🚧 This Package is deprecated. Please use the square-model-client or square-datastore-client packages. 🚧

SQuARE Skill Helpers

This package facilitates the interaction with UKP-SQuARE APIs. Currently, interacting with the Models and Datastores is supported. This helps when implementing Skills.

Installation

To install the latest stable version:

pip install square-skill-helpers

To install from the master branch:

pip install git+https://github.com/UKP-SQuARE/square-skill-helpers.git

Usage

After installing, the Datastore and Model API can be called easily. Before running the code, environment variables need to be set.

• KEYCLOAK_BASE_URL: The address where tokens can be obtained from. For UKP-SQuARE set this to https://square.ukp-lab.de
• REALM: The realm in which your Keycloak client resides. For UKP-SQuARE this is square.
• CLIENT_ID: The ID of your client. For UKP-SQuARE, you will receive this from the UI when creating a new skill.
• CLIENT_SECRET: The secret of your client/skill. For UKP-SQuARE, you will receive this from the UI when creating a new skill.

from square_skill_helpers import DataAPI, ModelAPI

if __name__ == "__main__":

    # 1. Initialize square api objects
    data_api = DataAPI()
    model_api = ModelAPI()

    query = "When was TU Darmstadt established?"
    # 2. Call DataAPI
    data = await data_api(datastore_name="nq", index_name="dpr", query=query)
    context = [d["document"]["text"] for d in data]
    context_score = [d["score"] for d in data]
    for i, (c, s) in enumerate(zip(context, context_score)):
        print(f"#{i} (score={s}) {c[:50] + '...'}")
    # #1 (score=70.325) "Technische Universität Darmstadt The Technische U...
    # #2 (score=69.928) "TU9, a network of the most notable German ""Techn...
    # #3 (score=69.208) "Texas Lutheran University Texas Lutheran Universi...
    # ...

    # 3. Call ModelAPI
    model_request = {
        "input": [[query, c] for c in context],
        "task_kwargs": {"topk": 1},
        "adapter_name": "qa/squad2@ukp"
    }

    model_api_output = await model_api(
        model_name="bert-base-uncased", 
        pipeline="question-answering", 
        model_request=model_request
    )
    print(model_api_output['answers'][0])
    # [{'score': 0.9184646010398865, 'start': 275, 'end': 279, 'answer': '1877'}]