Key value extraction improvements (#1573)

* Added example for image key value extraction evaluation Signed-off-by: Yoav Katz <[email protected]> * Removed unneeded comments Signed-off-by: Yoav Katz <[email protected]> * Model key value extraction as a first class task in catalog Signed-off-by: Yoav Katz <[email protected]> * Added documentation link Signed-off-by: Yoav Katz <[email protected]> * Added default template to extraction task Signed-off-by: Yoav Katz <[email protected]> * Changed order of results printout for clarity Signed-off-by: Yoav Katz <[email protected]> * Update docs/docs/examples.rst Co-authored-by: Elron Bandel <[email protected]> * Ensure key values are strings as expected by the metric Signed-off-by: Yoav Katz <[email protected]> * Moved to more standard font across Linux and Mac Signed-off-by: Yoav Katz <[email protected]> * Moved to default font Signed-off-by: Yoav Katz <[email protected]> * Doc improvements Signed-off-by: Yoav Katz <[email protected]> * Improved parsing and testing of processing. Signed-off-by: Yoav Katz <[email protected]> * Changed key value extraction to use json serializer Signed-off-by: Yoav Katz <[email protected]> --------- Signed-off-by: Yoav Katz <[email protected]> Co-authored-by: Elron Bandel <[email protected]>
IBM · Feb 6, 2025 · d21a884 · d21a884
1 parent d9e01f8
commit d21a884
Show file tree

Hide file tree

Showing 5 changed files with 77 additions and 6 deletions.
diff --git a/prepare/templates/key_value_extraction/templates.py b/prepare/templates/key_value_extraction/templates.py
@@ -1,5 +1,11 @@
 from unitxt import add_to_catalog
 from unitxt.processors import PostProcess
+from unitxt.serializers import (
+    DictAsJsonSerializer,
+    ImageSerializer,
+    ListSerializer,
+    MultiTypeSerializer,
+)
 from unitxt.struct_data_operators import JsonStrToListOfKeyValuePairs
 from unitxt.templates import (
     InputOutputTemplate,
@@ -10,7 +16,12 @@
         instruction="Extract the key value pairs from the input. Return a valid json object with the following keys: {keys}. Return only the json representation, no additional text or explanations.",
         input_format="{input}",
         output_format="{key_value_pairs_answer}",
-        postprocessors=[PostProcess(JsonStrToListOfKeyValuePairs())],
+        postprocessors=[
+            PostProcess(JsonStrToListOfKeyValuePairs()),
+        ],
+        serializer=MultiTypeSerializer(
+            serializers=[ImageSerializer(), DictAsJsonSerializer(), ListSerializer()]
+        ),
     ),
     "templates.key_value_extraction.extract_in_json_format",
     overwrite=True,

diff --git a/src/unitxt/catalog/templates/key_value_extraction/extract_in_json_format.json b/src/unitxt/catalog/templates/key_value_extraction/extract_in_json_format.json
@@ -10,5 +10,19 @@
                 "__type__": "json_str_to_list_of_key_value_pairs"
             }
         }
-    ]
+    ],
+    "serializer": {
+        "__type__": "multi_type_serializer",
+        "serializers": [
+            {
+                "__type__": "image_serializer"
+            },
+            {
+                "__type__": "dict_as_json_serializer"
+            },
+            {
+                "__type__": "list_serializer"
+            }
+        ]
+    }
 }
diff --git a/src/unitxt/serializers.py b/src/unitxt/serializers.py
@@ -1,5 +1,6 @@
 import csv
 import io
+import json
 from abc import abstractmethod
 from typing import Any, Dict, List, Union
 
@@ -61,6 +62,13 @@ def serialize(self, value: Any, instance: Dict[str, Any]) -> str:
         return ", ".join(str(item) for item in value)
 
 
+class DictAsJsonSerializer(SingleTypeSerializer):
+    serialized_type = dict
+
+    def serialize(self, value: Any, instance: Dict[str, Any]) -> str:
+        return json.dumps(value)
+
+
 class DialogSerializer(SingleTypeSerializer):
     serialized_type = Dialog
 

diff --git a/src/unitxt/struct_data_operators.py b/src/unitxt/struct_data_operators.py
@@ -23,7 +23,6 @@
     {"key1": "value1", "key2": value2, "key3": "value3"}
 """
 
-import ast
 import json
 import random
 from abc import ABC, abstractmethod
@@ -43,6 +42,7 @@
 from .operators import FieldOperator, InstanceOperator
 from .random_utils import new_random_generator
 from .serializers import ImageSerializer, TableSerializer
+from .type_utils import isoftype
 from .types import Table
 from .utils import recursive_copy
 
@@ -1025,16 +1025,21 @@ def process_value(self, table: Any) -> Any:
 
 
 class JsonStrToListOfKeyValuePairs(FieldOperator):
-    def process_value(self, text: str) -> List[Tuple[str, str]]:
-        text = text.replace("null", "None")
+    """Convert a Json string of representing key value as dictionary to list of key value pairs."""
 
+    def process_value(self, text: str) -> List[Tuple[str, str]]:
         try:
-            dict_value = ast.literal_eval(text)
+            dict_value = json.loads(text)
         except Exception as e:
             UnitxtWarning(
                 f"Unable to convert input text to json format in JsonStrToListOfKeyValuePairs due to {e}. Text: {text}"
             )
             dict_value = {}
+        if not isoftype(dict_value, Dict[str, Any]):
+            UnitxtWarning(
+                f"Unable to convert input text to dictionary in JsonStrToListOfKeyValuePairs. Text: {text}"
+            )
+            dict_value = {}
         return [
             (str(key), str(value))
             for key, value in dict_value.items()

diff --git a/tests/library/test_struct_data_operators.py b/tests/library/test_struct_data_operators.py
@@ -4,6 +4,7 @@
     DuplicateTableColumns,
     DuplicateTableRows,
     InsertEmptyTableRows,
+    JsonStrToListOfKeyValuePairs,
     ListToKeyValPairs,
     LoadJson,
     MapHTMLTableToJSON,
@@ -709,3 +710,35 @@ def test_insert_empty_table_rows(self):
             targets=targets,
             tester=self,
         )
+
+    def test_json_str_to_list_of_key_value_pairs(self):
+        inputs = [
+            {
+                "prediction": """
+{ "a": null , "b" : 3,  "c" : "word" }
+"""
+            }
+        ]
+
+        targets = [{"prediction": [("b", "3"), ("c", "word")]}]
+
+        check_operator(
+            operator=JsonStrToListOfKeyValuePairs(field="prediction"),
+            inputs=inputs,
+            targets=targets,
+            tester=self,
+        )
+
+        check_operator(
+            operator=JsonStrToListOfKeyValuePairs(field="prediction"),
+            inputs=[{"prediction": "bad input"}],
+            targets=[{"prediction": []}],
+            tester=self,
+        )
+
+        check_operator(
+            operator=JsonStrToListOfKeyValuePairs(field="prediction"),
+            inputs=[{"prediction": "3"}],
+            targets=[{"prediction": []}],
+            tester=self,
+        )