.. _DocKV :

Document Key-Value Extraction Agent 
====================================

The DocKV Agent is a smart AI-powered tool that extracts key-value pairs from a given image based on input keys and their short descriptions. Currently AgentForce support three differnet versions for the docKV agent. More versions will be coming soon.

1. Agents Input Dictionary
---------------------------

Firstly an Agents-input Dictionary has to be defined. For a DocKV agent, the agent input dictionary is of the format :

.. code-block:: console

    inputs_dict = {
        "fields" : {
            # The descriptions are supposed to be spatially aware.
            "field_1" : "description_1",
            "field_2" : "description_2",
            },
        "base64_image" : img_processed,
        "image_type" : "jpeg" # PNG or Whatever image is being used
    }

2. Agent's Process Functionality 
---------------------------------

After setting an inputs dictionary, we set up an agent_name. 

.. code-block:: console

    agent_name = latentforce/doc_kv/v1
    res = agent.process(agent_name = agent_name, agent_inputs = inputs_dict)
    print(res)

Available DocKV versions are listed down below. 

1. latentforce/doc_kv/v1
2. latentforce/doc_kv/v2
3. latentforce/doc_kv/v3