Skip to content

LLM-Driven Extraction of Unstructured Data — Built for API Deployments & ETL Pipeline Workflows

fix: Prompt studio tool export SPEC updated (#132)

Chandrasekharan M239b6c432585
diff --git a/backend/tool_instance/tool_instance_helper.py b/backend/tool_instance/tool_instance_helper.py
index df6d0fed9..f6c7fc0ca 100644
--- a/backend/tool_instance/tool_instance_helper.py
+++ b/backend/tool_instance/tool_instance_helper.py
@@ -4,11 +4,12 @@ import uuid
from json import JSONDecodeError
from typing import Any, Optional
+from account.constants import Common
from account.models import User
from adapter_processor.adapter_processor import AdapterProcessor
from adapter_processor.models import AdapterInstance
from connector.connector_instance_helper import ConnectorInstanceHelper
-from jsonschema import ValidationError
+from jsonschema.exceptions import UnknownType, ValidationError
from tool_instance.constants import JsonSchemaKey
from tool_instance.models import ToolInstance
from tool_instance.tool_processor import ToolProcessor
@@ -17,9 +18,8 @@ from unstract.sdk.tool.validator import DefaultsGeneratingValidator
from unstract.tool_registry.constants import AdapterPropertyKey
from unstract.tool_registry.dto import Spec, Tool
from unstract.tool_registry.tool_utils import ToolUtils
-from workflow_manager.workflow.constants import WorkflowKey
from utils.local_context import StateStore
-from account.constants import Common
+from workflow_manager.workflow.constants import WorkflowKey
logger = logging.getLogger(__name__)
@@ -340,10 +340,16 @@ class ToolInstanceHelper:
) -> tuple[bool, str]:
"""Function to validate Tools settings."""
tool: Tool = ToolProcessor.get_tool_by_uid(tool_uid=tool_uid)
- tool_name: str = tool.properties.display_name if tool.properties.display_name else tool_uid
+ tool_name: str = (
+ tool.properties.display_name
+ if tool.properties.display_name
+ else tool_uid
+ )
# Getting user from local_context store
user: User = StateStore.get(Common.USER_ID)
- schema_json: dict[str, Any] = ToolProcessor.get_json_schema_for_tool(tool_uid= tool_uid, user= user)
+ schema_json: dict[str, Any] = ToolProcessor.get_json_schema_for_tool(
+ tool_uid=tool_uid, user=user
+ )
try:
DefaultsGeneratingValidator(schema_json).validate(tool_meta)
return True, ""
@@ -351,3 +357,5 @@ class ToolInstanceHelper:
return False, str(e)
except ValidationError as e:
return False, str(tool_name + ": " + e.schema["description"])
+ except UnknownType as e:
+ return False, str(e)
diff --git a/unstract/tool-registry/src/unstract/tool_registry/dto.py b/unstract/tool-registry/src/unstract/tool_registry/dto.py
index 7479807d5..084ba0091 100644
--- a/unstract/tool-registry/src/unstract/tool_registry/dto.py
+++ b/unstract/tool-registry/src/unstract/tool_registry/dto.py
@@ -61,7 +61,7 @@ class Spec:
title: str = ""
description: str = ""
- type: str = ""
+ type: str = "object"
required: list[str] = field(default_factory=list)
properties: dict[str, dict[str, Any]] = field(
default_factory=dict[str, dict[str, Any]]
@@ -122,7 +122,7 @@ class Spec:
return self.get_adapter_properties(AdapterTypes.X2TEXT)
def get_ocr_adapter_properties(
- self,
+ self,
) -> dict[str, dict[str, Any]]:
return self.get_adapter_properties(AdapterTypes.OCR)
@@ -243,7 +243,6 @@ class Adapter:
TEXT_EXTRACTORS_KEY = "textExtractors"
OCRS_KEY = "ocrs"
-
language_models: list[AdapterProperties] = field(
default_factory=list[AdapterProperties]
)
@@ -284,10 +283,7 @@ class Adapter:
AdapterProperties.from_dict(text_extractor_dict)
for text_extractor_dict in text_extractors_list
]
- ocrs = [
- AdapterProperties.from_dict(ocr_dict)
- for ocr_dict in ocrs_list
- ]
+ ocrs = [AdapterProperties.from_dict(ocr_dict) for ocr_dict in ocrs_list]
return cls(
language_models=language_models,