="{task_data["name"]}" toName="{task_data["attribute_name"]}" choice="single" showInLine="true">'
     )
     values += [
-        f'{INDENT_SIZE*3}<Choice value="{label["name"]}"/>'
+        f'{INDENT_SIZE*3}<Choice value="{label["name"]}" style="background:{COLOR_LOOKUP[label["color"]]}50;margin:2px;outline: 2px solid {COLOR_LOOKUP[label["color"]]};outline-offset:-2px;" kern-color="{label["color"]}"/>'
         for label in task_data["labels"]
     ]
     values.append(f"{INDENT_SIZE*2}</Choices>")

controller/transfer/record_export_manager.py

@@ -81,9 +81,9 @@ def get_records_by_options_query_data(
     record_data_query = __get_record_data_query(project_id, row_options)
 
     # task columns part
-    classficiation_task_query = ""
+    classification_task_query = ""
     if task_options and sources_options:
-        classficiation_task_query += __columns_by_table_meta_data_query(
+        classification_task_query += __columns_by_table_meta_data_query(
             project_id, tables_meta_data
         )
 
@@ -96,7 +96,7 @@ def get_records_by_options_query_data(
     {select_part}
         {extraction_appends["SELECT"]}
     {record_data_query}
-    {classficiation_task_query}
+    {classification_task_query}
     {extraction_appends["FROM"]}
     WHERE basic_record.project_id = '{project_id}'
     """
@@ -139,7 +139,7 @@ def __extract_table_meta_classification_data(
 
         for source in sources_options:
             full_table_name = ""
-            addtional_confidence_table_name = ""
+            additional_confidence_table_name = ""
             tablename_dict = {}
             additional_confidence_table = {}
 
@@ -163,8 +163,8 @@ def __extract_table_meta_classification_data(
                 tablename_dict["table_type"] = "task_data"
 
                 if source.get("type") == enums.LabelSource.WEAK_SUPERVISION.value:
-                    addtional_confidence_table_name = f"{full_table_name}_CONFIDENCE"
-                    tables_mapping[addtional_confidence_table_name] = __create_alias()
+                    additional_confidence_table_name = f"{full_table_name}_CONFIDENCE"
+                    tables_mapping[additional_confidence_table_name] = __create_alias()
                     additional_confidence_table["original_table"] = tables_mapping.get(
                         full_table_name
                     )
@@ -176,20 +176,20 @@ def __extract_table_meta_classification_data(
                 alias = tables_mapping.get(full_table_name)
                 tables_meta_data[alias] = tablename_dict
                 if additional_confidence_table:
-                    tables_mapping[addtional_confidence_table_name] = __create_alias()
+                    tables_mapping[additional_confidence_table_name] = __create_alias()
                     tables_meta_data[
-                        tables_mapping.get(addtional_confidence_table_name)
+                        tables_mapping.get(additional_confidence_table_name)
                     ] = additional_confidence_table
                 if with_user_id:
                     additional_created_by_table = {}
                     additional_created_by_table["table_type"] = "user_data"
                     additional_created_by_table["original_table"] = tables_mapping.get(
                         full_table_name
                     )
-                    addtional_created_by_name = full_table_name + "__created_by"
-                    tables_mapping[addtional_created_by_name] = __create_alias()
+                    additional_created_by_name = full_table_name + "__created_by"
+                    tables_mapping[additional_created_by_name] = __create_alias()
                     tables_meta_data[
-                        tables_mapping.get(addtional_created_by_name)
+                        tables_mapping.get(additional_created_by_name)
                     ] = additional_created_by_table
     return tables_meta_data, tables_mapping
 
@@ -243,7 +243,10 @@ def __record_data_by_type(project_id: str, row_options: Dict[str, Any]) -> str:
 def ___record_data_by_slice(project_id: str, slice_id: str) -> str:
     slice = data_slice.get(project_id, slice_id)
     slice_type = slice.slice_type
-    if slice_type == enums.SliceTypes.STATIC_DEFAULT.value:
+    if (
+        slice_type == enums.SliceTypes.STATIC_DEFAULT.value
+        or slice_type == enums.SliceTypes.STATIC_OUTLIER.value
+    ):
         return __record_data_by_static_slice_query(project_id, slice_id)
     elif slice_type == enums.SliceTypes.DYNAMIC_DEFAULT.value:
         if not slice.filter_data:

graphql_api/mutation/data_slice.py

@@ -109,7 +109,12 @@ def mutate(self, info, project_id, embedding_id):
         auth.check_demo_access(info)
         auth.check_project_access(info, project_id)
         user_id = auth.get_user_id_by_info(info)
-        manager.create_outlier_slice(project_id, user_id, embedding_id)
+        data_slice_item = manager.create_outlier_slice(
+            project_id, user_id, embedding_id
+        )
+        notification.send_organization_update(
+            project_id, f"data_slice_created:{str(data_slice_item.id)}"
+        )
         return CreateOutlierSlice(ok=True)