Skip to main content
Use the document classifier node to classify your documents.
This feature is currently in public preview. Functionality and behavior may change in future updates.

​
Pre-requisites

Run the following command to enable watsonx Orchestrate Developer Edition to process documents:
BASH
orchestrate server start -e <.env file path> -d
Note: You need to configure a minimum allocation of 20GB RAM to your Docker engine during installation of watsonx Orchestrate Developer edition to support document processing features.
Note: To run the document classifier, you must define the WO_INSTANCE, WO_API_KEY, and AUTHORIZATION_URL credentials in your .env file. For more information on configuring the .env file, see Installing the watsonx Orchestrate Developer Edition.

​
Configuring document extractor node in agentic workflows

  1. Define document classes Create a class that defines the document classes to classify. Each document class must follow this structure:
    Python
    class CustomClasses(BaseModel):
    invoice: DocClassifierClass = Field(default=DocClassifierClass(class_name="Invoice"))
  2. Configure the document extract node
Include a call to the docclassifier() method in your agentic workflow to classify the document. This method accepts the following input arguments:
ParameterTypeRequiredDescription
namestringYesUnique identifier for the node.
llmstringYesThe LLM used for document classification.
display_namestringNoDisplay name for the node.
classesobjectYesThe document classification classes.
descriptionstringNoDescription of the node.
min_confidencefloatNoMinimum confidence threshold for classification.
review_fieldsList[string]NoThe fields that require user review.
input_mapDataMapNoDefine input mappings using a structured collection of Assignment objects.
enable_reviewboolNoEnables or disables the human-in-the-loop feature. Set to True to activate it and False to deactivate. The default value is False.
Note:The min_confidence and review_fields settings control the human-in-the-loop feature. This feature only works when you run the Flow from a chat session. If a field is extracted with confidence lower than min_confidence, and its name appears in review_fields, the agent opens a review window in the chat. You can then review and confirm the extracted values.
Example use of the docext node in a agentic workflow:
Python
from pydantic import BaseModel, Field
from ibm_watsonx_orchestrate.flow_builder.flows import (
    Flow, flow, START, END
)
from ibm_watsonx_orchestrate.flow_builder.types import DocClassifierClass, DocumentProcessingCommonInput, DocumentClassificationResponse


class CustomClasses(BaseModel):
    buyer: DocClassifierClass = Field(default=DocClassifierClass(class_name="Buyer"))
    seller: DocClassifierClass = Field(default=DocClassifierClass(class_name="Seller"))
    agreement_date: DocClassifierClass = Field(default=DocClassifierClass(class_name="Agreement_Date"))


@flow(
    name ="custom_flow_docclassifier_example",
    display_name="custom_flow_docclassifier_example",
    description="Extraction of custom classes from a document, specified by the user.",
    input_schema=DocumentProcessingCommonInput
)
def build_docclassifier_flow(aflow: Flow = None) -> Flow:
    # aflow.docclassifier return a DocClassifierNode object
    # DocumentClassificationResponse is the output schema of DocClassifierNode and it can be used as input schema for the next node

    doc_classifier_node = aflow.docclassifier(
        name="document_classifier_node",
        display_name="document_classifier_node",
        description="Classify custom classes from a document",
        llm="watsonx/meta-llama/llama-3-2-90b-vision-instruct",
        classes=CustomClasses(),
    )

    aflow.sequence(START, doc_classifier_node, END)
    return aflow
⌘I